RU2806618C2 - Pressure sensor for introduction to the human circular system - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: pressure sensor for insertion into the human circulatory system, a method for measuring pressure, an implantable medical device, and a pressure sensor reading system. The pressure sensor is a passive sensor and contains a housing, a magnetomechanical oscillation generator, a torque recovery unit. The body contains a diffusion-blocking layer covering a part of the body and configured to maintain a predetermined pressure inside the body. The oscillation generator is designed to convert an external magnetic or electromagnetic excitation field into a mechanical oscillation of a magnetic object and contains a magnetic object that provides a constant magnetic torque inside the housing. The housing part is flexible to allow changes in external pressure to be converted into changes in the mechanical vibration of the magnetic object. The magnetic object is positioned inside the housing in such a way that it can rotate beyond the equilibrium orientation if an external magnetic or electromagnetic excitation field acts on the magnetic object. The torque recovery unit contains an additional magnetic object for creating a magnetic field in such a position of the magnetic object that it provides a restoring torque. The torque recovery unit is configured to provide a recovery torque to return the magnetic object to an equilibrium orientation. The pressure sensor is designed in such a way that changes in external pressure can be converted into changes in resonant frequency. The implantable medical device includes a pressure sensor. The reading system contains a field generator, a converter, a processor and is designed for wireless reading of pressure sensor readings. The field generator is designed to generate a magnetic or electromagnetic excitation field to induce mechanical oscillations of the magnetic object of the pressure sensor. The transducer is designed to convert the magnetic or electromagnetic field generated by induced mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals. The processor is designed to determine the pressure value based on electrical response signals. The processor is configured to apply a compensation algorithm to correct the determination of the pressure value for the dependence of the electrical response signals on the distance between the pressure sensor and the field generator, the phase of the mechanical oscillation of the magnetic object, the orientation of the housing relative to the reading system, the amplitude of the mechanical oscillation of the magnetic object. When executing the method, an excitation field is generated to induce mechanical oscillations of the magnetic object of the pressure sensor. Transform the field into response signals. The pressure value is determined based on the electrical response signals.

EFFECT: pressure measurement sensor is provided for introduction into a human circulatory system, improving the accuracy of pressure measurement using such a sensor.

15 cl, 34 dwg

Description

Field of technology to which the invention relates

The invention relates to a pressure sensor for insertion into the human circulatory system. The invention also relates to a stent, a hepatic bypass device, a wire for the treatment of a cerebral aneurysm and a heart valve, respectively, containing a pressure sensor. The invention further relates to a sensing system, method and computer program for wirelessly sensing a pressure sensor.

State of the art

Patent document US 2007/0236213 A1 describes pressure measurement based primarily on mechanical resonators with attached magnetized material. A magnetic field can interact with a magnetized material and trigger mechanical vibrations. Mechanical vibrations are then detected by recording a time-varying field due to the oscillating mechanical structure. The recording device may be a coil or other suitable magnetometer. External pressure on such a device may change the actual spring constants and therefore result in a change in the resonant frequency that can be detected. In this way a pressure sensor is formed.

Although this system works in principle, it has a number of disadvantages and is not suitable for measuring pressure deep in a patient's body with sufficient accuracy and with a small enough device. The main problem is the use of a mechanical resonator. Typically, the maximum possible quality factor (quality factor) achievable in mechanical resonators is too low to operate effectively. There are some materials, such as fused silica, that provide high vibration quality. These materials are usually quite hard and do not allow for a high enough vibration amplitude (a large enough angle) to work effectively, i.e. to create a sufficiently large field variation. The next problem is the low sensitivity of the resonant frequency to external pressure, since only the elastic parameters change. This, combined with the low quality factor, results in the need for a fairly high signal-to-noise ratio, which in turn results in the need for a large amount of magnetic material, making the sensor large.

Moreover, implantable pressure sensors already exist, for example those developed by CardioMems and disclosed in US patent document 7,147,604 B1. These sensors operate on the principle of a resonant LC (inductive capacitance) device. The resonant frequency is shifted by mechanical motion caused by pressure, which in turn changes the value of L or C (or both). Although this system works, it cannot be reduced to the size needed to be introduced into the human circulatory system. The main reason is that in US Pat. No. 7,147,604 B1, the detected signal is set to be proportional to the large radius power of the pressure sensor coil. This creates a severe constraint on downsizing the pressure sensor.

Disclosure of the invention

It is therefore an object of the present invention to provide a pressure sensor which is designed to be inserted into the human circulatory system, has a very small size and yet provides high quality pressure measurement. The invention also relates to a stent, a hepatic bypass device, a wire for the treatment of a cerebral aneurysm and a heart valve, respectively, containing a pressure sensor. The invention further relates to a sensing system, method and computer program for wirelessly sensing a pressure sensor.

In a first aspect of the present invention, there is provided a pressure sensor for insertion into the human circulatory system, wherein the pressure sensor is a passive sensor capable of being wirelessly read by a sensing system located externally to the person, wherein the pressure sensor comprises a housing having a diffusion blocking layer, which covers at least a portion of the housing, and which is configured to maintain a predetermined pressure inside the housing, wherein the pressure sensor contains, inside the housing, a magnetomechanical oscillator with a magnetic object providing a constant magnetic moment, wherein the magnetomechanical oscillation generator is configured to convert an external magnetic or an electromagnetic excitation field into mechanical vibration of a magnetic object, wherein at least a portion of the housing is flexible to allow changes in external pressure to be converted into changes in mechanical vibration of the magnetic object, wherein the magnetic object is located within the housing such that it can be rotated beyond an equilibrium orientation, if the magnetic object is subject to an external magnetic or electromagnetic excitation field, wherein the pressure sensor further comprises a torque restoration unit configured to provide a restoring torque to return the magnetic object to an equilibrium orientation if the external magnetic or electromagnetic excitation field has rotated the magnetic object out of equilibrium orientation to provide mechanical oscillation of the magnetic object at a resonant frequency, wherein the torque restoration unit includes an additional magnetic object, which is a permanent magnet, for creating a magnetic field in such a position of the magnetic object that it provides restoring torque, and the pressure sensor is configured in such a way that changes in external pressure can be converted into changes in resonant frequency.

Since the pressure sensor uses a magnetomechanical oscillator (oscillator), which mainly uses magnetism to store vibration energy, high quality factor, high vibration amplitude and high sensitivity to external influences can be achieved at the same time by using a pressure sensor with a very small size suitable for insertion into the human circulatory system.

The dimensions of the pressure sensor are preferably such that it can be introduced into the circulatory system outside the main pulmonary artery. Preferably, the pressure sensor has an elongated shape with a maximum size less than or equal to 5 mm, even more preferably less than or equal to 4 mm, and a minimum size less than or equal to 1 mm. These dimensions mainly refer to the longitudinal direction and transverse direction of the pressure sensor. Thus, preferably, in the longitudinal direction, the size of the pressure sensor is less than or equal to 5 mm, even more preferably, less than or equal to 4 mm, and in the transverse direction, the size of the pressure sensor is less than or equal to 1 mm.

In a preferred embodiment, the pressure sensor comprises an external biocompatible coating. In addition, preferably the flexible housing portion includes a bellows to enable changes in external pressure to be converted into changes in mechanical vibration of the magnetic object. The bellows is particularly suitable for allowing changes in external pressure to be converted into changes in the mechanical vibrations of a magnetic object, since it provides flexibility sufficient to measure pressure even if it is coated, for example, with a diffusion blocking layer and/or a biocompatible coating. The bellows can be made from a fairly flexible material such as silicone rubber.

The pressure sensor may further comprise an outer casing over the bellows. An outer cover can be used to prevent blood clots (thrombi) from forming. The outer casing is preferably soft enough to allow the bellows to change in response to external pressure.

The diffusion blocking layer predominantly contains metal. For example, it can be in the form of a metal coating on the housing of the pressure sensor.

In one embodiment, the pressure sensor further includes an outer wire frame attached to the outside of the housing to allow a distance between the outside of the housing and the wall of the vessel to be maintained. A wire frame may be preferred for direct delivery into the vessel, and the frame can support the pressure transducer without blocking the vessel lumen. The frame may contain a ring-shaped or disk-shaped central part from which curved legs protrude. It can be made from a wire material such as nitinol due to its high flexibility and good biocompatibility. Other materials such as stainless steel or polymers can also be used. An alternative to wire are structures cut from sheet materials that can be bent into shape using a mold and heat treatment. Especially for polymers, injection molding is also possible. For connection to the sensor body, a ring or disk-shaped structure may act as an interface between the frame and the sensor body. The ring-shaped structure may be slotted into the cylindrical casing and may be secured by spring force and/or by gluing or welding. The disc-shaped structure can be glued or welded to the pressure sensor.

The magnetic object is located within the housing such that it can rotate beyond its equilibrium orientation if the magnetic object is subjected to an external magnetic or electromagnetic excitation field, wherein the pressure sensor further comprises a torque recovery unit configured to provide a recovery torque to return a magnetic object to an equilibrium orientation if an external magnetic or electromagnetic excitation field has rotated the magnetic object from an equilibrium orientation to cause the magnetic object to mechanically oscillate at a resonant frequency, wherein the pressure sensor is configured such that changes in the external pressure are converted to changes in the resonant frequency.

The torque restoration unit includes an additional magnetic object for generating a magnetic field at a position of the magnetic object such that it provides restoring torque. Moreover, in addition to the magnetic object, the additional magnetic object is also a permanent magnet. Moreover, the magnetic object is preferably a magnetic sphere. Also, the additional magnetic object may be a magnetic sphere. However, the magnetic object and also the additional magnetic object may also have a different shape. For example, they can be cylindrical. The magnetic object is preferably attached to one end of the thread, with the other end of the thread attached to the housing. Also, an additional magnetic object can be attached to one end of the thread, while the other end of the thread can be attached to the body. However, an additional magnetic object can also be captured.

In one embodiment, the magnetic object and/or the interior of the housing is coated with a slippery and non-stick material. Preferably, the slippery and non-stick material is graphite. Also, the additional magnetic object may be coated with a slippery and non-stick material. Preferably, a non-stick material is considered "slippery" if the coefficient of friction is dry, i.e. without lubrication, is less than 0.2 and even more preferably less than 0.1.

Typically, if a pressure sensor has been inserted into a person's body, a problem may arise if the body needs to be scanned using a magnetic resonance imaging (MRI) scanner. This problem does not pose a danger to the body, i.e. for the patient, since the pressure sensor is relatively small and therefore causes only small forces and torques that do not pose a threat to patients. Similarly, the magnetic resonance image produced by an MRI scanner cannot be damaged either because the pressure sensor is very small in size.

However, many clinical MRI scanners use high field strengths of more than 1.5 Tesla, and the strong magnetic field can destroy the pressure sensor by changing the magnetization of the magnetic object or damaging the mechanical structure within the device. For this reason, in one embodiment, the pressure sensor is designed such that the magnetic object can be aligned with the external magnetic field regardless of the position and orientation of the pressure sensor in the external magnetic field. For example, the pressure sensor may include an outer casing surrounding a housing, wherein the housing may rotate within the female casing, wherein the pressure sensor is configured such that changes in external pressure outside the female casing are converted to changes in external pressure occurring outside the housing and inside the female casing . The outer casing may be spherical or elliptical. In addition, the outer casing may also be filled with a fluid, said fluid preferably being a highly viscous fluid. The term "highly viscous" preferably refers to a viscosity that is such that at a magnetic flux density of 0.1 Tesla, the maximum rotational speed of the device is limited to less than 10,000 degrees per second or an angular velocity of no more than 160 s ^-1 . For typical device configurations, this means a viscosity of 1 to 100 Pa as a minimum. The main determining factor is the volume fraction of the magnetically hard material. There is also a maximum useful viscosity, which is approximately 100-1000 times the minimum useful viscosity.

The casing may be, for example, a very soft casing filled with a fluid, or a casing with openings to effectively convert changes in external pressure into changes in mechanical vibrations of the magnetic object. Preferably, the housing is considered "very soft" if the pressure change within the housing does not deviate more than 0.2 mbar (20 Pa) from the external pressure on the time scale of the expected pressure change, which is typically 0.01 to 1 s. The hardest known substance, brought into the correct form at a sufficiently small thickness, can form a "very soft" casing. The casing then preferably includes structures that act as a bellows.

The magnetic object is preferably a magnetic sphere that is attached to one end of the thread, with the other end of the thread attached to the interior of the housing, wherein in one embodiment the thread has a length of at least Pi/4 the diameter of the magnetic sphere. If also the additional magnetic object is a magnetic sphere, it may be attached to one end of the other thread, with the other end of the other thread attached to the body. It is also preferable that the other thread has a length of at least Pi/4 the diameter of, in this case, the other magnetic sphere. These filament lengths ensure free alignment of magnetic objects with the external magnetic field. The magnetic sphere and the other magnetic sphere preferably have the same diameter.

Moreover, in one embodiment, the magnetic object is a magnetic sphere that is attached to one end of the thread, with the other end of the thread attached to a length changing block configured to change the length of the thread and attached to the interior of the housing. The length changing unit may be, for example, a winding unit. The winding unit may include a winding mechanism. Preferably, the pressure sensor is designed in such a way that the length of the thread from the magnetic sphere to the length changing unit can be adjusted so that it is equal to a predetermined length. For example, the thread and/or length changing unit may include a stopper located and configured to stop while the length changing unit shortens the length of the thread, further shortening the thread if the length of the thread from the magnetic sphere to the length changing block has reached a predetermined length. In one embodiment, the stopper is located and configured to stop winding the thread if the length of the thread from the magnetic sphere to the winding unit has reached a predetermined length.

In one embodiment, the length change unit includes a spring having a spring force, the spring being configured and the thread attached to the spring such that the spring force pushes the thread out of the pressure sensor housing and into the length change unit to shorten the length of the thread in the pressure sensor body. , and in this case the length of the thread in the pressure sensor housing can be increased if the force acting on the magnetic object pulls the thread out of the length change block opposite to the spring force. The length changing unit may include a stopper configured and positioned to limit relaxation of the spring so that the thread inside the pressure sensor housing has a predetermined length if no force acts opposite to the spring force.

In a preferred embodiment, the pressure sensor is configured to compensate for the dependence of the resonant frequency on temperature. In particular, the pressure sensor includes a compensation element that is configured to change a resonant frequency in a first frequency direction depending on a change in temperature, which is opposite to a second frequency direction in which the resonant frequency of the pressure sensor would be changed depending on a change in temperature if the compensation the element was not part of the pressure sensor. Because the measuring device includes a compensation element that is configured to change a resonant frequency in a first frequency direction depending on a change in temperature, which is opposite to a second frequency direction in which the resonant frequency would be changed depending on a change in temperature if the compensation element were not part pressure sensor, temperature-induced shifts in the resonant frequency can be reduced or even eliminated. The first frequency direction represents the direction towards higher or lower frequencies, and the opposite second frequency direction represents the direction towards lower or higher frequencies, respectively.

Preferably, the compensation element comprises a magnetic material that changes its magnetization and therefore its resonant frequency depending on the temperature, wherein the magnetic material is selected and located inside the pressure sensor, in particular inside the housing, such that the direction of change of the resonant frequency is the first frequency direction . The compensating magnetic material is preferably located adjacent to the magnetic object and/or adjacent to the additional magnetic object, if present. This makes it possible to design a pressure sensor in such a way that undesirable temperature dependence can be significantly reduced or even eliminated in a technically relatively simple manner and without the need for large space inside the housing.

In one embodiment, the magnetic object is a magnetic sphere that is attached to one end of the thread, wherein the other end of the thread is directly or indirectly attached to the housing, wherein the magnetic sphere includes a through hole passing through the center of gravity of the magnetic object, wherein one end of the thread located and fixed in the specified through hole. This mount only reduces the magnetic dipole moment by a small fraction and therefore maintains a good signal. The shape of a magnetic object does not change much, which may be important in the case of spheres.

Additionally, in one embodiment, the magnetic object is a magnetic sphere that is attached to one end of the thread, wherein the other end of the thread is directly or indirectly attached to the housing, wherein one end of the thread is sandwiched between two magnetic components of the magnetic object that form the magnetic object . This method of fastening gives results almost as good as the through-hole method, but does not require special equipment to manufacture.

In one embodiment, the magnetic object is a magnetic sphere that is glued to one end of the thread, with the other end of the thread attached directly or indirectly to the housing. This method is technically very simple and makes full use of a magnetic object.

Moreover, in one embodiment, the magnetic object is a magnetic sphere attached to a non-magnetic object, wherein said non-magnetic object is attached to one end of the thread, while the other end of the thread is directly or indirectly attached to the housing. It also makes it relatively easy to attach the thread to a magnetic object.

In another aspect of the present invention, a stent comprising a pressure sensor is provided. For example, a pressure sensor may be located on the distal side of the stent to indicate restenosis within the stent. In one embodiment, the stent includes multiple pressure sensors to, for example, monitor pressure drop across the entire stent or a portion thereof as a parameter for in-stent restenosis. Early detection allows for adjustment of drug treatment or timely re-stenting, thereby avoiding unplanned hospitalization.

In another aspect of the present invention, there is provided a hepatic shunt device comprising a pressure sensor. For example, a pressure sensor may be located on the proximal side of the shunt device to monitor whether a decrease in pressure occurs, i.e. Is the shunt open? Also, the hepatic shunt device may contain several pressure sensors, in particular to monitor pressure drop. In addition, here too, early detection makes it possible to adjust drug treatment or timely re-stenting, thereby avoiding unplanned hospitalization.

In another aspect of the present invention, a wire for treating a cerebral aneurysm is provided that includes a pressure transducer. The wire can be used to skein, which should cause blood to clot to fill the aneurysm space. A pressure sensor can be used to indicate whether blood clotting is actually occurring, i.e. whether the variation (change) in pulsating blood pressure decreases.

In another aspect of the present invention, a heart valve is provided including a pressure sensor. For example, a first pressure sensor may be placed on the proximal side and a second pressure sensor may be placed on the distal side of the heart valve to monitor the pressure drop through the phases of the heartbeat. From the monitored dynamic pressure change, information about the valve's performance can be extracted. A localization sensor can also be placed directly on the moving part of the valve to not only convey pressure information, but also motion information through spatial localization and sensor orientation determination.

In another aspect of the present invention, there is provided a sensing system for wirelessly sensing a pressure sensor according to any one of claims. 1-11, and the reading system contains:

- a field generator for generating a magnetic or electromagnetic excitation field for inducing mechanical vibrations of the magnetic object of the pressure sensor,

- a transducer for converting the magnetic or electromagnetic field generated by the induced mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals,

- a processor for determining a pressure value based on the electrical response signals, wherein the processor is configured to apply a compensation algorithm to adjust the determination of the pressure value for dependence of the electrical response signals on at least one of the following:

a) distance between pressure sensor and field generator;

b) phase of mechanical oscillation of a magnetic object;

c) orientation of the housing relative to the reading system; And

d) amplitude of mechanical vibration of a magnetic object.

The field generator and transducer may be two separate units, or they may be combined, and if the field generator and transducer are combined, the same coils may be used to generate the magnetic or electromagnetic excitation field and to convert the magnetic or electromagnetic field generated by the induced mechanical oscillation magnetic pressure sensor object, into electrical response signals.

In a preferred embodiment, the processor is configured to apply a compensation algorithm to adjust the determination of the pressure value for the dependence of the resonant frequency on at least one of a) the distance between the pressure sensor and the field generator and b) the common mode excitation of the coil.

A large magnetic moment of a magnetic object is desirable because it will create a stronger response to be picked up by the transducer, which may contain suitable sensing coils. However, a large magnetic moment means a large restoring force, which means that the resulting oscillatory motion will have a large amplitude. When large oscillations occur, the restoring force decreases with large angular displacement. Therefore, for such an oscillation, the response frequency will depend on the restoring force, which depends on the distance between the field generator and pressure sensor coils. To solve this problem, the processor may be configured to adjust the determination of the pressure value for the dependence of the resonant frequency on the distance between the pressure sensor and the field generator.

Intravascular pressure changes during the cardiac cycle. The normal human heart rate is usually between 50 and 90 beats per minute, with possible maximums of up to 200 beats per minute. To determine pressure minima and maxima during the cardiac cycle, the measurement frequency should not be less than approximately 5 Hz. Preferably, the measurement frequency is from 10 to 20 Hz, most preferably above 40 Hz. On the other hand, to ensure a good signal-to-noise ratio, a very high quality factor Q of the oscillator is preferred, with a high quality factor Q meaning slow attenuation. Therefore, when the next measurement pulse is transmitted to the sensor, the fluctuations from the previous measurement pulse may not be completely canceled out, and they may affect the measurement. Thus, by compensating for this common-mode excitation of the coil, it is possible to better combine a high quality factor with a measurement frequency large enough to provide measurements of cardiac minima and maxima.

In another aspect of the present invention, there is provided a pressure measurement method for performing a measurement using a pressure sensor according to any one of claims. 1-11, wherein the pressure measurement method comprises:

- generating a magnetic or electromagnetic excitation field to induce mechanical vibrations of the magnetic object of the pressure sensor,

- conversion of the magnetic or electromagnetic field generated by the induced mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals,

- determining the pressure value based on the electrical response signals, wherein in the determining step, the pressure value is adjusted to depend on the electrical response signals from at least one of the following:

a) distance between pressure sensor and field generator;

b) phase of mechanical oscillation of a magnetic object;

c) orientation of the housing relative to the reading system; And

d) amplitude of mechanical vibration of a magnetic object.

Moreover, in another aspect of the present invention, there is provided a computer program that includes software code for causing the sensing system of claim 13 to carry out the steps of a pressure sensing method when the computer program is executed on a computer controlling the sensing system.

It should be understood that the pressure sensor of claim 1, the implantable device of claim 12, in particular the stent, hepatic bypass device, wire and heart valve, the sensing system of claim 13, the pressure measuring method of claim 14 and the durable storage medium, containing a computer program according to claim 15, have similar and/or identical preferred embodiments, in particular as defined in the dependent claims.

It should be understood that a preferred embodiment of the present invention may also be any combination of dependent claims or the above embodiments with a corresponding independent claim.

These and other aspects of the present invention will become apparent and will be explained with reference to the embodiments described below.

Brief description of drawings

The following drawings show:

in fig. 1 schematically and by way of example shows an embodiment of a pressure sensor in a situation with a first external pressure,

in fig. 2 shows schematically and by way of example the embodiment of FIG. 1 in a situation where the second pressure is greater than the first pressure,

in fig. 3 schematically and by way of example shows various embodiments of a pressure sensor with a bellows,

in fig. 4 schematically and by way of example shows an additional embodiment of a pressure sensor,

in fig. 5 schematically and by way of example shows additional embodiments of a pressure sensor with a bellows,

in fig. 6 and 7 show schematically and by way of example an embodiment of a guide wire with a pressure sensor,

in fig. 8 schematically and as an example shows an embodiment of a stent with pressure sensors,

in fig. 9 schematically and by way of example shows an embodiment of a heart valve with pressure sensors,

in fig. 10 schematically and by way of example shows an embodiment of a wire for the treatment of a cerebral aneurysm with pressure sensors,

in fig. 11 schematically and as an example shows an embodiment of a hepatic shunt with pressure sensors,

in fig. 12 schematically and by way of example shows an embodiment of a pressure sensor,

in fig. 13 shows schematically and by way of example the embodiment of FIG. 1 in a situation with a large external magnetic field,

in fig. 14 shows an embodiment of a pressure sensor with a spherical outer casing,

in fig. 15 schematically and by way of example shows an embodiment of a pressure sensor with an elliptical outer casing,

in fig. 16 shows schematically and by way of example an additional embodiment of a pressure sensor with relatively long filaments,

in fig. 17 schematically and as an example shows an embodiment of a measuring (sensing) device with winding blocks and stoppers,

in fig. 18 schematically and by way of example shows an embodiment of a winding unit,

in fig. 19 schematically and by way of example shows an embodiment of a temperature compensated pressure sensor,

in fig. 20 and 21 show schematically and by way of example a detection system for reading the resonant frequency of a sensor,

in fig. 22 shows schematically and by way of example the excitation pulses and the resulting induced voltages,

in fig. 23 is a schematic and exemplary illustration of a multi-coil array built into the mattress of an imaging system patient bed,

in fig. 24 schematically and as an example shows the coils of the detection system,

in fig. Figure 25 shows the frequency spectrum used to determine the resonant frequency,

in fig. 26 schematically and as an example shows an analog reception filter,

in fig. 27 shows, as an example, the frequency response of the passband of a type II Chebyshev filter,

in fig. 28 schematically and as an example shows a calibration installation for calibrating a pressure sensor,

in fig. 29 shows as an example the dependence of the detected response frequency of the sensor on the measured reference pressure,

in fig. 30 shows as an example i) the matching of the detected frequency response of the sensor to the measured reference pressure and ii) the calibration curve,

in fig. 31 shows the results of modeling the sensitivity of the sensor as an example,

in fig. 32 shows the noise level when measuring pressure as an example,

in fig. Figure 33 shows the measured dependence of the signal amplitude in different harmonics on the orientation of the sensor relative to one transceiver coil, and

in fig. 34 illustrates schematically and by way of example an additional embodiment of a pressure sensor.

Carrying out the invention

In fig. 1 schematically and exemplarily illustrates an embodiment of a pressure sensor 501 for insertion into the human circulatory system. The pressure sensor 501 contains a magnetomechanical resonator with two magnetic elements 507, 508.

The magnetic element 508 is suspended from the thread 506 and is thus free to rotate around the main axis of the resonator. In this embodiment, the additional magnetic object 507 is fixed. However, in another embodiment, the additional magnetic element may also be suspended on a thread and thus free to rotate around the main axis of the resonator.

At equilibrium, magnets 507, 508 are respectively aligned with an antiparallel orientation of their magnetization. An external magnetic field pulse can be used to trigger resonant rotational oscillations. The force of attraction determines the resonant frequency of vibration, which for a spherical suspended magnet is defined as:

where M _S is the saturation magnetization of the magnetic material, ρ is its density, r is the diameter of the sphere and B is the field created by a stationary (fixed) magnet. It can be approximated as a dipole field

where m is the magnetic moment of the magnet.

The change in field produced by the oscillating magnetic element may be detected by an induced voltage in one or more detector coils of the transducer, which is configured to convert the magnetic or electromagnetic field generated by the mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals. The time curve of the detected signal (see FIG. 13) can be Fourier transformed to obtain a spectrum (see FIG. 16), which allows the resonant frequency to be determined.

Due to low resonant frequencies of a few kHz, magnetic fields are not shielded by the metal, and therefore all non-ferromagnetic metals can be used as structural or coating materials. Likewise, the sensor can be placed in non-ferromagnetic metal objects without affecting its performance, as long as the thickness of the metal does not greatly exceed the depth of the surface layer (skin layer). At these frequencies, for very good conductors such as copper, the surface layer depth is on the order of one millimeter, while for nitinol the surface layer depth is about 10 millimeters.

A basic magnetomechanical oscillator contains two magnetic elements, and in equilibrium the magnetic elements are aligned with antiparallel magnetization. An external field pulse can be used to trigger rotational oscillation of a suspended sphere around the main axis of the resonator, while another sphere, i.e. additional magnetic object, fixed. If, in another embodiment, also another sphere is suspended in free space and can perform rotational oscillations, both spheres can perform resonant counter-oscillations.

Patent document US 2007/0236213 A1 describes primarily mechanical resonators with attached magnetized material. A magnetic field can interact with a magnetized material and trigger mechanical vibrations. Mechanical vibrations are then detected by recording a time-varying field due to the oscillating mechanical structure. The recording device may be a coil or other suitable magnetometer. External pressure on such a device may change the actual spring constants and therefore result in a change in the resonant frequency that can be detected. In this way a pressure sensor is formed.

Although such a system works in principle, it has, as also mentioned above, a number of disadvantages and is not suitable for measuring pressure deep in the patient's body with sufficient accuracy and using a small enough device. The main problem is the use of a mechanical resonator. Typically, the maximum possible quality factor (quality factor) achievable in mechanical resonators is too low to operate effectively. There are some materials, such as fused silica, that provide high vibration quality. These materials are usually quite hard and do not allow for a high enough vibration amplitude (a large enough angle) to work effectively, i.e. to create a sufficiently large field variation. The next problem is the low sensitivity of the resonant frequency to external pressure, since only the elastic parameters change. This, combined with the low quality factor, results in the need for a fairly high signal-to-noise ratio, which in turn results in the need for a large amount of magnetic material, making the sensor large. Another problem with the device disclosed in US 2007/0236213 A1 is the integration of a high strength permanent magnet into the device. The best permanent magnets are the sintered type. They are not compatible with the microelectromechanical systems (MEMS) manufacturing process. Thus, either the production is complex or the magnetic material of inferior quality must be used. A positive factor is the relatively high operating frequency of the sensor, disclosed in patent document US 2007/0236213 A1. The downside is that the noise in the enclosure also increases with frequency, and above a few hundred kHz there is no more gain observed. So the stated gigahertz resonant frequencies do not help. High frequency operation also requires rapid switching from transmit to receive mode, which is technically challenging. Another problem with the solution from US 2007/0236213 A1 is durability. If the product of amplitude and frequency is high enough, the spring material is subjected to significant stress, which can lead to failure.

These problems can be avoided by the design proposed, for example, in Fig. 1. Since energy is mainly stored in the magnetic field, it is relatively easy to achieve high quality factor. High vibration amplitudes are also possible. The thin thread is not subject to much wear. Resonance can be easily changed by changing the magnetic field by mechanically moving the magnets relative to each other. This is also easily matched to pressure changes (using materials with the right compliance and the right shape, as described below), so a fairly high frequency change can be achieved. The sensor can be made from the best available magnetic material, and the volume fraction of the magnetic material is high.

There already exist, as also explained above, implantable pressure sensors, for example those developed by CardioMeMS and disclosed in patent document US 7147604 B1. These sensors operate on the principle of a resonant LC (inductive capacitance) device. The resonant frequency is shifted by mechanical motion caused by pressure, which in turn changes the value of L or C (or both). Although this system works, it cannot be scaled down to the size required for the intended applications. This is an inherent problem with an LC oscillator. As the size decreases, the level of power that can be generated by the oscillator and the dynamic dipole moment generated by the power are reduced. This can be seen in the following equation. The quality factor of the resonator cannot be higher than the quality factor of the coil. The approximation of the quality factor of the coil can be written as:

where ω is the frequency, μ ₀ is the magnetic permeability of the vacuum, ρ is the resistivity, τ is the fraction of the radius consisting of the conductor, and r is the radius of the coil. It is assumed that the coil has a cylindrical shape with a diameter corresponding to the height. For a copper coil with a diameter of 1 mm at a frequency of 100 kHz, the quality factor is about 1. Such a coil obviously will not work. For the 1 cm (or larger) diameter coils used by CardioMeMS, the quality factor is greater than 100 at 100 kHz and greater than 1000 at 1 MHz. The above formula overestimates practically achievable values of quality factor Q, since it assumes that the entire volume is filled with conductive material, and neglects proximity and skin effects, as well as losses in the capacitor. However, these values result in a working system. Patent document US 7147604 B1 shows a measured quality factor of 48 in the range from 1 to 100 MHz. Since the dynamic dipole moment of an LC oscillator is the quality factor Q times the external magnetic field times the volume, the signal is set proportional to r ⁵ , while in the case of a mechanical oscillator (energy stored in elasticity) the signal is set proportional to r ³ , and in the case of embodiments described, for example, with reference to FIGS. 1 (magnetomechanical oscillation generator, energy stored in a magnetic field), the signal is set proportional to r ² , and the frequency is inversely proportional to the linear dimensions. Thus, the proposal presented here is very well suited for sensor miniaturization.

In a stationary (fixed) sphere embodiment, the stationary sphere may have a diameter of 620 microns, while the oscillating sphere 108 may have a diameter of 500 microns. The magnetic moment of the oscillating sphere 108 may be m≈70 μAm ² , the base frequency may be ƒ ₀ ≈2 kHz, and the quality factor may be approximately Q≈500. The signal to noise ratio (SNR) depends on the distance between a) the coil used to read the resonant frequency and b) the measuring (sensitive) device, as well as the parameters of the coil. For a 10 cm diameter portable coil with 200 turns and 10 ohm resistance, the theoretically achievable SNR at a distance of about 30 cm and with a sampling duration of 0.1 s is approximately 4000. However, typical SNR values of a fixed sphere demonstrator may range from 10 to 100, if practically no measures have been taken to suppress the background signal. Thus, the noise is mainly determined by the harmonic fluctuations of the mains power supply. For half-sphere diameters, i.e. 250 µm for an oscillating sphere, for example, the magnetic moment may be m≈9 µAm ² , the base frequency may be ƒ ₀ ≈4 kHz, the quality factor may remain unchanged, and the theoretical signal-to-noise ratio may drop to around 1000.

There are several ways to attach a thread to a moving magnetic object.

For example, you can use a through-hole mount. In this case, the hole is drilled through the center of gravity and approximately perpendicular to the magnetization. Although the magnet material is hard and brittle, there are several ways to drill holes, such as pulsed laser or electrical discharge machining (EDM). The thread is passed through the hole and glued in the right place. Passing is best done using a vacuum suction process. Several types of glue can be used. Light-curing adhesives are economical. They must have a low viscosity to fill the filament hole simply by capillary force. Alternatively or in addition, the thread may be attached to the magnetic object by mechanical means. For example, through a knot in the thread or some other thick part of the thread, such as a drop of glue or a (molten) ball formed by heat. The latter is especially easily made from supermolecular polyethylene (UHMWPE) fibers. This mounting method only reduces the magnetic dipole moment by a small fraction and, therefore, produces a good signal. The shape of a magnetic object does not change much, which may be important in the case of spheres.

A clamp mount can also be used. In this case, the magnetic object is separated into at least two components. Preferably, a separation plane is created that is orthogonal to the magnetization and parallel to the direction of filament attachment. The thread is placed on this plane. Precise alignment is not required. The second magnetic part is placed on top. Magnetic parts are usually held together by magnetic forces. Finally, glue is applied to secure everything in place. The preferred types of adhesive are the same as for the through-hole fastening process. Additionally, a groove can be machined into one or both of the magnetic objects to reduce the overall gap between the magnetic objects. This method produces results almost as good as the through-hole method, but does not require special equipment to manufacture. Typically, magnetic sub-objects are created not by dividing one complete magnetic object, but by grinding together two (identical) magnetic objects. The downside is that this process is more wasteful since two input objects are used, and the process can also be somewhat more labor intensive.

The cheapest way is to directly attach the thread to a magnetic object using a suitable adhesive. The magnetic object is held and aligned using some kind of tool. Both functions can be realized using suitable magnetic fields. The tool may be shaped like a funnel with a thread running through the funnel, with a magnetic object being attached to the opening of the funnel by magnetic forces. The glue is applied into the funnel and cured. Then the knot is removed from the tool and the unnecessary part of the thread is cut off. This method can be very cheap and makes full use of the magnetic object. The disadvantage is that a significant amount of material is added, which reduces the vibration frequency and takes up space in the finished device.

In another embodiment, a structure for fastening and additional gluing may be used. You can attach a thread to a magnetic object by first attaching it to a non-magnetic object, and then sticking the non-magnetic object to a magnetic object. The non-magnetic object may be manufactured by injection molding or an equivalent low-cost method. The shape of the non-magnetic object should allow the thread to be easily attached, i.e. it may have a hole or a clamping mechanism, perhaps even something as simple as a groove. The non-magnetic object is then glued to the magnetic object. Alternatively, it can be clamped or screwed to a magnetic object. This method is simple and cheap, but some applications may require too much extra space.

In principle, all the considered methods of attaching a thread to a magnetic object apply in the same way to attaching a thread to a body. However, since the body material is usually easier to work with, the through-hole method may be a good choice. Clamping is also a good option. This may be cheaper, but may be more difficult to achieve the final seal.

At least a portion of the housing is flexible to allow changes in external pressure to be converted into changes in mechanical vibrations of the magnetic object. Preferably the housing includes a deflectable membrane, as shown schematically and by way of example in FIG. 1 and 2. The deflection depends on the pressure applied to the sensor and changes the distance between the spheres. Decreasing the distance leads to an increase in the resonant frequency, and vice versa. In fig. 1 and 2 you can see the basic operating principle of the pressure sensor. An increase in pressure deflects the membrane 515 and reduces the distance between the spheres 507, 508, which results in an increase in the resonant frequency. In fig. 1 and 2 further show the housing 502 and the thread 506 by which the magnetic sphere 508 is attached to the membrane 515. FIG. 1, the pressure acting on the membrane and the resonant frequency are less than in FIG. 2.

In fig. 3 shows the designs of bellows, i.e. additional versions of the pressure sensor with a bellows. The bellows is subject to forces resulting from the pressure acting on the sensor, i.e. an increase in pressure compresses the bellows and reduces the intersphere distance. In fig. 3A shows a first bellows design. The bellows 703 is designed to utilize the space available around the filament 706 without increasing the outer diameter of the sensor, whereby the increase in pressure compresses the bellows 703 and reduces the distance between the magnetic spheres 707, 708, resulting in an increase in the resonant frequency. In fig. 3A further shows the housing 702 and stationary magnetic sphere 707 of the pressure sensor 701. The pressure sensor 701 further includes a thin metal coating 717 that acts as a diffusion barrier, i.e. diffusion blocking layer. It should be noted that all embodiments of the present invention include a diffusion blocking layer, even if it is not explicitly shown in all drawings for reasons of clarity.

In fig. 3B shows a coated sensor 801 that is similar to the pressure sensor shown in FIG. 3A, with an additional smooth and soft 818 cover over the bellows to avoid blood clots. The sensor 801 also includes a bellows 803 designed to utilize the space available around the filament without increasing the outer diameter of the sensor, whereby the increase in pressure compresses the bellows 803 and reduces the distance between the magnetic spheres 807, 808, resulting in an increase in the resonant frequency. In fig. 3B further shows the housing 802 and stationary magnetic sphere 807 of the pressure sensor 801. The pressure sensor 801 further includes a thin metal coating 817 that acts as a diffusion barrier, i.e. diffusion blocking layer.

In fig. 3C shows a pressure sensor 901, which is similar to the pressure sensor shown in FIG. 3B, additionally with a 3-piece 920 wire frame for direct delivery to the vessel. The frame fixes the sensor without blocking the lumen of the vessel. Thus, also in this embodiment, a smooth and soft cover 918 is present over the bellows 903 to avoid the formation of blood clots. The bellows 903 is designed to utilize the space available around the filament without increasing the outer diameter of the sensor, whereby the increase in pressure compresses the bellows 903 and reduces the distance between the magnetic spheres 907, 908, resulting in an increase in the resonant frequency. In fig. 3C further shows the housing 902 and stationary magnetic sphere 907 of the pressure sensor 901. The pressure sensor 901 further includes a thin metal coating 917 that acts as a diffusion barrier, i.e. diffusion blocking layer

The bellows can be manufactured in various ways. First, the bellows can be made from a fairly flexible material such as silicone rubber (see FIG. 4). In fact, it may just be a piece of silicone rubber. In fig. 4, the pressure sensor 1001 includes a cylindrical body 1002 having open ends closed by rubber elements 1009, 1003, the first rubber element 1009 holding a stationary magnetic sphere 1007, and the second rubber element 1003 holding a rotating-oscillating magnetic sphere 1008 using a thread 1006. The cylindrical rubber element 1003 acts as an expansion joint instead of a bellows.

However, when including at least one diffusion-dense layer in the bellows, i.e. by coating the bellows with a diffusion blocking layer, for example as described above with reference to FIG. 3A-3C, a simple tube is usually too hard. Therefore, a true bellows structure is preferred. Bellows are well known, and bellows of various shapes are possible. The origami-type structure is particularly suitable for use in pressure sensors. There are several ways to make a bellows. It can be simply manufactured through the injection molding process. This has the advantage that the bellows can be manufactured together with the casing in one step. However, since the membrane must be very thin, the manufacturing process is complex. An alternative is to create only the internal headspace of the bellows in a manufacturing process such as an injection molding or even a turning or milling process. The material should be easily soluble, such as polyvinyl alcohol or polystyrene. Some metals such as aluminum, iron or copper are also suitable. A bellows structure is applied to this material and the internal structure is removed with a suitable solvent and/or by the application of heat. Many deposition processes are suitable for creating a bellows. For example, a noble metal (gold, palladium, etc.) can be deposited electrochemically. Metals, compounds and polymers can be thermally deposited in a vacuum. Sputtering processes are suitable, as well as chemical vapor deposition (CVD, from chemical vapor deposition). Many others, such as simple painting, can also work. Although a pure metal bellows will work, it is best to combine the metal with a polymer because it produces a less rigid bellows. It also appears to be effective to embed at least two or more very thin metal layers. So, for example, it is good to first deposit (spray) a layer of gold-palladium, then use the CVD process to deposit parylene C (poly-para-xylylene), and then sputter-deposit the gold alloy on top again. This allows the diffusion barrier to function even if several cracks appear in the metal layer, since the gas has to diffuse over a long distance in the parylene layer, which is already quite resistant to diffusion. Additional layers may or may not be provided on top to increase biocompatibility, e.g. each of the described embodiments may comprise one or more outer biocompatible layers. Instead of using an internal form, it is also possible to use external forms. They must crack to release the bellows, but can be reused several times. Physical deposition methods may not be well suited for this manufacturing process, but chemical deposition and electrochemical deposition, for example, are suitable. Other deposition processes mentioned above can be used after the (unfinished) bellows is removed from the mold.

As described above, there are many ways to coat a sensor. It is particularly useful to recoat the end sensor with metal, for example as described above with reference to FIG. 3A-3C. This makes all possible connections diffusion tight. Here again physical or chemical vapor deposition is useful. On top of this layer (or alternatively), if necessary, a biocompatible coating such as Parylene C can be applied. Otherwise, a noble metal or titanium coating already provides good biocompatibility.

As shown in FIG. 3B and 10C, to avoid the formation of blood clots on the rather sharp edges of the bellows, a smooth and soft top layer 818, 918 may be added. The void between the soft layer 818, 918 and the bellows may be filled with a fluid, such as water or silicone oil.

In fig. 3C shows a sensor 901 with a 3-element wire frame 920 for direct delivery into a vessel. The frame fixes the sensor without blocking the lumen of the vessel. The frame usually consists of a ring-shaped or disk-shaped central part from which curved legs protrude. It can be made from a wire material such as nitinol due to its high flexibility and good biocompatibility. Other materials such as stainless steel or polymers will also work. An alternative to wire is structures cut from sheet materials, which are then shaped into curves using a mold and heat treatment. Injection molding would also be possible, especially for polymers. For connection to the sensor housing, a ring or disk-shaped structure acts as an interface between the frame and the sensor housing. The ring-shaped structure may be slotted into the cylindrical casing and may be secured by spring force and/or by gluing or welding. The disc-shaped structure can be glued or welded to the sensor.

To avoid forces resulting from contact with the vessel wall, the frame 920 connects to only one part of the sensor 901 and protects the space around the other part (see Fig. 3C). It may be connected to either a part containing a stationary magnetic element or a part containing a rotating magnet. The frame structure may also contain helical structures (single-wire or multi-wire) or mesh structures. These structures can be optimized for compression, for example, during transvenous delivery through a fine needle.

In fig. 5A-5D show schematically and by way of example additional embodiments of a pressure sensor. Here, a symmetrical sensor design is presented to minimize coupling of torque to the environment. In fig. 5A and 5B show a symmetrical sensor 1101 at low (FIG. 5A) and high (FIG. 5B) pressure. The symmetrical sensor 1101 contains a cylindrical housing 1102, with bellows 1103, 1104 located at opposite ends of the housing 1102, i.e. the end surfaces of the housing 1102 are held by bellows 1103, 1104. Magnetic spheres 1107, 1108 are attached to the end surfaces by means of corresponding threads 1105, 1106, and the magnetic spheres, as in other embodiments, are permanent magnets. The outer surface of the housing 1102 is provided with a thin metal coating 1117 that acts as a diffusion barrier, i.e. the outer surface of the housing 1102 is provided with a diffusion blocking layer 1117. In FIG. 5C and 5D show an additional embodiment 1201 that corresponds to embodiment 1101 but also has a wire frame attachment 1220 to maintain distance from the vessel wall. For the final design, the open ends of the wires must be joined to avoid the individual wires becoming entangled in the vessel structures during delivery through the stream. Thus, also in this embodiment, a symmetrical sensor design is present to minimize the coupling of torque to the environment. In fig. 5C and 5D show a symmetrical sensor 1201 at low (FIG. 5C) and high (FIG. 5D) pressure. The symmetrical sensor 1201 contains a cylindrical housing 1202, with bellows 1203, 1204 located at opposite ends of the housing 1202, i.e. the end surfaces of the housing 1202 are held by bellows 1203, 1204. Magnetic spheres 1207, 1208 are attached to the end surfaces using suitable threads, the magnetic spheres being permanent magnets as in other embodiments. The outer surface of the housing 1202 is provided with a thin metal coating 1217 that acts as a diffusion barrier, i.e. the outer surface of the housing 1202 is provided with a diffusion blocking layer 1217.

The pressure sensors described above may be incorporated, for example, into guide wire 1310, as shown schematically and by way of example in FIG. 6 and 7. The ends of the guide wire 1310 may be welded to the housing 1302 of the pressure sensor 1301 with a stationary magnetic sphere 1307 and a rotating magnetic sphere 1308 attached to the membrane 1304 by a thread 1306. The housing 1302 includes at least one hole 1303 that can be viewed as vents to provide fluid connection to an area outside of housing 1302 to enable pressure measurement. The dimensions shown in Fig. 6 and 7 are given as an example only. Sizes may also vary. However, the dimensions shown are suitable for use with a fractional flow reserve pressure sensor. Applying the scaling rules to the observed signal-to-noise ratio of the demonstrator indicates that the specified dimensions will provide sufficient signal-to-noise ratio and accuracy for remote operation at distances large enough to completely penetrate the patient. In this way, a pressure sensor can be built into the guide wire, thereby making the wire a pressure sensor.

It may be useful to connect pressure sensors to other implantable devices (see FIG. 8), for example, to monitor the pressure drop across the stent 1401. This may be useful for characterizing pressure profiles in and around stents, for example, to detect blockages or to monitor progression diseases. In fig. 8 shows the integration of a pressure sensor 1403 into a stent with a realistic size scale (stent length = 30 mm, stent diameter = 4 mm, wire diameter = 0.2 mm, sensor length = 1.2 mm, sensor diameter = 0.5 mm), and in Fig. 8A shows two sensors 1403 at the inlet and outlet of the stent 1402 that can be used to monitor pressure drop across the stent 1402 and therefore possible blockage. In fig. 8B shows that the stationary portion of the sensor 1403 must be coupled to the wire frame 1402. A covering material can be added to give the sensor a more streamlined shape for better integration into a stent (not shown). In fig. 8C is a view of the interior of the stent 1402. The movable sensing portion can be tilted slightly toward the interior of the vessel to avoid or delay tissue proliferation.

Applications include coronary stents, aneurysm agents (pressure monitoring can help detect endoleaks), transjugular intrahepatic portosystemic shunts (TIPS), or stents used in peripheral vascular disease. As above, a ring or disk-shaped structure can act as an interface between the device and the sensor with all the mounting options mentioned above. Similar attachment can be applied to other internal devices, such as coil wires, shunt grafts, or Amplatzer transmural devices. For larger devices, such as wire guides, fractional flow reserve guidewires, catheters, large bypass grafts, or artificial heart valves, a hole can be drilled into the device to accommodate the sensor. Within the opening, again only one side of the sensor is attached, for example by means of an adhesive or clamp, while the other side is free to move, for example in a fluid or directly in the blood. The fluid may be either an immiscible type, such as silicone oil or perforated polyvinyl ethers, or it may be separated from the blood by an additional thin and flexible membrane, or both.

All clinical applications take advantage of the fact that the sensor is passive and small in size. It can be placed inside the human body, while the sensing system can detect it wirelessly without contact with the body and from a distance. For most clinical monitoring applications, the sensor must remain stable in the human body for several months or several years. However, for guidewires and catheters, stability should only be ensured for a few hours. For a transvenously inserted transducer, stability over several weeks may also be sufficient as new transducers may be supplied periodically.

In fig. 9 schematically and by way of example shows an embodiment of a heart valve 2000 integrated with a stent, with FIG. 9, the stent material is designated by reference numeral 2011. The heart valve 2000 includes a valve structure 2001 with a stationary portion 2002 and a movable portion 2004. The heart valve 2000 includes pressure sensors in accordance with the described embodiments. A first pressure sensor 2020 is located on the low pressure side of the stationary portion 2002 of the valve 2000. In addition, a second pressure sensor 2008 is located on the movable portion 2004 of the valve 2000. These two pressure sensors are attached to the outer walls of the valve 2000. However, the pressure sensors may also be built into the valve body. structure, in which case there is a space inside the valve structure, closed by a membrane and filled with a fluid, and a corresponding pressure sensor is located in this space. Through the membrane and fluid, external pressure causes a change in pressure at the position of the corresponding pressure sensor within the corresponding cavity. In fig. 9, a third pressure sensor 2007 is located in a cavity 2005 covered by a membrane 2003 on the low pressure side within a stationary portion of the valve 2000. A fourth pressure sensor 2010 is located in a space within a movable portion 2004 of the valve 2000, which space is also filled with fluid and covered by a membrane 2021. An additional pressure sensor 2009 may be placed in the cavity 2006 of the stationary part of the valve structure on the high pressure side, again in this case the cavity is filled with fluid and closed by a membrane 2014.

In fig. 10 shows schematically and by way of example an embodiment of a wire for treating a cerebral aneurysm. Wire 2100 contains pressure sensors in accordance with the described embodiments. In particular, the first pressure sensor 2104 may be located at the first end of the wire 2100 on one side of the first end. In addition, an additional pressure sensor 2101 may be attached to the second end of the wire 2100, and an additional pressure sensor 2103 may be mounted at an intermediate portion of the wire 2100, wherein the wire 2100 may include an internal cavity in which the pressure sensor 2103 is located, wherein the internal cavity has a fluid connection to the area outside the wire 2100 through holes 2102.

In fig. 11 schematically and by way of example shows an embodiment of a hepatic bypass device 2200 comprising a wire structure 2203. In this embodiment, the wire structure 2203 has a first portion 2201 surrounded by a lining material and an exposed second portion 2202. In this embodiment, the first portion 2201 is lined with using PTFE (polytetrafluoroethylene). Moreover, in this embodiment, the first wire structure portion 2201 has individual wires, whereas in the second wire structure portion 2202, the wires 2203 are intertwined. The hepatic shunt device 2200, which may also simply be called a hepatic shunt, contains multiple pressure sensors. For example, the first pressure sensor 2204 is located adjacent to the corresponding wire of the first portion 2201 of the wire structure 2203 within the PTFE tube. The second pressure sensor 2205 is located "in wire" inside the PTFE tube, i.e. a pressure sensor 2205 is located between the two ends of the corresponding wires of the wire structure 2203. A third pressure sensor 2206 is located between two adjacent wires of the wire structure 2203 inside the PTFE tube and is also connected to these wires. The wires of the wire structure 2203 have a wave-like shape, with an additional pressure sensor 2207 located between the peaks or valleys of the corresponding waveform, wherein the pressure sensor, for example, can be connected to two adjacent peaks or valleys of the corresponding waveform.

In fig. 11 shows an additional pressure sensor 2208 adjacent to a wire structure 2203 within a PTFE tube. Also, the exposed portion 2202 of the wire structure 2203 may contain pressure sensors. For example, pressure sensor 2209 may be positioned between two adjacent woven wires and connected to those adjacent woven wires. An additional pressure sensor 2210 may be located adjacent to the wire, and a pressure sensor 2211 may be positioned between and coupled to two peaks or valleys in the corresponding wire waveform of the wire structure 2203.

It should be noted that in FIG. 8-11 the location of the pressure sensors is given as an example only, i.e. also, more or fewer pressure sensors may be located in the same or different positions on or within the corresponding device. It is also possible that the corresponding device contains only one pressure sensor. The one or more pressure sensors of the corresponding device are pressure sensors in accordance with at least one of the described embodiments.

In the following, it is assumed that the length of the sensor is always approximately twice the diameter. All sensors with a diameter of 0.3 mm or larger will allow real-time pressure monitoring (more than 10 readings per second) over a distance of more than 30 cm with a pressure error of less than 1 mbar and a pressure range of at least 400 mbar. These parameters allow you to measure blood pressure with clinically relevant accuracy.

The sensor may be built into, for example, a guide wire, as explained above with reference to FIG. 6 and 7. Typical guide wire diameters range from 0.33 to 1.0 mm, i.e. For thin wire with pressure sensors, the sensor diameter should be about 0.3mm or less. Thus, a sphere diameter of 0.25 mm would be possible, leading to the above estimates of frequency, signal-to-noise ratio, and quality factor. A theoretically achievable signal-to-noise ratio of ~1000 at 30 cm would be sufficient for all sensing situations. With larger wire diameters, larger sphere sizes can be used, reducing the need for optimal background suppression. Thus, sensor diameters between 0.3 and 1.0 mm are useful for integration into guide wires.

The sensor can also be built into the catheter. Here, when placing the sensor in the catheter lumen, the same argument applies as for the guidewire. Sensors may need to be embedded in the catheter wall material, resulting in tighter size restrictions. It may be feasible to create a sensor with a sphere diameter of 0.1 mm, but the effort to remove the background signal will be increased and/or the distance over which the sensor can be reliably read will be reduced. Alternatively, heart rate-synchronized averaging can be used to improve signal-to-noise ratio, but at the expense of time resolution. Thus, probe diameters between 0.1 and 1.0 mm are useful for catheter integration.

The sensor can also be placed on a stent. To minimize the effect on blood flow through the stent, the diameter of the probe should not be much larger than the diameter of the wire. Typical stent wire diameter is 0.2 to 0.5 mm. So this would be a useful range for sensor diameters. However, a larger sensor can also be integrated, optionally with an additional streamlined cover.

It is also possible to insert the sensor using a syringe, and the sensor can be inserted into a smaller vessel in the lung or liver area without risk to the patient. Typical insertion probe diameters range from 0.3 to 1.0 mm. The size of the frame must be adapted to the diameter of the vessel in which the sensor is to be optimally installed. Preferably, the diameter of the frame should be greater than 1 mm, since in smaller vessels the pressure may deviate from the required pressure present in the larger supply vessel. To facilitate delivery through the needle into the venous system, the frame must be radially compressible to the diameter of the transducer housing.

Because the pressure sensor contains a magnetic object, like a permanent magnet, there may be a problem when scanning the body with an MRI scanner. The problem may not pose a danger to the body, i.e. for the patient, since the pressure sensor is small and therefore causes only small forces and torques that do not pose a threat to patients. Similarly, the magnetic resonance image produced by an MRI scanner cannot be damaged either because the pressure sensor is very small in size. However, many clinical MRI scanners use high field strengths greater than 1.5 Tesla, and the strong magnetic field can destroy the pressure sensor by changing the magnetization of the magnetic object or damaging the mechanical structure in the pressure sensor. This will be described in more detail with reference to FIG. 12 and 13.

In fig. 12 schematically and by way of example shows a pressure sensor 1 without protection from the magnetic resonance imaging field. The pressure sensor 1 has two magnetic spheres 7, 8, suspended by means of a corresponding thread 5, 6, which is attached to the body 2 at the corresponding attachment points 3, 4. When excited by an oscillating external magnetic field, spheres 7, 8 begin a counter rotational oscillation around the axis of the thread. This resonant vibration creates a field that can be detected at a distance. The housing 2 is partially flexible so that the distance between the two spheres 7, 8 and hence the resonant frequency changes depending on the external pressure. The flexible part of the housing 2 is not highlighted in FIG. 12 for clarity reasons.

In fig. 13 pressure sensor 1 from Fig. 12 is placed in a strong magnetic field in the axial direction 9. This causes the spheres 7, 8 to be oriented in the direction of the magnetic field. However, in this implementation, the threads 5, 6 are too short to fully align the spheres 7, 8. If the sensor body 2 cannot move, either the threads 5, 6 break or, given the very strong magnetic field of high-field MRI scanners, the spheres 7, 8 change direction of magnetization, making device 1 inoperable.

One solution to this problem is to house the pressure sensor housing in an outer housing 10, 110, which is preferably a spherical or ellipsoidal housing, for example filled with a highly viscous fluid, so that the entire housing and spheres can be reoriented to align the magnetization of the spheres with the external field and , thus avoiding re-magnetization. The spherical outer housing 110 allows for arbitrary reorientation of the sensor and thus can also be used for a simpler sensor design with one stationary sphere 107 and another sphere 108 oscillating on a string 106, as schematically shown in FIG. 14. In FIG. 14, pressure sensor 111 is formed using a simpler magnetomechanical oscillator 101, which includes a housing 102 containing a stationary sphere 107 and another sphere 108 oscillating on a string 106, inside a spherical outer casing 110. For the design shown in FIG. 13, depending on the length of the threads 5, 6, a partial reorientation of the sensor 1 would be sufficient, and the outer casing 10 could be more ellipsoidal, i.e. could be made smaller in one or two directions, as schematically shown in FIG. 15, in which the pressure sensor is designated by reference numeral 11.

The pressure sensor is designed in such a way that changes in external pressure outside the female casing are converted into changes in external pressure occurring outside the housing and inside the female casing. For example, the outer casing may be a very soft casing filled with fluid, or a casing with openings to effectively convert changes in external pressure into changes in mechanical vibrations of the magnetic object.

The additional casing 10, 110 makes it possible to reorient the housing 2, 102 to align the magnetization of the sphere with the external field, while in FIG. 14 shows a spherical housing 10 allowing free reorientation of the sensor and therefore also suitable for designs with, for example, a fixed magnetic sphere, while FIG. 15, the required tilt is possible in a housing 110 that is flattened in one or two directions, i.e. has a smaller diameter than sphere 10 in FIG. 14. The elliptical body 110 is especially suitable for sensors with two spheres 7, 8 suspended on threads 5, 6.

In fig. 16, the pressure sensor 201 with threads 205, 206 is elongated compared to the pressure sensor 1 shown in FIG. 12 and 13, while the diameter of the tube remains constant. Thus, there is enough space for the spheres 207, 208 to align with an external magnetic field of any direction. The minimum string length is Pi/4 the diameter of the sphere 207, 208. The only problem that might arise is that the field was changed to wrap the string 205, 206 around the sphere(s) 207, 208. To make this unlikely, the spheres 207, 208 and the interior of housing 202 may be coated with a slippery and non-stick material such as graphite. The threads 205, 206 are attached to the body 202 at attachment points 203, 204.

In fig. 17, which shows schematically and by way of example another embodiment, the filaments 305, 306 of the pressure sensor 301 are themselves too short to make the device resistant to magnetic resonance imaging. However, attached to the corresponding thread 305, 306 and to the body 302 is a corresponding winding unit 314, 315. If the forces acting on the corresponding thread 305, 306 become too large, this block 314, 315 releases a larger length of thread 305, 306. Therefore, the corresponding sphere 307, 308 can rotate freely, and the problem is solved. For normal operation outside the field of the MRI machine, it is necessary to accurately determine the length of the filament 305, 306. This is done using a stopper 311, 312, which can be attached to the filament 305, 306, or using some kind of stopper in the winding unit.

In fig. 18 shows one possible winder 414. It contains a spring material 422 that holds the thread 405 through a pulley 427 and an attachment point 426. When the force on the thread 405 is low, the spring 422 is stopped by some of the stops 424, 425. The spring 422 presses against the stops 424, 425, and therefore the length of the thread 405 is fixed. If the force becomes greater, the spring material 422 flexes and the available length within the housing 420 increases. This design preferably allows the thread to be extended up to 1.5 sphere radii. This means that it is sufficient regardless of the length of the filament inside the housing 420. The filament 405 is directed from the housing 420 into the body of the measuring (sensing) device through the opening 421 of the housing. The spring is attached to the housing 420 at a spring attachment point 423.

The described pressure sensors are preferably configured to compensate for the dependence of the resonant frequency on temperature. This configuration of pressure sensors for compensating for temperature shifts in resonant frequency will be described below with reference to FIGS. 19.

Also in FIG. 19, pressure sensor 3001 includes a housing 3002 and a magnetic object 3004 located within the housing 3002 such that it can rotate beyond its equilibrium orientation if the magnetic object 3004 is subjected to an external magnetic moment. The pressure sensor 3001 further includes a torque recovery unit 3003 configured to provide a recovery torque to return the magnetic object 3004 to an equilibrium orientation if an external magnetic force has rotated the magnetic object 3004 out of the equilibrium orientation to allow rotational oscillation of the magnetic object 3004 excited external magnetic moment. In this embodiment, the housing 3002 is cylindrical, and the magnetic object 3004 can rotate about an imaginary axis of rotation through the center of the magnetic object 3004, wherein the magnetic object 3004 is rotationally symmetric about the imaginary axis of rotation. Specifically, in this embodiment, the magnetic object 3004 is a magnetic sphere.

The torque recovery unit 3003 includes an additional magnetic object 3003 for providing recovery torque. Specifically, a magnetic object 3004 is attached to one end of the thread 3007, while the other end of the thread 3007 is attached to the housing 3002. The thread 3007 is configured to prevent the magnetic object 3004 from contacting the additional magnetic object 3003 due to their magnetic attraction and allowing rotational vibration magnetic object 3004. In this embodiment, an additional magnetic object 3003 is permanently attached to the housing 3002 using adhesive 3009.

The magnetic object 3004 forms a first magnetic dipole, the additional magnetic object 3003 forms a second magnetic dipole, and the magnetic object 3004 and the additional magnetic 3003 are arranged such that in the equilibrium orientation the first and second dipoles point in opposite directions. The first magnetic object 3004 and the second magnetic object 3003 are permanent magnets, with the north pole of the magnetic object 3004 facing the south pole of the additional magnetic object 3003 in equilibrium orientation, and vice versa.

The housing 3002 is cylindrical, wherein the cylindrical body 3002 includes two end surfaces 3030, 3031, and wherein an additional magnetic object 3003 is fixedly attached to the first end surface 3030, and the end of the thread 3007, which is opposite to the end attached to the magnetic object 3004, is attached to the second end surface 3031 of the cylindrical body 3002.

In this embodiment, the second end surface 3031 of the housing 3002 is formed by a flexible portion 3008 of the wall of the housing 3002, wherein a magnetic object 3004 is attached to the flexible portion 3008 by a thread 3007 such that external pressure acting on the flexible portion 3008 from outside the housing 3002 causes a change distance between the magnetic object 3004 and the additional magnetic object 3003. Due to this change in distance caused by external pressure, the strength of the magnetic field generated by the additional magnetic object 3003 at the position of the magnetic object 3004 changes, and therefore the resonant frequency changes. Thus, the resonant frequency changes depending on the external pressure so that the pressure sensor 3001 can be used to measure the external pressure as another physical quantity. Thus, the flexible wall portion 3008 of the housing 3002 can be considered as a measuring element that is configured to change the resonant frequency depending on the external pressure.

The pressure sensor 3001 further includes magnetic material 3005, 3006 located adjacent to the additional magnetic object 3003. This magnetic material 3005, 3006 influences the magnetic field generated by the additional magnetic object 3003, wherein the influence of the magnetic material 3005, 3006 depends on temperature to change the strength magnetic field at the position of the magnetic object 3004 and therefore to change the resonant frequency if the temperature changes. The magnetic material 3005, 3006 is designed such that its magnetization decreases with increasing temperature. Moreover, the magnetic material 3006 is configured such that its magnetization direction is opposite to the magnetization direction of the additional magnetic object 3003, and the magnetic material 3005 is configured such that its magnetization direction and the magnetization direction of the additional magnetic object 3003 are the same. The magnetic materials 3005, 3006, which are soft magnetic materials, therefore affect the resonant frequency as a function of temperature in opposite frequency directions, i.e. one of these magnetic materials leads to a change towards higher frequencies as the temperature increases, and the other of these magnetic materials leads to a change towards lower frequencies as the temperature increases.

The pressure sensor is preferably designed in such a way that the resonant frequency is independent of temperature. However, for example, the flexible housing wall portion 3008, which may be formed by the membrane, may have temperature-dependent flexibility so that the resonant frequency may also generally be temperature dependent. Also, other parts of the pressure sensor may be affected by temperature, and this dependence may also affect the resonant frequency. To compensate for this unwanted temperature-dependent frequency shift, the magnetic materials 3005, 3006 can be designed to provide the same frequency shift in the opposite frequency direction depending on the temperature change. In particular, the magnetic materials 3005, 3006 may be selected and positioned to eliminate any temperature dependence of the resonant frequency of the pressure sensor 3001. It is also possible that only one of the magnetic materials, i.e. only a magnetic material that decreases the resonant frequency with increasing temperature, or only a material that increases the resonant frequency with increasing temperature, is used to reduce or even eliminate the temperature dependence of the resonant frequency of the pressure sensor 3001. One or both of the magnetic materials 3005, 3006 can be considered as compensation elements to compensate for temperature-induced resonant frequency shift.

Fig. 20 schematically and exemplarily illustrates a detection system 1501 for detecting a resonant frequency of a corresponding sensor, for reading the corresponding sensor, i.e. reading system for wireless reading of the corresponding pressure sensor. In fig. 21 shows a prototype detection system 1501 as an example. The detection system 1501 primarily comprises at least one magnetic field generator and at least one magnetic field sensor, i.e. a transducer for converting the magnetic or electromagnetic field generated by the induced vibrations of the magnetic object of the pressure sensor into electrical response signals. The operating frequency band is in the low kHz range and must be wide enough to cover the responses of multiple sensors operating in parallel at different frequencies, and perhaps also the higher harmonics of the sensor's resonant frequency, for example to improve the signal-to-noise ratio. Transmission field amplitudes are a few milliteslas at most, while field amplitudes to be detected range from 1/10 nT to several nT. Many different field generators can be used (oscillating permanent magnets, coils with/without cores, magnetostrictive field modulators, ...), as well as many different magnetometers (based on the Hall effect, magnetoresistive sensors of various types, magnetic resonance sensors, superconducting magnetometers (SQUID, from English superconducting quantum interference device), etc.). Technically, the simplest systems are coreless conductive loops for transmitting and receiving magnetic fields. The coils are generally good enough for sensor applications. The magnetic field generating coil can also be used to receive the magnetic field. However, different coils can be used for these tasks, which provides some advantage. In fig. 20 and 21 show a detection system that is a single-channel transceiver system, where multiple channels can operate in parallel to obtain spatial information.

In fig. 20, detection system 1501 includes a transmit coil 1503 that is coupled to a microcontroller 1507 via a digital-to-analog converter (DAC) 1506 and an audio amplifier 1502 to generate an external magnetic torque for a pressure sensor 1520, which may be any of the described pressure sensors. The receive coil 1504 is also connected to the microcontroller 1507 through a low noise amplifier 1505 and an analog-to-digital converter (ADC) 1508 to sense the resonant frequency. The microcontroller 1507 is coupled to a display computer 1509. The microcontroller 1507 is configured, for example, to generate and receive signals, estimate and control frequency, and optionally measure a reference pressure. In fig. 21 also shows a transmit/receive decoupler.

The microcontroller 1507 generates transmit pulses (see top line 1350 in FIG. 22), which are amplified by an audio amplifier 1502 and then transmitted to a transmit coil 1503, which may also be called a drive coil. This implementation uses a separate receive coil 1504 that is decoupled from the transmit coil 1503 via two additional decoupler coils 1510, which are not shown in FIG. 20 for reasons of clarity. The received signal is fed into a low-noise amplifier 1505 and sent to the ADC 1508 of the microcontroller 1507, where the time curve, typically 1/20 of a second, is sampled at a rate of about 20 ksa/s. In fig. 22 shows, in addition to the transmit pulses 1350, which may also be called excitation pulses, the induced voltage 1351 in the receive coil 1504 due to the oscillation of the sphere in the sensor and, therefore, due to the response of the sensor. The interval between excitation pulses 1350 may be continuously adjusted by microcontroller 1507.

The advantages of a multi-coil system will be discussed below. In a single coil system, the relative orientation of the sensor relative to the coil may be such that the coil cannot control the oscillations of the magnetic sphere, nor can it sense the change in the generated field. Thus, to avoid the user having to reorient the sensing system relative to the sensor, a multi-coil system may be desirable. The coils must have different spatial sensitivity distributions to be able to generate the optimal excitation field vector in all situations. The use of multiple coils, in addition, allows the sensor to be localized by determining the position and orientation of the oscillating magnetic dipole in space. The different amplitudes of the received signals, together with the known sensitivities of the coil elements, can be matched to a dipole model to determine position and orientation parameters. An example of a multiple coil system for implementation in a pillow or mattress is shown in FIG. 23. With multiple receive coils and channels available, additional information can also be used to improve background suppression, as described below.

In fig. 23, multiple coils 1652 form a multi-coil array that is built into the patient bed mattress 1651 of an imaging system 1650 similar to a C-arm system. 1652 coils are predominantly aluminum coils with x-ray absorption of less than 10 percent. Therefore, there is no need to increase the patient dose if 1652 coils are used.

Below, the transmission system based on the detection system coil will be described in more detail by way of example. The coil-based transmission system includes a transmission amplifier and a transmission coil. Optionally there is also a matching circuit and a “jamming” circuit. Since the shape of the transmitted signal is not very critical in the sensor application, a variety of amplifiers (Class A, Class B, Class AB, Class D, etc., using transistors, vacuum tubes, thyristors and many other components) are suitable for this task. Since signal quality is not critical, the lowest-loss amplifier topology can be selected, which is a half-bridge or full-bridge amplifier using low-resistance switches. The preferred switches are MOSFET or IGBT type transistors. The matching circuit in the simplest case is a simple capacitor connected in series to an inductor. Provided that the amplifier is operated with sufficient supply voltage, such a matching capacitor may be omitted or the capacitance may be chosen so high that the resonant frequency of the capacitor coil is well below the operating frequency. The matching circuit is of interest for another reason. Medical equipment must always operate in a safe manner, so voltage reduction is of interest. By placing the capacitor in the middle of the coil so that current flows through one coil section, then through the matching capacitor, and then into the second coil section, voltage peaks can be reduced. This is even more true if the coil is divided into several sections, each of which is connected to a corresponding capacitor. This turns the coil and matching network into an integrated unit. It is convenient to control the field amplitude using pulse-width modulation, i.e. the amplifier increases/decreases the current through the coil for only a fraction of a cycle, or quickly switches between increasing/decreasing current. Since the exact waveform is less important to the sensor application, it is best achieved by changing state only 2 times during a half-wave (or 1 time in the case of full power, where the pulse length is identical to the half-wave length). Ideally, the amplifier should be able not only to increase or decrease the current, but also to maintain it more or less constant or at the level required by the matching network. This is achieved by the correct switching sequence of transistors in a half-bridge or full-bridge circuit. Typically, the amplifier supply voltage should be fairly low and in the range below 50 V. Additionally, the matching network should be configured so as not to exceed this 50 V limit at any two points. In both cases it is even better not to exceed 24 V. This means that the number of turns should be kept to a minimum. However, peak operating currents should exceed 10 A, preferably 100 A.

Transmission/reception isolation will be described below. It is important that there is not too much noise from the transmission system, i.e. from the field generator, entered the receiving system, i.e. into a transducer for converting the magnetic or electromagnetic field generated by the induced mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals while the transmitting system is not in transmitting mode, i.e. no excitation field is generated. In addition, the transmission amplifier should not short-circuit the received signal or even partially attenuate it. There are several possibilities to achieve this goal. If we have different transmitting and receiving coils, then these two coils can be decoupled geometrically (see Fig. 24).

In fig. Figure 24 shows the implementation of the design of a gradiometric receiving coil for suppressing the transmitted and background signal in the receiving path. Here, large 1452 coils were chosen, which allowed the sensor to be read at a distance of about 30 cm above the top coil. The gradiometric design uses a geometric decoupling technique: transmit coil loops 1451 are coupled to produce parallel fields, while receiver coil loops 1450 are coupled to receive field gradients and suppress uniform fields. This transmit/receive system provides internal geometric isolation through the use of parallel transmit loops and antiparallel receive loops, which can be called a gradiometric configuration. This leads to an internal geometric connection. This coreless coil system is highly linear. In fig. 24 also shows a DC blocker 1455 with an audio amplifier 1454 and a low pass filter 1453. Bottom part of Fig. 24 illustrates the outer turns of the receive coil 1450 and the inner turns of the transmit coil 1451.

In particular, in FIG. 24, the bottom left image is a close-up of the middle portion of the top coil assembly. In the bottom left image you can see that there is actually only 1 turn of the 1451 transmit coil peeking out from the bottom. The rest is hidden by a receiving coil wound with much thinner wire. The DC blocking circuit 1455 is simply used to precondition the signal before the audio amplifier, since the signal for the audio amplifier can be generated by a simple PWM output. The low pass filter 1553 is a filter between the output of the audio amplifier 1454 and the transmit coil 1451. It has two purposes. Firstly, it avoids the introduction of high-frequency noise, and secondly, it allows you to combine two output channels of an audio amplifier into one.

Geometric isolation is not always possible, especially if an array of transmitters and receivers is used. In this case, a transformer can be introduced with terminals connected to both the transmitting and receiving circuits. This transformer provides decoupling of the transmitting and receiving system. This transformer solution can be used even if a combined transmit/receive coil is used. The transformer can be replaced by a capacitive (or even resistive) isolation network with either combined or separate transmit/receive coils. The disadvantage of compensation methods is that they require a fairly large space, increase noise and, in the case of capacitive decoupling, narrow the frequency range of the detection system. A more reliable and cheaper solution is to add a circuit that completely mutes the transmit amplifier during reception. To do this, antiparallel diodes can be added to the output of the amplifier. Diodes with low zero-voltage capacitance, such as PIN diodes, are especially useful. This provides high impedance in the absence of current. To further enhance this, an electronic switch can be installed at the output of the amplifier, short-circuiting any residual noise signals during reception. The diodes still provide the required high impedance. It is also possible to design a special amplifier that is completely silent and provides high impedance when not operating. With half bridge and full bridge designs this can be achieved by having absolutely no switching in any component during receive, using low output capacitance transistors, having approximately half the supply voltage at the output(s) in receive mode, having no noise from the input jacks (optical isolation) and the presence of a highly filtered supply voltage (high filtering or no switching of the power supply during reception).

The receiving system based on the detection coil is discussed below. The reception amplifier must be of a low noise type. However, the requirements are not so high that it requires the use of unusual receiving transistors. Standard low noise bipolar or silicon JFET transistors are usually good enough. The only special characteristics are that the amplifier must withstand the transmit pulses and begins to operate shortly after the transmit pulse. There are several ways to achieve this goal. In the case of decoupled transmit/receive systems (including a combined transmit/receive coil with an isolation network), the receive amplifier does not need special characteristics to achieve this goal. If there is no decoupling, the amplifier can be configured for a transmit pulse. This can be done by adding a suitable capacitor at the input of the amplifier and anti-parallel diodes at the second pin. This provides a suitably high impedance in case of transmission and shorts everything at high voltages to harmless levels for the amplifier. Naturally, the added capacitor must be rated for the maximum transmission voltage. The capacitance value should be so high that the signal at the amplifier does not decrease too much when receiving. For a JFET based amplifier this is usually not critical. The antiparallel diodes can be supplemented or replaced with a suitable electronic switch such as a MOSFET output optocoupler. This has the advantage of further reducing the input voltage. If done correctly, the receiving amplifier will not be saturated and will function as soon as the transmitted signal is sufficiently attenuated.

The interface to the digital system is discussed in more detail below, first describing the output and processing of the digital signal. Although an analog timer system can generate output signals, a digital system such as a digital signal processor (DSP) or field programmable gate array (FPGA) is typically used. Depending on the type of output amplifier, different outputs can be used. For an analog amplifier, some type of ADC may be used. Since the quality of the output signal is not very critical, a simple PWM-type analog output may be sufficient. Digital amplifiers are best connected using a digital output line. However, you can also use the analog output for them and implement a switching circuit generator on the amplifier. With the best matching amplifier, half bridge or full bridge, it is most appropriate to create the switching circuit directly in the digital system. In addition, switching circuits for protecting the input of the receiving amplifier and suppressing noise at the output of the transmitting amplifier can also be generated directly by the digital system. A common feature of all output options is that they must be fast enough to accurately maintain phases across different excitations of the same sensor or between different sensors. Thus, the output should be able to switch updates on the raster more often than 1/10 of the full time period, better - more often than 1/100 of the full time period. For a sensor at, say, 2 kHz, this means refreshing the raster at more than 220 kHz, even better than 200 kHz. This does not mean that switching state changes must be possible at raster points every time. So, for example, it is possible to have a serial interface for each amplifier that communicates a new switching state to the amplifier, and a protocol for making that change at a specific time over the same serial interface. This is especially useful for this type of amplifier, which by default does not function during the receive phase. A 1-bit serial interface can be implemented for this, requiring only one optocoupler on the amplifier. This makes it easy to achieve noise immunity on the digital side, since the parasitic capacitance in a single optocoupler can be very low.

The analog-to-digital interface will be discussed below. Analog to digital conversion is fairly standard. Since the signal has low bandwidth, at least if only one sensor is used, it would be possible to mix the signal to near DC current and sample that signal. However, the sensor signal has a fairly low frequency, typically below 10 kHz. Today, there are many suitable ADC chips available for direct sampling. Moreover, digital signal processing is very poor compared to analog filters, it is best to use highly oversampled ADCs. At least 10 times the sensor frequency should be used, but 100 or 1000 times are also valid options. Highly oversampling simplifies and reduces the cost of the ADC input filter design, since only the sensor signal frequency must pass, while the signal must not pass above the Nyquist frequency. However, filtering below the sensor frequency is also useful to avoid the usual high background signals there. A high background signal can reduce the available gain in front of the ADC, increasing the noise contribution of the ADC. The ADC noise (effective bits) and samples must meet the required dynamic range and noise expectations. This means that the ADC should not saturate when the maximum expected signal and all noise components are present. At the same time, the ADC sampling noise must be so low that the overall noise does not increase. Here, noise refers to all the unwanted components of the recorded signal that arise from real noise sources, such as coil resistance or the behavior of the receiving amplifier. It also includes interference components that cannot be eliminated by suitable filtering and background subtraction. Typically, with modern ADC chips, this requirement can be met, for example, by using 18-bit ADCs at 2 MS/s. To save money, it may be useful to use a lower specification ADC but add a gain control to still achieve good overall performance.

Data processing will be discussed below. Before data evaluation, the raw ADC data must be processed. Since highly oversampling is desired, the first processing step may be a thinning step. The main advantage of this is to reduce the size of the data and therefore the required computing power for further steps. Optionally, the thinning step may include other filters, e.g. bandpass filters around the expected signal frequency. This can simplify further processing steps and reduce the dynamic range of the signal, which in turn can save processing power (variables with fewer bits). Another optional data processing step is the use of an inverse nonlinear filter to reduce the nonlinearity of the receiving system. This means that the nonlinearity of the complete receiving system is measured and a computational filter is created to remove the effect of the nonlinearity. This is especially useful if low-cost components are used, as they tend to have more non-linear behavior. This non-linear filter can alternatively be used in the first stage of processing. If more than one received signal is used, then additional signal processing steps are performed. If at least one receive channel does not detect a sensor signal and thus provides a measurement of the background signal, this (and all other such signals) are correlated with the receive signal and the correlated components are subtracted from the channels carrying the signal. This subtraction can be done in the time domain or the frequency domain, or a combination of both. If there are no channels without a sensor signal, a data processing strategy sometimes called a "virtual gradiometer" can be used. It decomposes multiple channels into virtual channels, which are linear combinations of physical channels, to minimize interference from signals that are not generated by the sensor. The coefficients for linear combinations can be found by correlating the channel signals, excluding the sensor signal band.

In addition, the implementation of data evaluation will be explained below. Since changing the pressure across the sensor changes the distance between the magnetic spheres and therefore the resonant frequency of the magnetomechanical oscillator, frequency is the main parameter that needs to be extracted from the received sensor signal. Due to the high quality factor of the resonator (time constant up to seconds), subsequent excitation pulses are usually reproduced until the oscillations are completely damped (see Fig. 22) and, therefore, it is necessary to have the correct phase and timing to amplify the existing oscillation. This requires real-time frequency extraction between successive excitations. Frequency can be extracted either by a comparison algorithm that minimizes phase differences between the measured signal and precomputed time curves covering a range of frequencies, or by Fourier analysis, which is the preferred method. High-resolution frequency information can be obtained by zero padding the time domain or interpolating the frequency domain and then locating the resonant peak in the spectrum using either a peak finding technique or a curve fitting procedure. To further improve the accuracy and reliability of the frequency determination, the higher harmonics of the detected resonant signal (see spectrum in the upper right corner of Fig. 25) can be included in the estimation, for example, using a weighted frequency estimate based on several harmonics, or by checking the consistency of the frequency determination between several harmonics.

In the example to which FIG. 25, the second harmonic signal is an order of magnitude smaller than the base frequency signal. Therefore, better filtration is required. Various filtering stages can be used to optimize the signal at the resonant frequency and its higher harmonics, such as analog excitation filters such as DC blocker and low pass filter, analog receive filters such as bandpass filters, and digital receive filters , such as filters with infinite impulse response (IIR filters, from the English infinite impulse response) for real-time processing (Chebyshev filter of the second kind of the sixth order). In fig. 25, the central position of the resonant peak f0 is determined by the largest peak in the filtered spectrum. From f0 the instant of time of the next in-phase excitation pulses is calculated. The system's repetition rate ranges from 5 to 30 Hz, providing real-time frequency response tracking (see Figure 32).

In fig. Figure 25 shows the spectrum of the signal with and without digital bandpass filtering (1051 and 1050, respectively). The marked points are within the range selected for evaluation. Different dot symbols represent different types of filters, which don't really show a significant difference and can therefore be ignored. In fig. 26 bandpass filter 1054 is connected to a commercial low noise audio amplifier 1053, type DLPVA-100-BUN-S from FEMTO Messtechnik GmbH. In fig. 27 the actual rejection spectrum of a 40 dB digital filter is compared with the range of the selected band. The two implementations show no noticeable difference. The displayed filter is applied to the data shown in FIG. 25, which results in a difference between 1050 and 1051.

Based on the specific frequency and known time stamp of the received signals, the correct timing for the next block of excitation pulses can be calculated. The number and duration of excitation pulses are selected in such a way as to generate oscillations with a sufficiently high amplitude to obtain a sufficient signal in the receiving coils.

To calibrate a pressure sensor, it is necessary to obtain a frequency response for many well-defined pressures acting on the sensor. For this purpose, a high quality pressure sensor can be connected to a pressure chamber containing the sensor (see FIGS. 28 and 29). Based on one or more measurements over an appropriate pressure range (up to 400 mbar above ambient pressure to safely cover the blood pressure range), a frequency-pressure calibration curve can be determined, as shown in FIG. 30. Depending on the properties of the sensor, a simple fit to the calibration curve may be sufficient. However, a real sensor may exhibit hysteresis behavior, dynamic response of the membrane or other mechanical elements of the sensor, or temperature dependence. Thus, it may be necessary to fit a model based on the physical parameters of the sensor to the measured calibration data to obtain a highly accurate sensor calibration.

In fig. 28 schematically and by way of example shows a calibration apparatus for calibrating a pressure sensor, the calibration apparatus including a pressure applicator 1070, a pressure chamber 1071, a reference pressure sensor 1072, which may be a commercial pressure sensor, and a data recording and display unit 1073. The pressure applicator 1070 applies the specific pressure measured by the reference pressure sensor 1072 while the resonant frequency is measured such that during this calibration procedure the measured resonant frequencies can be assigned to actual pressure values. In fig. 29 is a schematic diagram of a data recording and display unit 1073 showing a measured reference pressure 991 and a measured sensor response frequency 990, i.e. in fig. 29 shows a real-time display of the detected sensor response frequency as a function of pressure measured using a commercial reference pressure sensor. In fig. 30, the bottom portion shows the corresponding calibration curve 890, and the top portion shows a comparison between the reference pressure 892 and the resonant frequency 891, which are in good agreement. The sensitivity of this example calibration curve is 0.17 Hz/mbar, which corresponds, if a frequency resolution of about 20 MHz is assumed, to a pressure resolution of about 0.1 mmHg.

The sensitivity of the sensor is determined by the magnitude of the change in frequency per change in pressure. This depends on several parameters such as the design of the sensor, the stiffness of the membrane, the size of the magnetic elements and the distance between the magnetic elements. For a given sensor design and size, modeling can determine the optimal distance between magnetic elements, as well as the required membrane properties for optimal deflection as a function of applied pressure. In fig. Figure 31 shows an example simulation predicting sensor sensitivity over the frequency range of interest for two sensor sizes. In this figure, curve 437 corresponds to the demonstration sample, and curve 438 corresponds to the target size described above for use in fractional flow reserve (FFR).

In addition to the relative sensitivity of the sensor as defined above, the absolute sensitivity or pressure resolution depends on the noise level when determining the frequency. In the example shown in FIG. 32, the noise level is about 0.2 Hz and limits the resolution of the sensor to about 1 mbar, with curve 441 showing the frequency response of the sensor and curve 442 showing the measured reference pressure. In fig. Figure 32 shows the noise level of the initial demonstrator, which limits the pressure resolution to approximately 1 mbar. This is already sufficient for most medical applications. Improved background noise removal strategies can further increase resolution.

A sensing system processor for wirelessly sensing a corresponding pressure sensor may be configured to apply a compensation algorithm to adjust the pressure value determination for a resonant frequency dependent on at least one of a) the distance between the pressure sensor and the field generator and b) the common-mode drive of the coil. This compensation will be explained in more detail below.

Presented here are methods for compensating for the effects of distance as well as orientation on frequency in a magnetomechanical resonator for measuring position and parameter, i.e. pressure. This compensation is only required when a physical parameter such as pressure is being measured and the information is encoded into the frequency of the oscillator. For the localization of the oscillator, which will be explained below, frequency effects are either irrelevant, i.e. either irrelevant (sensitivity coding is discussed below) or negligible (gradient field coding is also discussed below). For localization using the gradient field method, which also affects the sensor frequency, these compensations are not required, since only the change in frequency over a subsecond period of time needs to be estimated. This change does not depend much on the amplitude of the oscillations.

The signal of magnetomechanical oscillation generators is detected by the voltage u _i (t) induced in the i-th coil as a result of a change in the field due to the oscillatory motion of the magnetic moment m(t) of the suspended magnetic sphere in position r ₀ :

where B _S,i (r) is the sensitivity of the i-th detector coil at position r, which basically remains constant over time. At the last stage, the magnetic moment was replaced using

Where is a unit vector describing the spatial orientation of the magnetization, M _sat is the saturation magnetization of the material used (usually from 1.30 to 1.45 T/μ ₀ for NdFeB), and V _sphere is the volume of the magnetic object.

From (4) it follows that in order to induce a high voltage in the receiving coils, it is desirable to have a large dynamic magnetic moment. Since the volume of the spheres must be small in most applications, the signal can be increased by using a large amplitude of oscillation, resulting in a large . However, the restoring torque does not increase linearly with the angle ϕ (i.e. the amplitude of oscillation) between the restoring field B _rest created by the stationary sphere and the magnetization m of the oscillating sphere:

Considering the torque due to friction with damping coefficient C, and the torque required for the angular acceleration of a sphere with mass m _s and radius r _s , , we can create an equation of motion:

Approximation of the small angle sin ϕ≈ϕ and replacement m=M _sat V _sphere lead to

The high quality factor of the system makes it possible to further approximate C≈0 and allows one to calculate the angular resonant frequency as:

Since microoscillators are typically driven to amplitudes much greater than 10°, this approximation is generally incorrect. For larger angles, the restoring torque is smaller and hence the frequency decreases, resulting in an amplitude-dependent frequency ω(ϕ _max )=ω ₀ k(ϕ _max ), with k(ϕ _max )<1. A change in the restoring torque during oscillation also introduces nonlinearity into the sensor response, which is manifested in the presence of higher harmonics of the base frequency in the spectrum.

In addition to the nonlinear restoring torque, the force between two magnetic spheres depends on the relative orientation of their magnetization:

For this sensor design, the force is always directed along the vector of connection of the two magnetic spheres, but its magnitude tends to zero at an oscillation amplitude of 90° and even changes from attraction to repulsion at higher angles. If the thread on which the sphere is suspended elongates due to the applied force, then the decrease in the average force at high vibration amplitudes increases the distance between the spheres, reducing B _rest and, therefore, the frequency of vibration. Not only can the filament change its length, but potentially other structures in the sensor as well.

If the excitation fields created by the transmitting coils had a constant amplitude, the amplitude ϕ _max of oscillations would decrease with increasing distance between the coil and the sensor (decreasing the excitation field) and, therefore, the frequency would decrease. The amplitude also depends on the relative orientation between the coils and the sensor, as shown in FIG. 33.

In fig. Figure 33 shows the measured dependence of the signal amplitude in various harmonics on the orientation of the sensor relative to one transceiver coil. If the excitation field is aligned parallel to the orientation of the magnetic dipole, no excitation occurs and the signal is zero. With orthogonal alignment of the field and dipole, the largest amplitude of oscillations is achieved. It should be noted that the spatial structure of the even harmonics is aligned orthogonally relative to the odd harmonics. This can be seen by the zero amplitude of the 2nd harmonic at an orientation corresponding to the maximum in the base signal (1st harmonic) and the 3rd harmonic. The amplitude ratio plot (center plot) highlights this difference in orientation relationships: the ratio of the 2nd harmonic to the 1st harmonic varies from zero to a maximum value (or singularity), while the ratio of the 3rd harmonic to the 1st harmonic is permanent. Knowing that the dynamic response at even harmonics is oriented orthogonal to the dynamic response at odd harmonics can be used to determine the 3rd orientation angle of the sensor.

Thus, a naive determination of frequency will result in readings that depend not only on the physical quantity, but also on the position and orientation of the sensor. This is generally undesirable and mechanisms are therefore preferably used to reduce this effect.

There are two strategies to mitigate this effect. One is to ensure that the amplitude of the oscillation remains constant for all valid positions and orientations (i.e., in the field of view). Another is to calculate the virtual frequency from one or more frequency readings, where the virtual reading is independent of the amplitude of the oscillations. The selected virtual frequency may be a resonant frequency with very low amplitude, i.e. frequency with zero amplitude as given in equation (8). Naturally, a combination of these two strategies can be used.

Below will be described the control of the amplitude of oscillations, which uses the position and orientation of the pressure sensor, known from localization.

Conceptually, the simplest way to correct for frequency offset is to use a known position and orientation relative to the transmit/receive coil array. This position can be obtained using sensitivity encoding or gradient field encoding or a combination thereof, described below. Once the position and orientation are known, the absolute dynamic dipole moment can be determined. Usually the dynamic dipole moment is already a suitable parameter when determining the position. Otherwise, it is derived from the known sensitivity of the coils given the relative position and orientation of the sensor and the recorded signal level, as given in equation (3). From the dynamic dipole moment and the known static dipole moment of the sensor, the oscillation amplitude can be deduced. (If the static dipole moment of the sensor is previously unknown, it can be inferred from several dynamic dipole moments measured at different excitation amplitudes as the achieved or extrapolated maximum possible dynamic dipole moment.) The amplitude of the transmission field can now be adjusted to achieve a predefined maximum oscillation angle _ϕmax . A slightly simpler, but also slightly less reliable method is to simply adjust the transmitted amplitude to such a value that the excitation field component at the sensor is always constant. Typically, the excitation process is dominated by one directional field component (amplitude in one direction relative to the sensor). The amplitude depends on the different simultaneous excitation currents in several excitation coils. The amplitude and phase of the current in the coils are adjusted to provide the required amplitude at the required position and orientation. How this is done is well known in the field of electromagnetism. Since the solution to this problem and the previous method for adjusting the transmitted amplitudes are ambiguous, at least one second criterion is used for the current optimization. "Ambiguous" in this context means that many combinations of amplitudes and phases in the transmitting coil currents have the same desired result. The second criterion is usually to minimize the unwanted side effect of arousal. Examples of unwanted side effects are total power dissipation, maximum power dissipation in a single coil, maximum temperature in a single coil (which also depends on the drive history), heat or forces in surrounding objects. Naturally, any other related criterion or combinations can be used for this optimization. It is also acceptable to use not the optimal solution, but simply the first feasible solution found by a specific optimization algorithm (for example, gradient descent or simulated annealing, etc.).

Zero amplitude frequency detection using the position and orientation of the pressure sensor known from the localization procedure will be described below.

The second approach, in which the amplitude is not adjusted but the frequency is converted, for example, to a frequency with zero amplitude, can also be described for the case of a known position relative to the coil array. This approach requires a sensor model. This model describes the relationship between frequency and angle of oscillation directly or in some equivalent form, such as frequency and the strength of a dynamic dipole moment. This model can be in the form that the frequency is expressed as the frequency with zero amplitude minus a function depending on the amplitude of the oscillations. More complex descriptions can also work, such as multiplication by a correction term similar to k(ϕ _max ) introduced after equation (6) above. Both frequency and amplitude can be interpreted as an average or fitted curve over a predefined integration period or predefined sequence length. Since the position and sensitivity of the coil are known, the amplitude of the oscillations can be determined from the recorded signals as described above. Thus, since the actual frequency and amplitude are known, the frequency with zero amplitude can be derived by inverting the model function. The inversion can be performed analytically or using well-known numerical methods. Mathematically, instead of a frequency with zero amplitude, you can also use any other related quantity, for example, a frequency with an amplitude of 10°, as well as quantities that do not reflect frequency at all, for example, the absolute distance between spheres. However, such a transformation does not change the fundamental nature of this assessment process.

Generally, even when using the zero amplitude extrapolation method, some kind of amplitude adjustment is preferable for most applications. If the sensor is close to the coil array, less current can be used through the coils than when the sensor is further away from the array. The zero amplitude extrapolation method has the advantage that it can operate even if the sensor is so far away that the preset oscillation amplitude for the constant amplitude method can no longer be maintained due to coil current limitations.

In the above descriptions it was assumed that the amplitude of the oscillations could be inferred based on the known position and orientation of the sensor and the induced voltage in the receiving coils. However, any other method that can provide amplitude information can be used to compensate for amplitude-dependent frequency using either constant excitation or extrapolation methods. Some alternative ways to determine this amplitude will be presented below. These methods can be useful when only a small number of coils are available and the exact relative position cannot be determined. The compensation methods themselves will not be repeated in the text below.

Below we will describe how the amplitude of oscillations can be determined using the amplitudes of the harmonics of the base frequency.

One way to determine the amplitude of oscillations is to evaluate the harmonics of the induced signal in the coils. As nonlinear oscillators, magnetomechanical sensors generate harmonics at a resonant frequency in a dynamic dipole moment. These harmonics are collected in the pickup coil(s). Preferably, care should be taken not to suppress these multiples of the base frequency during the sampling and filtering stage. The harmonic spectrum depends on the characteristics of the sensors. There may be sensors that predominantly generate odd harmonics (at 3ω ₀ , 5ω ₀ , ...), and there may be sensors that generate even and odd harmonics (at 2ω ₀ , 3ω ₀ , 4ω ₀ , ...). However, mixed types can be created. The dynamic dipole moment of odd harmonics tends to align with the dynamic dipole moment of the fundamental frequency, while the even harmonics tend to align perpendicular to the dynamic dipole moment of the fundamental frequency and perpendicular to the rotation axis. Therefore, odd harmonics are conceptually the easiest to use because the ratio of, say, the dynamic dipole moment of the third harmonic to the base frequency dipole moment is reflected as a corresponding ratio in the recorded voltages in one coil, for example estimated as the amplitudes of the spectral peaks. However, since the gain in the receiving system can be frequency dependent, a correction is preferably applied to determine the true ratio of the 3rd harmonic dynamic dipole moment to the base frequency dynamic dipole moment. This ratio can be measured over a predetermined integration period. For each sensor, this ratio can be calibrated to the amplitude of the oscillations or directly to the frequency offset and therefore corrections can be applied. In the case of even harmonics, the situation is somewhat more complicated, since the direction of the dynamic dipole moments does not coincide with the dynamic dipole moments of the base frequency. Thus, here it is usually necessary to use more than one coil, or the orientation of the coil relative to the sensor must be determined by other means. While a large set of coils (e.g. >=6) may allow both the position and orientation of the sensor to be recovered, multiple coils (e.g. 3-5) should at least allow the orientation of the sensor relative to the coils to be recovered using methods similar to location determination described in this document. The true ratio of dynamic dipole moments for even harmonics can then be determined using the coil sensitivities. The intermediate step of determining the orientation can be omitted, and a direct map of the ratios of the base frequency amplitudes and the harmonic amplitudes in the coils can be established using linear algebra methods. It should be understood that the methods described here in the frequency domain can be compared with methods in other bases, such as the time domain. In the time domain, frequency analysis is compared with waveform analysis. These matching methods are well known in the mathematical literature.

Determination of the oscillation amplitude based on the envelope function in the time domain will be described below.

Another way to determine the amplitude of the oscillations is to use the nonlinear behavior of the signal attenuation. Sensor attenuation is usually nonlinear. Nonlinear attenuation means that when the accumulated energy is doubled, the average power dissipation of the sensor does not double, but increases by a factor greater than two. The reason for this may be the stretching of the thread due to the force modulation described above. Equation (9) shows that at low oscillation amplitudes, the attractive forces between magnetic objects are largely constant, but at higher amplitudes this constancy is not maintained. This change in force depends, to a first approximation, on the square of the oscillation amplitude, which corresponds to the approximation of the cosine function by a parabola. This quadratic dependence is the reason for the nonlinearity of scattering. The changing force between the magnetic object periodically stretches the thread(s), resulting in dissipation. Other effects can also lead to non-linear behavior. Together, these effects cause the shape of the decay envelope curve over a given time to depend on the initial amplitude. Thus, if the sensor has a constant initial amplitude of oscillation, and the distance and/or orientation of the sensor is changed relative to the receiving coil(s), a scaled version of the initial attenuation envelope is found. However, when the excitation amplitude of the sensor changes, not only the overall amplitude of the attenuation curve changes, but also its shape. This means that amplitude effects and distance/orientation effects can be disentangled, and hence the initial amplitude of the oscillations can be reconstructed, for example, by using a lookup table of pre-recorded decay curves. This again leads to the possibility of zero amplitude frequency detection or constant amplitude controlled excitation as described above. This method only requires one coil to work. However, it is somewhat sensitive to sensor movement during acquisition, so this also changes the shape of the envelope. Therefore, it is preferable to include a model of probable sensor movements in the assessment. For example, if it is known that the sensor will not perform rapid accelerations, it is useful to adjust the decay curve envelope to account for the constant motion assumption.

Determining the amplitude of oscillations based on the amplitude response of the signal to changes in excitation fields will be explained below.

Another way to determine the amplitude of oscillations is to analyze the response of the sensor signal to different excitation forces. In this case, the current pulses are systematically varied and the response of the sensor(s) to various excitation pulses is evaluated. The current, duration and phase of the transmit pulse, or a combination thereof, may vary. For example, suppose there are two excitation pulses. If the distance is large and the amplitude of the local field is small, the two pulses are designed to generate twice the amplitude that one pulse would produce. However, if the distance is small and the local field at the sensor is large, the amplitude will be less than twice the amplitude. This leads to a characteristic reduction in the reception voltage by a factor of two compared to the expected one. Thus, the amplitude (Fourier) ratio(s) of the received sensor signal for a given drive pattern is a measure of the drive amplitude and can again be used to extrapolate to a zero-amplitude frequency and/or to obtain a constant drive amplitude. In addition, other quantities such as frequency and decay time can be estimated. The ratios of these quantities are also characteristic of the amplitude of the oscillations and can be used to extrapolate to a frequency with zero amplitude.

Determining the correct parameter based on a complete model of all contributing factors will be described below.

All the methods described above are only evaluation methods, and some methods require changes in the transmitted field pulses. No changes to the system hardware are required to perform these evaluations. Therefore, it is logical to implement them all. This can be done by simply running the evaluations in parallel and combining the results in a way that minimizes noise, i.e. obtain a weighted average according to the relative noise. Although this is relatively simple and easy to implement, better results can be expected by using a truly comprehensive mathematical approach, which will be described below. On the other hand, a mathematically complex approach is much more difficult to implement and may require too many computational resources on low-cost computer hardware. The basis for a correct mathematical approach is the mathematical model of the sensor. This model predicts the response of the sensor to the excitation fields, the current state of the sensor and the measured parameter, i.e. sensor environment.

The sensor state can only represent the current deflection angle and rotation speed of the suspended sphere. However, particularly but not exclusively for pressure sensors, this may also include the elastic state of structures that may deform under external or internal changing forces. Thus, it is necessary to include membrane and filament hysteresis models. This model can take various mathematical forms, but the most common way is to formulate the model as a set of differential equations. A model of the transmit and receive coils, including filter and amplifier characteristics, must then also be generated. This can be formulated in differential equations, although the Fourier parameter representation here is also not unusual if the transmitting and receiving systems are sufficiently linear in nature. Finally, it is necessary to provide a model of the transmit and receive sensitivity of the coil. This can be simply a set of spatial points with an associated sensitivity and an interpolation algorithm between the points. It can also be based on coil modeling based on Biot-Savart's law. This model can now predict the voltage response of the sensor at any given location and orientation with a given history of excitation pulses and external parameters. Thus, the procedure is to change the position and orientation of the sensor, and to have the sensor influence the physical parameters in the simulation so that the recorded signal and the simulation match best. Many well-known optimization techniques, such as gradient descent or random walks, can be used. Agreement can be defined as the root mean square value of the sum of the difference between the measured sample points and the modeled sample points. The best match will be if this value is the smallest. The best fit may be modified by introducing additional constraints, such as a model of expected relative positions and orientations, or a constraint on the maximum expected accelerations of the sensor and/or a model of the measured quantities, which, for example, provides a constraint on the maximum rate of change of those quantities. Additional sensor inputs, such as from accelerometers in a portable coil system, may also be used for at least one independent distance and orientation change inputs. Because full model-based estimation processes are computationally expensive, they can be combined with one or more of the previous methods to provide a good starting point for further optimization.

The processor may also be configured to compensate for gravitational effects, as will be explained below.

In a pressure sensor, the weight of the movable segment of the sensor can affect the pressure reading: if it is on top of the stationary part of the sensor, it compresses the sensor and thus causes an apparent increase in pressure, if it is located at the bottom, it results in an apparent decrease in pressure. The force of gravity (weight) on a 0.5 mm diameter NdFeB sphere is about 5 µN. In comparison, the change in force on the cylindrical back side of the sensor (conservatively assuming the diameter is the same as a sphere) created by a 1 mbar change in pressure is about 20 µN. Thus, the force difference between the up and down orientation of the sensor in air would limit the accuracy to 0.5 mbar. To mitigate this problem, a correction can be applied based on the spatial orientation of the sensor obtained using one of the methods described for (position and) orientation recovery herein. In a liquid medium such as blood, the effect of weight can be minimized by matching the density of the sensor segments to the density of the liquid so that buoyancy compensates for gravity.

The processor may also be configured to compensate for the Earth's magnetic field and other static field effects.

Static background fields add to the field of a stationary magnetic object and thus modulate the restoring field B _rest seen by the oscillating magnet. This changes the resonant frequency according to equation (8) and therefore introduces measurement errors through changes in the frequency of the oscillator. This does not matter for localizing the oscillator. For 0.5 mm diameter magnetic spheres made of NdFeB with a saturation magnetization of 1.3 T/μ ₀ , the fields generated by the stationary sphere at the center of the oscillating sphere are 16.1 mT and 6.8 mT for center-to-center distances 0 .75 mm and 1.0 mm, respectively. The Earth's magnetic field ranges from 25 to 65 μT. The frequency difference between parallel and antiparallel alignment of the static field component with a maximum earth magnetic field of 65 µT will create a frequency difference of about 5 Hz and 9 Hz for the above distances of 0.75 mm and 1.0 mm, respectively. Regarding the typical frequency resolution of 10 to 100 MHz and the prototype sensitivity of -0.3 K/Hz for the temperature sensor and 20 mbar/Hz for the pressure sensor, this worst-case calculation leads to significant errors in the measured values. Various mitigation strategies for this are presented below.

A mitigation on the sensor side is to use a design that uses two suspended spheres with the same magnetic dipole moment and moment of inertia (or a suitable ratio of the two quantities) instead of a single sphere. Since the counter-oscillation occurs at the same frequency, the effect of a static first-order bias field, such as the Earth's magnetic field, is canceled out.

Another mitigation strategy is to use absolute field sensors in the detector system to measure the magnitude and orientation of static background fields. Based on the sensor orientation determined using the methods described herein, a frequency or field correction can be calculated to obtain the correct pressure, temperature, or other parameters from the sensor. To measure static background fields, any magnetic field sensor with sufficient sensitivity and footprint that can be integrated into a detector system can be used. One cost-effective option would be 3-axis Hall effect sensors. An alternative would be a 3-axis temperature-compensated microbot array with a well-defined frequency with zero amplitude. By changing their corresponding frequencies, the magnitude and orientation of the background fields can be determined. Ideally, their resonant frequencies are chosen such that they do not interfere with the frequency of the sensor of interest. Instead of correcting for frequency offset, the estimation can also use the coils of multi-coil detection systems to generate small offset fields to balance the Earth's magnetic field and other background fields. If non-uniform fields exist in the field of view due to the presence of ferromagnetic material, multiple sets of 3-axis magnetic field sensors can be used to characterize spatial variations in the field. Based on the interpolated map (mapping) of the background field obtained from these measurements, a correction for the sensor at a known position and orientation can be calculated, or appropriate offset correction fields can be applied, or a mixture of the two correction methods can be used.

Pressure sensors, as well as markers, must have a high quality factor and must have a large frequency sweep in order to be sensitive to the measured value in the range required for a particular application. A high quality factor is especially important at high oscillation amplitudes, when the highest signal is generated. Since two magnetic objects have strong attractive forces, and these forces increase greatly with decreasing distance (up to the 4th power of distance, see equation (9)), both properties can deteriorate. Large forces result in relatively strong tension in at least one thread holding at least one magnetic object. This tension in itself does not lead to dissipation. However, especially at large oscillation amplitudes, the forces between the magnetic objects decrease and, consequently, the tension of the thread periodically decreases. This causes the filament to periodically lengthen and shorten, which usually generates heat. Therefore, power is extracted from the oscillator. The forces also depend strongly on the distance from the magnetic objects and become very large as the objects move closer to each other. This behavior is especially problematic for pressure sensors. The force between magnetic objects is functionally equivalent to external pressure. Thus, if external pressure increases, magnetic objects move closer, which in turn increases the apparent pressure. This effect is compensated for by using a measurement calibration curve to determine pressure, but this can lead to a situation where the sensor reaches a critical point in which magnetic objects are quickly attracted to each other and finally touch. This results in a condition where the sensor no longer works. This can be avoided by simply making the pressure sensor diaphragm or bellows more rigid. However, this reduces the sensitivity of the sensor, i.e. frequency shift by applied pressure. To solve this problem, a method for reducing force and changing force is described. It consists of only a piece of magnetic material that is magnetized in the opposite direction compared to another magnetic object, as shown in FIG. 34.

In fig. 34, pressure sensor 4001 includes a magnetic object 4008, which is a permanent magnet suspended by a thread 4006, which is preferably a high tensile wire, to a flexible portion 4010 of the housing 4002. The flexible portion 4010 is preferably a membrane, which may be a latex membrane. The remainder of the housing 4002 may be made of metal or polymer. Housing 4002 may be filled with gas or it may provide a vacuum space, the interior space being designated 4009 in FIG. 34. An additional magnetic object 4007 is attached with an adhesive 4011 to the inner end surface of the housing 4002. The two magnetic objects 4007, 4008 are typically magnetized in opposite directions. However, the stationary magnetic object 4007 also includes a reverse magnetization portion 4012.

Thus, if two magnetic spheres are involved, in this example at least one sphere receives a cap magnetized in the opposite direction. The lid is located next to another magnetic sphere. If one sphere is stationary (fixed) and the other is oscillating, it is best to have a cover on the stationary sphere. Thus, the dynamic dipole moment of the sensor is not reduced. Only the oscillation frequency is slightly lower. However, it is also possible to swap the roles of the spheres. The oppositely magnetized part is so small that at all working distances the total (net) force between the magnetic objects is still attractive. If the area with reverse magnetization is small enough, the attraction condition can be satisfied until the magnetic objects touch. There are several ways to make caps with opposite magnetization. One is to simply add some magnetic material on top of at least one magnetic object. The magnetic material can be soft magnetic or hard magnetic. It could be a solid continuous magnetic object, magnetic paint, or anything in between. Magnetic material tends to align itself so that it itself produces an opposing magnetization. In addition, it also tends to stick to a magnetic object. However, this additional material must be adhered to the magnetic object, especially if the two main magnetic objects may touch each other from time to time. To maintain the originally desired shape, some material may be removed from the magnetic object to be modified, for example by grinding. There is an alternative way to form a zone of opposite magnetization. It can be created simply by oppositely magnetizing the desired zone of a magnetic object. This can be achieved by sending strong current pulses through a conductor near a magnetic object. However, this is not very practical due to excessive heat. This can be achieved more simply by simply heating the converted part of the magnetic object to a temperature close to or above the Curie temperature. This will lead to a change in magnetization. The effect can be enhanced by applying a pulsed or constant magnetic field in the opposite direction. The fields may also include strong gradients when using any hard or soft magnetic material near the affected area. Since the heating must be fairly localized, the temperature increase must be very rapid so that the total energy deposited into the magnetic object is low and does not bring it entirely close to the Curie temperature. A suitable heating source may be a laser. Resistive or inductive heating methods may also work.

It will be explained below how the pressure sensor can be localized.

For a tracking system, it is necessary to determine the orientation and three-dimensional position of the marker, which may also contain a magnetomechanical oscillator, but also for a pure measurement system, it may be necessary to determine the orientation and position of the oscillator to provide improved sensor reading accuracy, as discussed above. For localization, two independent localization strategies can be used. In some situations, one strategy may be sufficient; in other situations, a combination of both methods may be useful to improve accuracy or identify systematic errors (for example, a strong ferromagnet in the work area) that lead to conflicting results from the two methods.

One strategy relates to localization based on coil sensitivity. This approach takes advantage of the fact that each i-th coil in a coil array has a different B _S,i (r) spatial sensitivity profile depending on its position and orientation. According to equation (3), one oscillator then produces a response with a characteristic amplitude for each coil, which is determined by the corresponding orientation of the dynamic dipole moment oscillation generator relative to B _S,i (r). To restore the position and orientation of the sensor, it is necessary to determine a set of direct functions given by equation (4). Ultimately, a mapping is required between the position and orientation coordinates of the 6 markers and the voltage amplitudes at the base frequency or higher harmonics for all receiving channels. The following equations describe how to remove the time dependence from equation (4) so that only amplitudes need to be considered. We start by including all the arguments, i.e. vector r=(x,y,z) ^T position and vector ϕ=(ϕ, θ, ψ) ^T orientation:

The required coil sensitivity profiles can either be calculated from known coil geometries, measured at given positions and then interpolated, or determined using both, i.e. model that can be fitted to experimental results with appropriate fitting parameters. For magnetization oscillations, an explicit description of the frequency ω and amplitude α ₀ of oscillations in the marker frame will be

where the prime indicates the local frame of the marker; expansions of trigonometric functions for small amplitudes α ₀ oscillations were used. Then the temporary change will be

where the first term characterizes the basic frequency response, and the 2nd term characterizes the frequency response of the 2nd harmonic. Using the rotation matrices R(ϕ), one can calculate the magnetizations for the general orientation in space, i.e. Thus, starting from (11), the voltage amplitudes for the base frequency and the 2nd harmonic frequency can be defined as

And

respectively. Accordingly, the total voltage for the i-th coil will be

From a set of forward functions (14) and (15) and measured response amplitudes, the position and orientation of the marker can be calculated by solving a system of equations using a nonlinear solver, which is a standard mathematical method. The accuracy of the solution will improve as the number of receiver coils increases, as well as the orthogonality (i.e., the magnitude of the differences) between their respective coil sensitivities increases. The discrepancy between the 6 unknowns and a larger (or smaller) number of receiving channels can be taken into account by solving the system of equations using the least squares method.

Localization can also be performed based on gradient field encoding. Because coil sensitivity localization is based on the amplitude distribution produced by the coil array, marker frequencies can be manipulated to obtain independent position information. For this purpose, a non-uniform magnetic field is generated, ideally having a constant field gradient over the workspace, for example, by applying low-frequency currents to selected coils of the coil array. This additional field changes the restoring field B _rest acting on the oscillating sphere, and therefore its frequency (equation 9). Due to the non-uniform nature of the field, the change in frequency will depend on the position and orientation of the marker. By sequentially applying multiple encoding fields (eg, a field gradient applied in 6 different orientations), all three positions and two of the three marker orientation parameters can be determined. The remaining angle can be taken from the higher order sensor's response to external magnetic fields, however, at the cost of the higher field strength required to generate sufficient higher order contribution. The basic coding idea is related to gradient coding in MRI; thus, both frequency and phase encoding can be performed.

For frequency encoding, a non-uniform field is applied during signal readout to obtain the required frequency offset. For the required spatial resolution, the applied coding field strength must be adapted to the frequency sensitivity of the marker and the frequency resolution provided by the system. Assuming that the frequency sensitivity of an NdFeB marker with a sphere diameter of 0.5 mm is , for a spatial resolution of Δr=1 mm and an expected frequency resolution of Δƒ=10 mHz, a field gradient of approximately

This gradient value is approximately 100 times smaller than the gradient of typical MRI systems. Thus, special water-cooled gradient coils are not required, but transceiver array coils can be used to generate the field.

For phase encoding, the non-uniform encoding field is applied before the signal is read, i.e. The position-dependent frequency offset is applied only for a short window during which the position-dependent phase offset of the signal increases. In the event that phase resolution is not sufficient for accurate localization, the duration and/or amplitude of the phase encoding pulses can be varied across successive excitations so that ambiguities in the phase rise (greater than 2 pi) can be discerned. In this way, complete spatial information is obtained within a few readings. Phase encoding with one non-uniform field configuration (eg, one spatial axis encoding) can be combined with frequency encoding with another non-uniform field configuration (eg, orthogonal spatial axis encoding) for efficient localization. If the approximate position of the marker is already known from a sensitivity encoding approach (which is faster due to its parallel nature), it will be sufficient to use only a few phase encoding steps, which provide the missing high-resolution components (high spatial frequency) but not the full spatial information.

As described herein, comparison of localization results obtained using gradient field encoding and sensitivity encoding can be used to identify systematic errors, such as those resulting from background fields. In addition, it should be noted that the linear response to low-frequency external fields of sensors using two suspended spheres may be suppressed; in this case, the higher order frequency response can be used for localization or performance testing. However, the field sensitivity of these oscillators is much lower, so higher gradient fields will be required to encode a gradient field.

Parameter determination and position determination for tightly coupled sensors will be described below.

Determining a position (meaning 3 position parameters and 3 orientation parameters) and measuring a parameter (such as pressure or temperature) is especially difficult if only a few coils are used. However, using just a few coils is cost effective and also preferable in some applications due to space constraints. Therefore, it is desirable to modify the detection procedure and hardware to handle only a few coils. One way to do this is to use multiple markers and/or sensors in a connected manner. "In a connected manner" here means that multiple sensors/markers, each operating at a distinct known frequency, are combined with a fixed relative orientation at a node. Typically, the sensors are attached to a rigid frame, but technically only the relative positions of the sensors/markers need to be known at assessment times. With a sufficient number of sensors, the position can be determined using only two coils. This is most easily seen when compared to traditional electromagnetic navigation systems. They usually consist of several, usually more than 6, transmitting coils and one receiving coil, which is positioned and the orientation of which is assessed. However, the rotation of the coil around its axis (the axis of the dynamic dipole moment) cannot be detected due to the rotational symmetry of the coil. In this comparison, a set of tightly coupled sensors can be thought of as a transmit array, and a single transmit/receive coil as a marker. Thus, it is possible to place the sensor/marker array somewhere in a ring around the axis of the dynamic dipole of the transmitting coil. It should be noted that if the coil is not round, the rings are not perfect circles in space, but this does not change the argument. So the position cannot be determined with a single coil, but with two coils (with a non-parallel dynamic dipole moment), the symmetry is broken and the position and orientation of the sensor/marker array can be determined. Estimation of various sensor signals is best accomplished using an integrated modeling approach, which is described elsewhere in this document. In short, they generate a model of each sensor/marker in the array, i.e. in the form of differential equations. This model predicts the response of a sensor to a given excitation. Together with a model of the transmitting/receiving system (including amplifiers, filters and coils), the overall response of the array can be predicted. Knowing the past excitation pulses (usually only the pulses for a few decay periods need to be known), the expected received signal for the sensor position and the parameter value can be calculated. The position/orientation of the sensor, as well as the physical parameter that the sensor measures, are now optimized to minimize the difference between the calculated and actually received signal. It is also possible to incorporate prior knowledge into this procedure, i.e. allow only the maximum speed of movement of the sensors relative to the coils. The only difference here from the previous described method is that this process is not performed for a single sensor, but for a set of related sensors in an array or for several arrays at the same time. The sensor array also contains a set of prior knowledge, namely the relative positions and orientations of the sensors/markers in the array. It is especially useful to use a full parametric approach, or at least zero-amplitude frequency extrapolation approaches, since it is difficult to ensure that all multiple sensors operate simultaneously at the required amplitude. However, the full model approach requires quite a lot of computational effort. To reduce the required computational power, it may be useful to first use the already explained approaches to estimate one sensor/marker separately and use their results as seeds for the final reconstruction of position and value based on the full model.

Several aspects of calibration will be explained below, focusing primarily on calibration in the presence of conductive and soft magnetic ferromagnetic material.

The presence of conductive and, in particular, soft magnetic ferromagnetic material can interfere with localization by distorting the fields created by the oscillating magnet of the marker or sensor and/or by distorting the field(s) created by the transmitting coil(s). To a lesser extent, sensor readings may also be affected, especially since the accuracy of compensation for amplitude effects may be reduced. Therefore, a field calibration procedure is desirable. In addition, it is also advantageous to have a measure for determining whether field disturbances may currently occur. So, first, methods for detecting disturbance problems are discussed. Typically, a localization system uses an array of transmit/receive coils. The coils may be separate transmit-only and receive-only coils, or the same coil may be used for both functions. In any case, in this configuration one coil can transmit and all other coils can directly receive the transmitted signal. The received signal is compared with stored reference values. If the actual received signal deviates too much from the stored values, some action is triggered, such as an inaccuracy warning, triggering a self-calibration process, or suggesting a calibration process that includes user interaction, or a combination of these. It is also possible to transmit with several coils simultaneously. The transmission pulse must contain multiple frequencies. This can be achieved by generating pulses or using frequency sweep or some intermediate well known in the literature. Frequency analysis is important because eddy currents flowing through conductive structures are highly dependent on frequency. Thus, a significant change may be that the ratio of the received signal at two different frequencies exceeds some limit. It may also be significant if at least one spectral component changes by a given amount. However, a uniform change across the entire spectrum may be due to a change in gain, for example in a receiving amplifier. Thus, if, for example, reception amplifiers are designed in such a way that gain changes are likely, this effect can be used to set a new gain value in software to compensate for this gain change. The above argument holds in the same way if a gain change is expected to occur in the transmit amplifier but not in the receive path. Here, as a correction, the transmitted amplitude changes in the computational model (which leads to changes in the amplitudes of the sensor oscillation, etc.). It is also theoretically possible to measure the impedance of a single coil and use its change as an indicator of changes in the eddy current environment. However, the ability to measure impedance does not come naturally with electronics and requires specialized equipment. To detect changes in eddy currents in the environment, not only coil connections can be used, but also the known properties of sensors/markers in the operating range. In particular, it is possible to embed sensors directly into an array of transmit/receive coils. Even one sensor/marker is useful. For example, if a system incorporates one marker in a fixed position relative to the coil(s), the change in marker response is indicative of a modified eddy current environment. It is even more advantageous to use a sensor/marker that is sensitive to low frequency magnetic fields, but insensitive or only slightly sensitive to other physical properties that can change quickly. This marker not only indicates a static magnetic field, but also the presence of ferromagnetic material. To detect ferromagnetic material, the coils will be powered not only by current at the oscillating frequencies of the sensor/marker, but also by current at a much lower frequency. The current can be supplied coil by coil or using multiple coils. If the measured sensor response (i.e., the change in frequency due to an applied low-frequency magnetic field) does not match the stored expectation, it is likely that the ferromagnetic material is distorting the field. If the system has enough coils, it is not even necessary to install the field dependent sensor/marker at a known position. With enough coils, the position of the marker can be determined using the sensitivity of the coils at the oscillation frequency of the sensor/marker and independently using the sensitivity of the sensor/marker to near-DC current magnetic fields (gradient field encoding). If the positions obtained by the two methods diverge, the eddy current (or ferromagnetic) medium has changed. However, it is even better if not one such sensor is built into the system, but several. It is also better to have them in known positions than in unknown positions. But it is also useful to know at least some of the properties of the provisions instead of having no information about the provisions. A practical way of partial knowledge is to place sensors/markers on a rigid structure that provides a known and time-stable position and orientation relative to each other. Such a calibration “frame” with sensors/markers can be placed permanently or from time to time in the working volume of the localization system. If the localization system detects relative positions and orientations that deviate from expectations, the system experiences disturbances caused by eddy currents or ferromagnetic material. If, again, the sensors/markers are also sensitive to near-DC magnetic fields, and there are enough coils in the coil array, then the relative positions of the sensors/markers can be determined independently at very low frequencies, at which only ferromagnetic material causes disturbances to the fields , and at the resonant frequency of the sensor/marker, at which both ferromagnetism and eddy currents lead to field disturbances. Consequently, information can be obtained about the nature of the objects causing the disturbances, for example if ferromagnetic material contributes to such disturbances. Again, the best approach for detecting disturbances is a complete mathematical model of the transmit/receive amplifiers, coils, and marker/sensor(s). This model also includes known positions and orientations, both absolute and relative. In the first stage, all positions/orientations and physical parameters are optimized so that errors are minimized. This stage includes prior knowledge, for example, of fixed-position markers attached to the coil array and relative positions within potential frames. As an additional note, a "frame" does not need to be introduced just for calibration purposes, but a marker consisting of multiple oscillators can itself act as a frame. In the second step, the total weighted error between the expected signals and the delivered signals is calculated. If the error exceeds a specific threshold value, it is concluded that some material is causing the fields to be disturbed. The nature of the error (i.e. whether it occurs on AC-sensitive components or DC-sensitive components) can determine the nature of the material causing the disturbance.

The latter method of determining the presence of field disturbances is also a good starting point for methods to compensate for the effects of field disturbances. The method is most easily illustrated by assuming that there is a conductive material present that produces eddy currents, but not a ferromagnetic material. By applying the model described above, we obtain correct positions from estimating near-DC current-dependent signals (gradient field encoding) and incorrect local field positions and amplitudes at the sensor frequency and its harmonics (coil sensitivity encoding). Therefore, we can distort high-frequency fields so that they match expectations. After applying the distortion, all sensor positions and readings will be improved. It is useful not to rely solely on position estimation based on near-DC current magnetic fields, since AC sensitivity encoding is much faster. The most important part of this compensation method is determining the correct AC field distortion model. A simple solution is to parameterize the field shift function, for example using simple 3D polynomials. This means that the field value is not used for the actual position, but for the position transformed by the 3D polynomials. This is computationally efficient, but may lack physical understanding, and it is unclear how, for example, coil coupling measurements could be incorporated into this framework. Therefore, it is better to use models that are closer to physical reality. For example, it is better to use a field model of conductive plates in the vicinity of a coil system to induce the required field distortions. So basically the position, angular thickness and size of some virtual plates are changed until the model expectations and the measured data match. How to model such conductive plates is well known in the electromagnetic modeling literature. This type of modeling has the added benefit that it is easy to include shapes of objects that are likely to be encountered in a particular environment. Thus, if a special device is brought close to the field of view, such as the C-arm of an X-ray machine, and this device is known and can be simulated previously, then the system software only needs to optimize the exact orientation and position. Another advantage is that the estimated position of the object causing the disturbance can be displayed by the system, or the data is transmitted to a second system that performs the mapping task. Thus, the user may be specifically pointed out to objects that are interfering with the measurement, and the user may want to move or remove them. During this process, the coil communication data essentially acts as an array of metal detectors. The inclusion of a ferromagnetic material is conceptually the same as with a conductive material that creates eddy currents. However, modeling a ferromagnetic material is a little more computationally complex, and since there may not be a specific starting position defined by special markers, it may not result in an accurate position. But again, it is best to model an array of ferromagnetic materials, such as sheets and rods, and place and deform them in the simulation around an array of coils. Here it is very useful for the model if a database of probable ferromagnetic objects is provided. In addition, the process of measuring mutual coupling can be complemented by measuring the harmonic generation in the coil environment. The presence of harmonics is a strong signature of a soft ferromagnetic material, and the measured signals provide valuable information about the size and position of objects.

The generation of excitation pulses will be described below.

The system preferably includes software for generating the timing and shape of the excitation pulses. This drive pulse generator is preferably aware of the hardware capabilities. Various types of amplifiers and filters are possible. One type of amplifier is capable of generating a current signal that closely follows a fairly random path. In this document they are referred to as "analog amplifiers". Another type of amplifier is only capable of increasing current at a predetermined rate, decreasing it at the same rate, and making it more or less constant. Essentially, these amplifiers apply a positive or negative voltage to the coil or act as a short circuit. In this document they are referred to as "digital amplifiers". Digital amplifiers can have different switching speeds, i.e. permissible number of state changes per unit of time. If the switching speed is much higher than the oscillation speed, the digital amplifier again acts as an analog amplifier. Therefore, this type of amplifier can be conceptually considered as an analog amplifier. If the switching speed is approximately the same as the marker/sensor oscillation frequency, the processing should be slightly different. However, this is a more complex situation, so the entire discussion will focus on this. This type of amplifier has some advantages over analog ones. The main advantage is that the efficiency of this amplifier is usually very high and 98% efficiency can be easily achieved. Another advantage is that interaction with the computing system is very simple. There may be a matching circuit between the amplifier and the coil. The simplest matching circuit is simply a capacitor connected in series to a coil. When using a matching circuit, the maximum current through the coil at a given amplifier supply voltage increases. However, such a matching circuit has the disadvantage of blocking low-frequency currents. Some sequences may require low frequency currents. Solutions to this problem can be twofold. First, a matching network can be provided that is transparent at high and low frequencies. An example of such a circuit would be a series-connected coil or coil capacitor circuit in parallel with the first matching capacitor. Another method is to have a switch that bypasses the matching circuit, and when close to DC current is required, the switch closes. A capacitor can also be built into the bypass path if the resonant frequency is low enough. Likewise, a variety of different matching frequencies can be provided by using a variety of switches and capacitors. Another thing to note is that even if the circuit is set to near DC current, some current at the resonant frequency of the marker/sensor is still available. It should be noted that it is not necessary to use constant currents during reading. There are two main elements that provide this capability. First, direct currents should not be allowed to interfere with readings. There is a serious problem if the transmit and receive coils are combined. The DC source can provide a short circuit path for the signal. This should be avoided and a proper matching network can avoid this. The matching network must provide a sufficiently high impedance between the coil and the DC source. This can be achieved by using an additional coil in series with an inductance of the order of that of the transmit/receive coil. The inductor may have a parallel switch to short it if it is not needed. There are many other solutions available. The second condition is that the DC source does not create too much noise, i.e. Noise from the current source did not interfere with accurate marker/sensor measurements. This can be achieved by using a suitable analog filter in case of DC transmission. This filter can be bypassed during AC transmission pulses using a suitable switch (eg MOSFET optocouplers). It may also be advisable to avoid switching in the DC source while receiving a signal and simply use a slowly decaying current in the coil. It may also be advisable to perform just a few switching steps during reception and simply delete the received data if it is corrupted. The DC field sources can also be completely separate coils, or the field generators can be (moving) permanent magnets. This avoids most problems. An additional problem associated with the presence of constant currents during signal reception is that the coils may provide a different environment for the sensors. This means that, for example, some coils may be shorted to AC currents and the AC field no longer penetrates the coils, changing the field values in neighboring coils. This effect must be taken into account when calculating sensor positions and values. Two main field elements interact with the sensor/marker. One of them is the amplitude of a current close to direct current, i.e. current value averaged over a time of about 0.1 second (from about 0.01 second to about 1 second). The other is the Fourier amplitude (as a complex value since phase is important) at the resonant frequency of the sensors/markers. So the first task is to match the two values to generate a sequence.

The comparison, if required, of Fourier amplitudes and currents with a specific time domain pulse pattern will be described below.

It is also useful to create a software subsystem that performs this exact type of matching, i.e. a piece of software that takes the required near-DC currents and the desired Fourier amplitude (and frequency) as input, and which generates time-domain pulse trains. It is also desirable for this software to return information about whether the required values can be achieved within limits imposed by the hardware, such as maximum currents or maximum coil heating, or regulatory limits such as patient heating or peripheral nerve stimulation. Instead of simple yes/no information, information about the severity of unwanted side effects may be provided. This information can be provided for each individual transmission channel (for each transmit coil). An additional return value may be the actual best-fitted output DC current and Fourier amplitude(s). The input data may not only be a combination of a single frequency and Fourier amplitude, but also multiple Fourier amplitudes at different frequencies. The maximum pulse train length can also be a parameter that is an input to this function. The internal processing is as follows: In the case of analog amplifiers, the first result can be generated simply by performing an inverse Fourier transform of the required Fourier amplitudes (and DC values) for the required transmission time. If this process results in a waveform that cannot be realized due to some limitations, this is reported and perhaps a scaled version is planned for generation. Possible filter characteristics are taken into account by appropriate convolution. If there are several switchable filter states, they can all be tested and the one that places the least demands on the amplifier can be selected. It should be noted that there are several heuristic approaches available, so in most cases not all filter states need to be evaluated. For example, filters with frequencies far from resonant can be omitted if better ones are available. For a digital amplifier, the inverse Fourier transform (including filtering effects) provides a good starting point for optimization. In this first approximation step, the resulting peaks in the time spectrum are approximated by two (or at best several) sloping regions and flat regions in between. So, for example, the half-cycle of a sine wave starting at zero and ending at zero is approximated first by a flat part (zero), then an upward sloping part (rise), then a flat part, then a downward sloping part (descent), and finally a flat part (zero). region. The different sections are arranged in time to reach approximately the same area. After this first approximation step, a second step is performed in which the positions of the start of the inclined part and the flat part are shifted to achieve the best fit to the required Fourier values. The best option may be the smallest sum of squares (complex) values of the difference between the desired and achieved Fourier components. You can use all the usual optimization algorithms, such as gradient descent.

Below we will describe the comparison of the required Fourier values on the sensors / markers with the currents in the coils.

The next higher level of abstraction of a pulse generation program is a piece of software that requires specific Fourier values and field directions at a specific position as input and converts them into coil current requirements. The estimation algorithm usually provides some measure of the position and orientation of the sensors/markers. The position is not and should not be a position in three-dimensional space. However, 3D position is the ideal case.

For example, if only one coil is present, it may be possible to determine the field value only in the sensitive direction on the sensor. However, this also results in some virtual position and orientation in 3D space. Thus, these situations do not require special handling in the software. In this case, the conversion to the coil current requirements is the result of an optimization process. There is a model that calculates the Fourier components of the field in specific spatial positions from the currents in the coils. This is the basis for optimization, in which the Fourier components of the coil current are optimized to generate the required field components. There is usually no clear way to generate the required fields from coil currents. It may also be that the required currents are incompatible with limitations in the hardware system. The lower-level software returns values describing the negative effects, and the software uses this information to optimize the currents. The goal of the optimization is to find a good compromise between the achieved Fourier field components at the sensors/markers and the negative effects. This means that the deviation from the desired fields and side effects are combined into a single number, and a maximum or minimum is determined for that number using standard optimization algorithms. The combined number may be a weighted sum of squares. Naturally, a huge number of working mathematical combinations can be found for this entity. Finally, this part of the program returns to the calling program (higher level) the achieved fields in positions and quality values for its optimization.

The generation of the required Fourier field values for markers/sensors will be described below.

At this level of abstraction, the software system actually deals with the measurements that need to be made. So the input to this program is the current requirements for what needs to be measured, how accurately and how quickly. These requirements depend on the actual application of the sensors/markers and are therefore not part of this document. Requirements can be very different. For example, if only one sensor is involved, then it is required, for example, to measure one value as accurately as possible, say, every 0.1 second. If the application is a tracking solution with multiple associated markers, the desired result might be that, say, every 0.1 seconds a position update is performed for the entire marker node, regardless of which of the markers/sensors in it contribute to signal (based on coil sensitivity) and that an independent position check using a gradient method is required every 1 second. This program also has access to the current state of the sensors/markers (position/oscillation parameters, etc.) and the simulation model described elsewhere in this document. From this, the optimal Fourier values of the excitation field can be calculated, including the direction for each sensor/marker. These parameters can be passed down to the previously described lower software levels (with the desired execution somewhere in the future) to eventually generate currents. In the case of a single sensor, this will work immediately and the plan can be written to the hardware output buffers. However, for example, for tracking a marker node, there is likely no pulse shape that ideally excites all the individual sensors/markers. In particular, phase is not suitable for all individual sensors/markers. Therefore, the software may have to try to concentrate the optimal excitation on only a subset of the available sensors/markers and try to find a solution that produces a workable pulse sequence. This is the general operating principle of optimization for this software. It tries to change the required excitation of different sensors and focuses on a few to still get the desired result. Conceptually, the simplest approach is to look at all possible subsets of sensors/markers and test which excitation subset provides the best information about the desired parameters. Since there are many possible subsets, the program needs to add some heuristics to reduce complexity. For example, one can first observe which other markers/sensors are also excited if a given one is excited, and they can always be grouped together. If a suitable solution is found, it can be written to the output buffer. Additional logic may be required to enable near-DC magnetic fields, depending on the hardware implementation. If the hardware is capable of applying DC magnetic fields during signal acquisition, the software does not need to do anything special other than apply one or more gradients during readout. However, if the DC gradients and readout are inconsistent, an additional optimization step is required that generates the correct DC field or gradient at some point in time between excitation pulses. The logic behind the optimization remains the same. The parameters are changed until the simulation predicts a good enough measurement value for the application.

The generation of the startup sequence will be described below.

The algorithm usually assumes that there is already enough knowledge about the available sensors to optimize the sequence. This is usually not fully available at the beginning of the sequence. For example, for a given application (device), it may be known how many sensors/markers should be present in the device and in what range the frequencies may be. But the exact frequencies and positions are not known. Therefore, a special startup sequence is needed that tries to find all possible sensors in all possible positions. The simplest possible startup sequence is as follows. The work volume is divided into a 3D spatial or abstract grid. An abstract mesh is a mesh that should be used when there are not enough coils for a full 3D encoding. Each spatial point is divided into different directions. The program looks at each position and each angle at that position and applies the maximum transmit power for a given frequency and preset transmission time. The system then registers potential signals from sensors/markers. Typically, one transmission pulse excites not only one marker, but also many others at the same time. However, this procedure ensures that even the sensor/marker with the weakest possible signal will also be detected. An optional next step is to excite each sensor individually with different amplitudes. Nonlinear properties can be extracted from this. Another optional step is to excite each sensor/marker in the presence of a DC field or measure the phase of the signal after the DC field (again in different directions) to determine the sensitivity of the sensors/markers to DC magnetic fields. These basic procedures can be greatly speeded up by using some knowledge of the system. For example, it is likely that if a distant volume is already being searched for a sensor/marker, many or all of the closer volumes have assumed the maximum possible amplitude for at least some angles. Therefore, for closer volumes, only a few remaining parameters need to be applied. The same logic can be used to evaluate the nonlinear nature of sensors/markers or their response to DC magnetic fields.

Strategies for high time resolution measurements are explained below.

For many applications, high time resolution is desirable. One example is the measurement of intravascular pressure, which is modulated by the cardiac cycle, which can reach rates of up to 200 beats per minute in humans. To determine pressure minima and maxima during the cardiac cycle, a minimum measurement frequency of about 5 Hz is required, preferably above 10-20 Hz, most preferably above 40 Hz. On the other hand, to ensure a good signal-to-noise ratio, a very high quality factor Q of the oscillator is required. A high quality factor Q means slow attenuation. Therefore, when the next measurement pulse is transmitted to the sensor, the oscillations from the previous measurement pulse are not completely canceled out and they may affect the measurement.

Thus, strategies are required to achieve high time resolution using magnetomechanical oscillators, both for localization and characterization. The simplest approach to obtain high time resolution is to simply reduce the repetition time. Repetition time refers to the time interval between successive excitation pulses. After each excitation pulse, the frequency and amplitude are determined, from which physical values and position can be calculated, as described elsewhere. However, the quality factor of the sensors/markers is usually relatively high, and the amplitude of the oscillations does not decrease much during the next excitation pulse. To always obtain the required sensor/marker excitation, the phase of the next excitation must be taken into account. Typically we require "in-phase excitation", that is, excitation in such a way that the sensor/marker receives energy from the very beginning of the excitation pulse. How to optimize timing (selection of timings) is described elsewhere. Common-mode excitation minimizes the transmission energy, and hence the duration of the excitation pulse can be minimized. This increases the overall signal to noise ratio.

High repetition rates have some disadvantages. First, during and shortly after excitation pulses, the system typically cannot take on values, and therefore the signal-to-noise ratio may not be optimal.

Secondly, each transmission pulse destroys some knowledge about the oscillation phase of the sensors. Only if the excitation pulse and the orientation of the sensor are strictly controlled and precisely known, the phase information can be preserved to some extent, however this is a technologically challenging task. Phase information over a longer period can be useful because it encodes information about the average frequency (hence the average physical quantity). Measuring the average physical quantity is much more accurate when estimating a double-length interval than simply estimating the first half and the second half independently and averaging the two results. Therefore, it may be advisable to not have as many excitation pulses as measurement pulses, but to extract more than one measurement value from one signal pulse. This can be done by simply dividing the signal into several parts and evaluating each part separately. This simple approach does not take into account that the measurements get better if a longer data set is used. To account for this, the set can be divided into a hierarchy of subsets and each subset within each hierarchy is scored and the means scaled to fit longer data sets. Thus, for example, a data set (one unperturbed decaying signal) is first evaluated as a whole. It is then split into two, and the two split data sets are evaluated separately. The same number is then added to each result so that their average matches the average of the full set. This process can be repeated to end up with 4, 8, and so on subsets. This approach can be refined mathematically prior to estimation based on the full model. To do this, a model is created for the evolution of the physical parameter (and possibly also including the spatial movement of the sensor). This model may be a polynomial of some degree or some other suitable mathematical function. The function must describe the physical nature of the measured quantity in such a way that only a small number of parameters are required. So, for example, when the parameter is blood pressure, a Fourier series may be better modeled because it describes the pressure waveform of the heart rate better than polynomials. The parameters are then changed to fit the measured data set as best as possible. If discrete measurement points are ultimately needed, they can simply be calculated using the model outputs for specific points in time.

It should be noted that all described embodiments include a diffusion blocking layer that covers at least a portion of the housing, and which is configured to maintain a predetermined pressure within the housing, even if the diffusion blocking layer is not shown in all drawings for reasons of clarity. Also, a pressure sensor configuration that provides temperature compensation for the resonant frequency may be applied to any embodiment described.

Other variations to the disclosed embodiments may be understood and implemented by those skilled in the art in practicing the claimed invention based on examination of the drawings, disclosure, and appended claims.

In the claims, the word “comprising” does not exclude other elements or steps, and the use of the singular does not exclude the plurality.

One block or device can perform the functions of several elements specified in the claims. The fact that certain measures are stated in mutually different dependent claims does not mean that a combination of these measures cannot be used to advantage.

Determinations such as determination of resonant frequency based on the induced signal, determination of pressure based on the resonant frequency, determination of the calibration curve, etc., performed by one or more blocks or devices may also be performed by any other number of blocks or devices. Control of the detection system can be implemented in the form of program code using a computer program and/or in the form of special hardware.

The computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium provided with or as part of other equipment, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.

Any reference symbols in the claims should not be construed as limiting the scope of the invention.

The invention relates to a passive pressure sensor for insertion into the human circulatory system and for wireless reading of its readings using an external reading system. The pressure sensor contains a housing with a diffusion-blocking layer to maintain a preset pressure inside the housing and a magnetomechanical oscillator with a magnetic object that provides a constant magnetic moment. The magnetomechanical oscillation generator converts an external magnetic or electromagnetic excitation field into mechanical oscillations of a magnetic object, wherein at least a portion of the housing is flexible to allow changes in external pressure to be converted into changes in mechanical oscillations of the magnetic object. A pressure sensor can be very small and still provide high quality pressure measurements.

Claims (39)

1. A pressure sensor for insertion into the human circulatory system, wherein the pressure sensor is a passive sensor, made with the possibility of wireless reading of its readings by a reading system located outside the person, wherein the pressure sensor further comprises a housing containing a diffusion-blocking layer covering at least a portion of the housing and configured to maintain a predetermined pressure inside the housing, wherein the pressure sensor further comprises a magnetomechanical oscillation generator containing a magnetic object that provides a constant magnetic moment inside the housing, wherein the magnetomechanical oscillation generator is configured to convert an external magnetic or electromagnetic excitation field into a mechanical oscillation of the magnetic object, wherein at least a portion of the housing is flexible to allow changes in external pressure to be converted into changes in mechanical vibration of the magnetic object, wherein the magnetic object is located inside the housing in such a way that it can rotate beyond the equilibrium orientation if the magnetic object is exposed to an external magnetic or electromagnetic excitation field, wherein the pressure sensor further comprises: a torque restoration unit comprising an additional magnetic object for generating a magnetic field at a position of the magnetic object such that it provides a restoring torque, wherein the torque restoration unit is configured to provide a restoring torque to return the magnetic object to an equilibrium orientation, wherein the pressure sensor is designed in such a way that changes in external pressure can be converted into changes in the resonant frequency. 2. The pressure sensor according to claim 1, wherein the flexible housing portion includes a bellows for converting changes in external pressure into changes in at least the amplitude or resonant frequency of mechanical vibration of the magnetic object. 3. The pressure sensor according to claim 2, additionally containing an outer cover over the bellows. 4. The pressure sensor as claimed in any one of the preceding claims, wherein the diffusion blocking layer comprises a noble metal or titanium or a metal coated with a biocompatible coating. 5. The pressure sensor according to any one of the preceding claims, comprising an outer wire frame attached to the outside of the housing to enable the distance from the outside of the housing to the wall of the vessel to be maintained. 6. The pressure sensor as claimed in any one of the preceding claims, wherein the magnetic object and/or the interior of the housing is coated with a slippery and non-stick material. 7. A pressure sensor as claimed in any one of the preceding claims, designed such that a magnetic object can be aligned with an external magnetic field regardless of the position and orientation of the pressure sensor in the external magnetic field. 8. The pressure sensor of claim 7, comprising an outer casing enclosing a housing, wherein the housing is rotatable within the enclosing casing, wherein the pressure sensor is configured such that changes in external pressure outside the enclosing casing can be converted into changes in external pressure occurring outside the housing and inside the surrounding casing. 9. Pressure sensor according to any one of paragraphs. 7 and 8, in which the magnetic object is a magnetic sphere that is attached to one end of the thread, wherein or the other end of the thread is attached to the inside of the housing, the thread having a length of at least Pi/4 the diameter of the magnetic sphere, or the other end of the thread is attached to a length-changing unit configured to change the length of the thread and attached to the inside of the housing. 10. The pressure sensor according to any of the previous paragraphs, configured to compensate for the dependence of the resonant frequency on temperature. 11. The pressure sensor according to claim 10, comprising a compensation element configured to change the resonant frequency in a first frequency direction depending on a change in temperature, which is opposite to a second frequency direction in which the resonant frequency of the pressure sensor would be changed depending on a change in temperature, if the compensation element were not part of the pressure sensor. 12. An implantable medical device containing a pressure sensor according to any one of paragraphs. 1-11, which is preferably one of the following: a hepatic bypass device, a cerebral aneurysm wire, an artificial heart valve, or a stent. 13. Reading system for wireless reading of pressure sensor according to any one of claims. 1-11, containing: - a field generator for generating a magnetic or electromagnetic excitation field for inducing mechanical vibrations of the magnetic object of the pressure sensor, - a transducer for converting the magnetic or electromagnetic field generated by the induced mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals, - a processor for determining a pressure value based on the electrical response signals, wherein the processor is configured to apply a compensation algorithm to adjust the determination of the pressure value for dependence of the electrical response signals on at least one of the following: a) distance between pressure sensor and field generator; b) phase of mechanical oscillation of a magnetic object; c) orientation of the housing relative to the reading system; And d) amplitude of mechanical vibration of a magnetic object. 14. A method for measuring pressure for performing a measurement using a pressure sensor according to any one of claims. 1-11, In this case, the method for measuring pressure contains the following steps: - generating a magnetic or electromagnetic excitation field to induce mechanical vibrations of the magnetic object of the pressure sensor, - conversion of the magnetic or electromagnetic field generated by the induced mechanical vibrations of the magnetic object of the pressure sensor into electrical response signals, - determining the pressure value based on the electrical response signals, wherein in the determining step, the pressure value is adjusted to depend on the electrical response signals from at least one of the following: a) distance between pressure sensor and field generator; b) phase of mechanical oscillation of a magnetic object; c) orientation of the housing relative to the reading system; And d) amplitude of mechanical vibration of a magnetic object. 15. A non-transitory medium containing a computer program containing program code for causing the sensing system (1501) of claim 13 to perform the steps of the pressure measurement method of claim 14 when the computer program is executed on a computer controlling the sensing system (1501).

Images