JP2013029512A - System and method for portable electronic device that detect attitude and angular velocity using magnetic sensor and accelerometer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system which determines motion information including a dynamic attitude and an angular velocity of a body.SOLUTION: The system includes a magnetic field detection device 71 which measures intensities and/or directions of a magnetic field in three substantially orthogonal directions on a main-body reference coordinate system; an acceleration detection device 72 which is adapted to measure all accelerations of the body on the main-body reference coordinate system; and a processor 75 which calculates an attitude and an angular velocity by combining measurement data on all the accelerations, measurement data on the magnetic field, and a kinematical model in a filter.

Description

  The present invention relates to an input technique for an electronic device. More specifically, the present invention relates to an angular velocity and a dynamic attitude angle for generating an input signal corresponding to an attitude or attitude change, or an angular velocity, to an application program executed in the electronic device itself. The present invention relates to an electronic device or apparatus adapted to be generated using a three-axis magnetometer and a three-axis accelerometer.

  Portable devices and, in particular, but not limited to, portable wireless devices (mobile phones, cell phones, cordless phones, text message devices, pagers, talk radio, portable navigation systems, portable music players, portable video images Players, portable multimedia devices, personal digital assistants (PDAs), portable games, etc.) are increasingly used in everyday life. As technology advances, portable electronic devices will come to implement more applications while being reduced in size and weight. Typically, the user interface and power supply make up the bulk of the portable device's volume and weight.

  The user interface of the portable device, and more specifically, the signal input unit of the user interface is very important for the operation and operability of the portable device. Traditionally, user command input and data input to a portable device has been performed using an input device such as a keyboard or keypad, mouse, joystick, touch pen or digital pen, or a gesture using the device itself. . Arrow buttons, thumbwheels, gaming handles, and other devices may be included in the portable device for screen scrolling and menu navigation.

  However, as portable devices become more sophisticated and smaller, input with traditional keypads, arrow buttons, thumbwheels, or digital / touch pens is inconvenient when the component parts are minimal. Can be unrealistic or not fun. More complicated menus, 3D maps, and modern games that require more sophisticated navigation exacerbate this problem.

  In order to generate input signal data relating to motion for applications implemented in portable devices, development of motion sensing devices such as accelerometers and gyroscopes and their implementation in portable devices have been proposed. An accelerometer measures the acceleration of a specific force, ie, “total” acceleration, including gravitational acceleration. Typically, a three-axis accelerometer measures three orthogonal components of the total acceleration vector. A magnetometer is, for example, a magnetic field sensing device that measures the strength and / or direction of a local magnetic field. A triaxial magnetometer detects the strength and / or direction of a magnetic field by measuring the local x-axis component, y-axis component, and z-axis component of the magnetic field vector.

  A magnetic compass is a device that can provide an azimuth angle of the earth's magnetic field relative to the north pole. A simple magnetic compass can be a simple magnetic field sensing device that provides an azimuth, or “heading”, of the earth's magnetic field with respect to the north pole when placed in a horizontal plane. A rate gyroscope is a specific device that measures the angular velocity applied to the sensitivity axis.

  In recent years, MEMS (microelectromechanical system) accelerometers and magnetometers have been widely combined to form an “electronic compass”, “digital compass”, or “orientation sensor”. The acceleration data is used to calculate roll and pitch angles. Also, the calculated roll and pitch angles are used to convert the geomagnetic field measurement data obtained by the magnetometer measured in the carrier's main frame reference coordinate system to the magnetic field measurement data in the local level reference system. used. The measurement data of the magnetic field in the local level reference system is used to calculate the yaw angle, i.e. the heading angle.

  For example, US Pat. No. 7,138,979 to Robin et al. Discloses a method and system for generating an input signal based on the orientation of a portable device. Robin discloses a gyroscope and / or accelerometer that uses a camera to detect device changes in spatial orientation and to generate a position signal representative of the changes. According to Robin, the input signal can be used for moving the cursor, manipulating game elements, etc.

  U.S. Patent Application Publication No. 2006/0046848 to Abe et al. Discloses a game suitable for execution on a portable device with a vibrating gyro sensor. A vibratory gyro sensor detects angular velocity from changes in vibration resulting from Coriolis forces acting in response to changes in direction. According to Abe's teaching, the gyro sensor detects the angular velocity of rotation about an axis perpendicular to the game display screen. Two-dimensional rotation angle data is calculated from the angular velocity data.

  However, the gyro sensor disclosed by Robin and Abe is expensive and relatively large in size and weight. Robin and Abe also deal with the two-dimensional “orientation” of the portable device rather than the three-dimensional “posture” of the portable device. Accordingly, it would be desirable to provide a method, device and system for generating input signal data for a three-dimensional attitude of a portable device. It would also be desirable to provide devices and systems that generate input signal data that are more economical, smaller, and lighter than conventional devices with gyro sensors.

  Traditionally, gyroscopes have been an important component of inertial attitude detection systems to provide angular velocity and dynamic angles. However, providing a 3-axis gyroscope can significantly increase cost, power consumption, and size. This is not desirable in battery powered portable devices. Furthermore, current low cost MEMS gyroscopes themselves are subject to many operational problems, such as bias drift, and are still not as convenient for portable consumer electronic systems as magnetometers and accelerometers. According to the present invention, angular velocity and dynamic angle can be detected using an electronic compass, as described below.

  In contrast to a gyroscope, an electronic compass is beneficial in that it can detect yaw, pitch and roll angular velocities as well as inertial position. The gyroscope does not provide complete angular position information, but rather provides only relative changes in angular position information.

  Also, gyroscopes tend to be relatively large compared to magnetometers. For example, a three-axis magnetometer can be manufactured up to about 0.2 inches x 0.2 inches x 0.04 inches (about 5 mm x 5 mm x 1.2 mm). A three-axis gyroscope will be significantly larger.

  Conventional attitude sensing devices include a two-axis or three-axis accelerometer, a three-axis magnetometer, and a three-axis gyro to provide a complete motion state, ie, pitch, roll, and yaw. Despite accelerometers becoming cheaper, gyroscopes remain several times more expensive than accelerometers due to their technical and manufacturing complexity.

  Furthermore, in an ideal free space, i.e., under the condition that gravitational acceleration is zero and no magnetic field is present, the six degrees of freedom of motion information can be achieved with a two-axis or three-axis accelerometer and a three-axis magnetometer. Can be collected using. However, on Earth, the presence of known gravitational accelerations and magnetic fields can serve as a useful reference, so motion information can be determined using a different method than that in free space. As a result, the magnetic field sensing device that replaces the gyroscope should be very inexpensive.

  An "electronic compass", i.e. an electronic compass, a digital compass, or an orientation sensor, inter alia, combines an accelerometer and a magnetometer in a typical carrier body to provide a magnetic heading angle . However, there are two major drawbacks to using an electronic compass for motion-based applications that can be performed on a carrier. First, angular velocity is not measured directly. Second, tilt angle (roll and pitch) measurements are only accurate when the carrier is stationary and therefore cannot compensate for dynamic effects. Failure to compensate for dynamic effects results in inaccurate magnetic heading angles.

  Townsend et al., US 2003/0158699 (hereinafter referred to as “Townsend”) includes a static roll, a static pitch, and a static yaw angle. An orientation system that combines accelerometer measurement data and magnetometer measurement data is disclosed. However, Townsend is not taught how to determine dynamic roll, dynamic pitch, and dynamic yaw angle and angular velocity. Therefore, Townsend's system suffers from both the disadvantages of a typical electronic compass.

  US Pat. No. 7,216,055 to Horton et al. (Hereinafter referred to as “Horton”) includes a three-axis gyroscope, a three-axis accelerometer, and a posture and nose reference system (including a magnetic field sensing device). AHRS) is disclosed. Unfortunately, gyroscopes significantly increase cost, size and power requirements. Horton has built-in gyro data to obtain roll, pitch, and yaw angle solutions. In addition, Horton includes a Kalman filter to estimate roll, pitch, and yaw angle errors. Similar to the gyro bias, this is used to cancel the attitude drift that continues to drift over time.

  Kalman filters are well known to those skilled in the art and are used in a variety of applications. However, the Kalman filter provides a standard computational framework with state vector equations and measurement vector equations. As a result, the equation of state and measurement equation should be specified for each individual application.

  In consumer applications, where cost is a very important factor, a low-cost solution that satisfies functional requirements will be the key to successful commercialization. Thus, not only to determine the posture and angular velocity of an object, but also to a posture detection and motion detection device for measuring magnetic field strength and acceleration around or in three orthogonal axes with respect to the object It is desirable to provide.

US Pat. No. 7,138,979 (Publication date: November 21, 2006) US Patent Application Publication No. 2006/0046848 (Publication date: March 2, 2006) US Patent Application Publication No. 2003/0158699 (Publication date: August 21, 2003)

  A posture detection and motion detection system for a portable electronic device such as a mobile phone, a game device, etc. is disclosed. The system, which can be implemented in a portable electronic device, comprises a 2-axis or 3-axis accelerometer and a 3-axis magnetic field sensor, like an electronic compass. First, measured values from the accelerometer and the magnetic field sensor are processed by a signal processing unit that calculates an attitude angle and an angular velocity. These data are then converted into input signals for a specific application program associated with the portable electronic device.

  More specifically, the system measures dynamic pitch / roll / yaw angles and angular velocities, and does not use a gyroscope to measure the kinematic characteristics of posture from accelerometers and magnetometers. It can be measured directly by using it in combination with data.

The foregoing and other objects, features and advantages of the present invention will become more apparent from the following description of preferred embodiments of the invention as illustrated in the accompanying drawings, in which like numerals represent like parts throughout the different views. It will be clear from the detailed description.
It is a figure which shows the attitude | position angle of the rigid object in the space based on a prior art. FIG. 3 is a block diagram illustrating a generator for generating an input signal according to the prior art. FIG. 2 is a diagram of an apparatus associated with a 3D map application using the techniques of the present invention. FIG. 2 is a diagram of an apparatus associated with a flight simulator game application using the techniques of the present invention. 6 is a flowchart of a method for supplying a posture signal and a posture change signal to an application program according to the present invention; (A) is a figure which shows the pitch angle with respect to a carrier main body reference system (carrier body frame of reference) and a local navigation tangent frame of reference (local navigation tangent frame of reference) based on a prior art. (B) is a figure which shows the roll angle with respect to a carrier main body reference system and a local navigation contact reference system based on a prior art. (C) is a figure which shows the yaw angle with respect to a carrier main body reference system and a local navigation contact reference system according to the prior art. 1 is a view of a posture and motion detection system according to the present invention. It is a figure of the filter of the 1st attitude and angular velocity concerning the present invention. It is a figure of the filter of the 2nd attitude and angular velocity concerning the present invention.

  The present invention relates to an attitude detection device that detects the attitude of an object and a motion detection device that detects an angular velocity of an object such as a carrier. The posture and motion detection device includes a 3-axis magnetic field sensor such as a 3-axis magnetometer, and a 2-axis or 3-axis acceleration detection device such as a 2-axis or 3-axis accelerometer. More specifically, the posture and motion detection device mounted on the carrier can receive input signals corresponding to posture and angular velocity for other applications in the carrier (eg, 3D (3D) maps, games, etc.). An electronic compass is used to determine the carrier attitude and angular velocity to generate.

(Reference system)
The posture of the rigid object 10 in space can be described by three angles: yaw, pitch and roll angle (see FIG. 1). Typically, these angles are referenced to a local tangent plane, eg, a plane perpendicular to the Earth's gravity vector, or the Earth's ecliptic plane. Yaw (Ψ) is defined as the angle measured in the true tangential plane, ie, clockwise from the earth's magnetic pole axis in the forward direction of the object 10 in the local tangential plane. The pitch (θ) is defined as the angle between the longitudinal axis of the object and the local horizontal plane. Conventionally, in aerospace applications, a positive pitch is called “up” and a negative pitch is called “down”. The roll (φ) is defined as the angle of rotation about the longitudinal axis between the local horizontal plane and the actual plane of the object. Conventionally, in aerospace applications, positive rolls are called “right wing down” and negative rolls are called “right wing up”.

6A-6C show a local navigation contact reference system (T), a standard body (“object” or “carrier”) reference coordinate system (B), and a local-level frame of reference. All three of these can be used to describe the pose of an object. The local navigation tangent reference system (T) is defined as the north (N) axis, the east (E) axis, and the lower (D) axis. Therefore, the axes in the local navigation tangent reference system (T) are denoted as X ^N , Y ^E , and Z ^D. The axis of the body reference coordinate system (B) is a positive Y ^B axis passing the positive X ^B axis passing through the forward object, the "right" direction, and is defined as the positive Z ^B axis "down" The

Optionally, the three axes of the acceleration sensing device and the magnetic field sensing device can be arranged along the body reference coordinate system (B). However, any offset or misalignment does not affect the results or teachings of the present invention. The x-axis (X ^L ) of the local level reference system (L) is defined as the projection of the x-axis (X ^B ) of the main body reference coordinate system on the local horizontal plane. The y-axis (Y ^L ) of the local level reference system ( ^L ) is perpendicular to the x-axis (X ^L ) of the local level reference system (L). The z-axis (Z ^L ) of the local level reference system ( ^L ) is downward.

(Input signal generation system)
FIG. 2 shows a block diagram of a typical input signal generation system 20. The three-axis magnetic field sensor 22, the three-axis accelerometer 24, the A / D converter 26, and the data processing unit 28 are configured and arranged so as to supply solutions of posture and angular velocity. The conversion device 29 uses the solution of the posture and the angular velocity in order to generate an appropriate input signal 27 for the application program 21.

The sensing devices 22 and 24 were measured as the attitude of the sensing devices 22 and 24 changed, that is, as the sensing devices 22 and 24 rotated around at least one of the X, Y, and Z axes. magnetic field strength M _x, M _y and M _z, and the acceleration a _x, and generates an output signal proportional to the a _y and a _z. Typically, the magnetic field sensor 22 detects the _{M x, M y, M z} , accelerometer 24 senses the _{A x, A y, A z} .

  The six magnetic field strength and acceleration parameters are transmitted to the processing unit 25. The processing unit 25 may be implemented in one or more processing devices 22, 24, or may be a separate electronic device, a local electronic device, or a remote electronic device. The processing unit 25 includes a signal and data processing unit for processing the measured parameter data. For example, the processing unit 25 may include an A / D converter 26 for analog to digital (A / D) conversion, a data processing unit 28 for processing data, and the like.

  More specifically, the data processor 28 may be adapted to process magnetic field and acceleration measurements for calculating attitude angles and angular velocities. Subsequently, these data can be input to a converter 29 adapted to convert the data into an input signal 27. Subsequently, the converted input signal 27 is transmitted to the electronic processing device 21 including an application or a driving program that operates the converted attitude angle and angular velocity data and reflects the data in the motion state.

According to the prior art, three-axis magnetic field sensor, X-axis, Y-axis, and the respective magnetic field strength M _x about the Z axis, M _y, may be adapted to measure M _z. On the other hand, the three-axis accelerometer can be adapted to measure the respective accelerations A _x , A _y , A _z in the X-axis, Y-axis, and Z-axis. Therefore, the pitch of the object 10 in the space is calculated by the following equation.

Then, the roll of the object 10 in the space is calculated by the following equation.

here,

Corresponds to the measured acceleration value (in g) from the triaxial accelerometer obtained in the carrier body reference system. Thus, both pitch and roll can be determined using a 2-axis or 3-axis accelerometer. The roll in equation (2) uses a biaxial accelerometer and the carrier pitch angle is relatively small, that is, when the carrier pitch angle is less than 20 degrees, even if the following equation is used: Can be calculated.

  The yaw calculation is slightly complicated and requires measurement data from both the accelerometer and the magnetic field sensor. More specifically, yaw can be calculated using the following equation:

and

here,

Corresponds to the geomagnetic field from the triaxial magnetometer measured in the carrier body reference system (B). Approximate angular velocities can be obtained by calculating the time derivative of the angular change using the following equations, respectively:

Here, ω _x , ω _y , and ω _z correspond to the angular velocities of the rotation of the object around the X axis, the Y axis, and the Z axis, respectively.

  This approach has the excellent advantage of low computational load. However, there are drawbacks of latency, dynamic error, neglect of roll / pitch / yaw coupling, and vulnerability to changing hard iron / soft iron conditions, which are unacceptable for some applications. This approach assumes that the rotational angular velocity of the body axis is equal to the rate of change in Euler's attitude angle, which is roughly correct. It is sufficient for those skilled in the art that with portable devices, the angle can be continuously changed by hard iron and soft iron strains caused by local concentration of hard / soft magnetic material (eg iron). Can be understood.

  The present invention provides highly accurate angular velocity and attitude measurements in harsh situations that can involve extremely dynamic manipulation and vibration.

(Improved attitude and angular velocity)
In order to achieve accurate and immediate attitude and angular velocity, the major improvements to the latest include the implementation of attitude and angular velocity filters in the system. The posture and angular velocity (denoted as first and second posture and angular velocity filters) are used as part of the filter equation of state, as described in more detail below. Two preferred embodiments of the filter are disclosed.

  Referring to FIG. 7, triaxial angular velocity data, dynamic roll, dynamic pitch, and dynamic heading angle measurement data to provide input data related to posture and movement to a program or application. A system 70 adapted to determine the is shown as another preferred embodiment of the devices 22, 24, 26 and 28 (FIG. 2). The system 70 is a three-axis magnetometer 71 adapted to dynamically measure the geomagnetic field and generate a magnetic field measurement data signal 79 represented in the carrier body reference system (B), the carrier body reference system. A triaxial accelerometer 72 coupled to generate total acceleration measurement data 78 represented in (B), and a data signal 78 to calculate triaxial angular velocity, triaxial acceleration, and attitude angle, and A processor 75 for receiving 79 is provided.

  Magnetometer 71 and accelerometer 72 have been described in detail above, except where it is necessary to explain their relationship and interaction with each other and their relationship to processor 75. It will not be described again in more detail. The processor 75 is particularly adapted to receive and store data from the magnetometer 71 and the accelerometer 72. Further, the processor 75 calculates and outputs triaxial angular velocity data and dynamic roll / pitch / heading angle data for input to an application or program running on the object or carrier. Adapted to use the data. The processor 75 uses one of the attitude and angular velocity filters 80, 90 to best combine the geomagnetic field measurement data 79 and the total acceleration measurement data 78 with the carrier attitude and angular velocity kinematic model. preferable.

  The earth's magnetic field is subject to local magnetic field distortions such as hard iron and soft iron distortions. In small portable device applications, hard iron and soft iron strains from the host system may continue to change. As a result, the processing device 75 includes a magnetometer automatic calibration module 84 (FIG. 7) and a compensation module 85 (FIG. 7). The magnetometer auto-calibration module 84 estimates magnetometer errors, including changing hard and soft iron distortion effects in harsh environments, based on the principle that the scalar length of the local magnetic field is constant. The compensation module 85 uses the estimated magnetometer error including hard iron and soft iron to correct the measured raw data of the magnetic field.

  Also, in small portable device applications, it is often the case that low cost accelerometers are not well calibrated in the factory. As a result, accelerometers can have significant bias drift over time and temperature ranges of operation. To further improve accelerometer accuracy, the processing device 75 includes an accelerometer automatic calibration module 76 (FIG. 7) and a compensation module 77 (FIG. 7). By taking advantage of the fact that the raw output of the accelerometer reflects only gravitational acceleration when the carrier is stationary, the accelerometer auto-calibration module 76 estimates the accelerometer error source. The compensation module 77 uses the estimated accelerometer error to correct the measured raw data for all accelerations.

  The magnetometer auto-calibration module 84 and the accelerometer auto-calibration module 76 restrict the measurement of the geomagnetic field vector and the trajectory of the measurement of the gravitational acceleration of the earth over time in a limited terrain area. Use. If there is no error in the measured value, the trajectory should be spherical. However, when there is a sensor error including a hard / soft error, the locus is an ellipsoid. Sensor error estimation can be accomplished by parameter identification of ellipsoids that satisfy arbitrary trajectory constraints. The magnetometer auto-calibration module 84 and accelerometer auto-calibration module 76 are various while supporting the possibilities of new hardware while ensuring compatibility, flexibility and adaptability of the system 70 to the host system. It is provided to improve performance in various operating environments.

  As a preferred embodiment, a framework approach based on the Kalman filter is used to simultaneously estimate roll / pitch / heading angles and angular velocities from geomagnetic field measurements and acceleration (specific force) measurements. Is used for the filter. A Kalman filter is an indirect measurement observed over time, a recursive optimal estimator that finds important parameters from indirect measurements including noise and other errors, simply from data measurements. It's not just for removing noise. As new measurements become available, they can be processed. The estimated important parameters are formed into a state vector. Indirect and noisy observations are formed in the measurement vector.

  State equations and measurement equations need to be established for the Kalman filter. The calculation of the Kalman filter includes two phases (phase): time propagation and measurement update. In time propagation, the state vector from the previous measurement time point to the current measurement time point is transmitted as the previous estimated value of the state vector at the current measurement time. The measurement update uses the current measurement vector to correct the previous estimate of the state vector.

  The posture of the carrier can be expressed mathematically using a posture quaternion (posture quaternion). According to Euler's theorem, the components of the attitude quaternion are expressed as follows:

  The cosine rotation matrix from the carrier body system to the local navigation system is directly formed using the attitude quaternion as shown in the following equation.

  Dynamic roll, dynamic pitch, and dynamic yaw angle are the cosine rotation matrix

And can be expressed as:

  The kinematic differential equation of the posture quaternion can be defined as:

Therefore, the differential equation of posture quaternion includes a matrix representation Ω _angularRate (4 × 4 asymmetric matrix) of angular velocity. The angular velocity matrix Ω _angularRate is composed of angular velocities for each coordinate axis, and is expressed as the following equation.

However,

  Referring to FIG. 8, the first posture and angular velocity filter 80 uses a state equation consisting of a differential equation of posture quaternions and a kinematic differential model of angular velocity. The kinematic differential model of angular velocity can be modeled using lower and / or higher order processes. In a preferred embodiment,

Can be modeled as a first order Markov process.

  The equation of state is given by

The state vector for the first posture and angular velocity filter 80 is expressed by the following equation.

  The roll / pitch / yaw calculated directly by the accelerometer data and magnetometer data is expressed as a pseudo-angle that is used as a measurement vector for the first attitude and angular velocity filter 80. The measurement equation of the first posture and angular velocity filter 80 uses equation (8) representing the relationship between posture quaternion and roll / pitch / yaw. The pseudo angle measurement is calculated as:

here,

Corresponds to the geomagnetic field from the triaxial magnetometer measured in the carrier body reference system (B),

Corresponds to the measured acceleration value (in g) from the triaxial accelerometer obtained in the carrier body reference system.

The pseudo roll, pseudo pitch, and pseudo heading calculation module 89 performs generation of filter 80 measurements, thereby calculating measurement vectors (pseudo roll, pseudo pitch, and pseudo heading). Module 89 may be a stand-alone device or part of processing device 75. The measurement update module 82 updates the state vector measurement to produce a computed output (Y _k ) 88 when the measurement vector is available. The updated state vector (X _k ) 86 is fed back to the time propagation module 83 recursively in order. The time propagation module 83 propagates the state vector between the measured values of the filter.

  Referring to FIG. 9, the second attitude and angular velocity filter 90 uses a differential equation of attitude quaternion, a kinematic differential model of angular velocity, and a state equation consisting of a magnetometer error. The state vector of the filter 90 of the second attitude and angular velocity is expressed by the following equation.

here,

Is a magnetometer error that, in the preferred embodiment, can be modeled as a constant value. Using a constant value is straightforward and consistent with the actual error behavior of the magnetometer, but one skilled in the art understands that a more accurate higher order model with increased computational load can be used. Has been.

  The measurement equation of the second attitude and angular velocity filter 90 uses the transformation equation of the magnetic field vector and the acceleration vector from the local navigation tangent reference system to the body reference coordinate system, and can be modeled as: .

Where Z _mag and Z _accel are respectively a three axis magnetic field measurement vector (after compensating for any estimated error including hard iron and soft iron in magnetometer error compensation module 85), and It represents a measurement vector of triaxial acceleration. C ^T _B represents a C ^B _T (formula (7)) transposed matrix.

Corresponds to the components of the geomagnetic field vector measured northward (N), eastward (E), and downward (D) in the local navigation tangent reference system (T).

These known local magnetic field vectors are, for example, a world magnetic model 73 (FIG. 7), such as the world magnetic model provided by the National Geospatial Information Authority (NGA) of the United States, the British Defense Geographic Center (DGC), etc., or Preferably, it is from a similar global magnetic field model. The seven magnetic field components calculated from the NGA / DGC world magnetic model are as follows.
-F-total magnetic field strength-H-magnetic field strength in the horizontal direction-X-magnetic field component in the north direction-Y-magnetic field component in the east direction-Z-magnetic field component in the vertical direction-I (DIP)-geomagnetic dip angle-D (DEC)-Geomagnetic declination (magnetic fluctuation)
The world magnetic model 73 automatically applies declination correction to the magnetic heading so that the measurement reference is true north, ie, geographical north rather than magnetic north. The earth's local magnetic field in the local navigation tangent reference system (T) is accurately known via the world magnetic model 73 for any position on the earth. In many applications, the carrier location solution is available from the GPS receiver 74 (FIG. 7). The world magnetic model 73 can be used to provide accurate local magnetic field data in the local navigation reference system (T) based on latitude / longitude / altitude data generated by the GPS receiver 74.

  Optionally, the local magnetic field vector may be pre-loaded with a substantially “universal” value in the second attitude and angular velocity filter 91. Although this option degrades performance, this option is preferred when GPS location information is not available. The above is not intended to be exhaustive to the precise shape disclosed, nor is it intended to limit the invention to the precise shape disclosed. The embodiments have been selected and described to provide an illustration of the principles of the invention and its application. Modifications and variations are within the scope of the present invention.

  The time propagation module 93 is adapted to time propagate the state vector between the measurements of the filter 90. Furthermore, the measurement update module 92 is adapted to update the state vector measurements.

  As disclosed above, if position information is not available for the world magnetic field model, the measurement vector and measurement equation designed for the first attitude and angular velocity filter is the second attitude and angular velocity. It can be formed and used for other filters.

  Since the systems and measurement equations of both the first and second attitude and angular velocity filters 80, 90 are nonlinear, the first and second attitude and angular velocity filters 80, 90 are expanded Kalman filters, sigma point Kalman filters. Can be implemented using a non-linear Kalman filter such as.

  In fact, using both the first and second attitude and angular velocity filters 80, 90, the static gravitational acceleration measurement from the triaxial accelerometer is dynamically caused by the carrier's own acceleration. It will be distorted. To reduce the effects of carrier acceleration, filter adaptive control modules 81 (FIG. 8), 92 (FIG. 9) monitor attitude and angular velocity filters 80, 90 by monitoring the length of acceleration vector measurements. Automatically adjust the gain. For example, when the length of the acceleration vector measurement is greater than 1 g, the gains of the filters 80 and 90 are reduced.

  Large time-varying magnetic field disturbances can have a significant effect on filter estimates. For example, if the carrier passes near a large magnetic source, this causes a large distortion of the earth's magnetic field, and the magnetometer measures this large distortion. Once the carrier has been exposed to a magnetic field that exceeds the magnetic field strength of the magnetic disturbance source, the geomagnetic field can be accurately measured. The filter is designed to react very quickly to situations where the measurement information is not reliable.

  In order to reduce the adverse effects due to magnetic field moment disturbances, a filter gain adaptive control 81 (FIG. 8) is provided to achieve a highly accurate state estimate. The filter gain adaptive control module 81 automatically adjusts the gain 87 of the filter 80 for the pseudo heading measurement data 88 by monitoring the scalar length of the magnetic field vector measurement. For example, when detecting a change in the scalar length of the magnetic field vector measurement, the filter adaptive control module 81 adjusts, ie, reduces, the filter gain 87 with respect to the pseudo heading measurement 88.

  Here again, the measured value using the second attitude and angular velocity filter 90 is formed by the measured value of the earth's magnetic field and consequently suffers a magnetic field disturbance. In order to reduce the effects of magnetic field disturbances, the filter gain adaptive control 92 is designed to achieve a highly accurate state estimate.

  The filter adaptive control module 92 automatically adjusts the gain 97 of the filter 90 for magnetic field measurements by monitoring the scalar length of the magnetic field vector measurements. For example, if a change in the scalar length of the magnetic field vector measurement is detected, the gain 97 of the filter 90 for the magnetic field measurement is reduced.

  As a preferred embodiment, the length (magnitude) of the geomagnetic field vector measured by the three-axis magnetometer and the length (magnitude) of the total acceleration vector measured by the three-axis accelerometer are bandpass digital Filtered by the filter. This removes both noise and DC offset (DC offset is the average amplitude of the signal) in order to obtain a reliable indication of magnetic field disturbances and acceleration changes for the filter adapted control modules 81 and 92. Can be done. The bandpass digital filter has a low cutoff frequency and a high cutoff frequency. Thereby, only signals with frequencies between the low cutoff frequency and the high cutoff frequency pass through the band pass frequency.

(Exemplary use of this technology)
The application of the electronic compass in the mobile phone 30 is shown in FIG. For this disclosure, the cell phone 30 is further adapted to execute a three-dimensional (3D) map program to allow the user to rotate the cell phone (and virtual space map) around all three axes. Has been. A conventional mobile phone with motion detection or a conventional mobile phone without motion detection has at least six input devices, for example two buttons for X-axis rotation, in order to realize the generation of input signals. You will need one button, two buttons for Y-axis rotation, and two buttons for Z-axis rotation.

However, if an electronic compass is used as the motion detection device, a direction pointing arrow button is not necessary. More specifically, with an electronic compass, as the mobile phone 30 is rotated, the sensor signal provides an attitude angle (Φ, θ, Ψ) and angular velocity (ω _x , ω _y , ω _z ). Can be processed. The posture angle and angular velocity can be input to a converter 29 that converts the posture angle and angular velocity into an appropriate input signal 27 to the application program 21.

In short, the generation of the input signal 27 does not require a directional arrow button, but rather the user can input sensor signals such as M _x , M _y , M _z , A _x , A _y and A _z . In order to generate, the attitude of the mobile phone 30 may be simply changed. If the application program is a 3D map application, it is possible to rotate the map around three axes. It is beneficial that the surface area of the panel that is required for conventional navigation buttons is not required. As a result, the surface area that should be used for the navigation buttons can be used for other purposes, and the mobile phone 30 can be made smaller.

  FIG. 4 shows an application for a flight simulator game that can be executed by the portable game machine 40. Although the game machine 40 is a flight simulator for the present embodiment, those skilled in the art can apply the teaching of the present invention to a myriad of game machines 40 and game programs including three-dimensional and attitude control. You can fully understand what you can do.

  Conventional gaming machines that control the attitude of an airplane require a number of input devices, such as buttons on the surface of the gaming machine, or alternative joysticks that are optionally coupled to the gaming machine. In contrast, according to the present invention, the electronic compass controls the attitude of the airplane by rotating the gaming machine itself along one or more of the X, Y, and / or Z axes. An aircraft attitude input signal is generated that can be used to

  Having described a system for motion and orientation detection and a portable electronic device having such a system, a method for supplying input signals of orientation and orientation changes to an application program, and determining the inertial orientation and inertial orientation changes of an object. A method for changing the operation demonstrated on the application program executed by the object and a method for generating an input signal for the application program executable on the portable electronic device are described below. Referring to the flowchart in FIG. 5 and FIG. 2, the method includes the step of mounting a two-axis or three-axis accelerometer and a three-axis magnetic field sensor on the portable electronic device (step 1). The method includes adapting the two-axis or three-axis accelerometer to generate a first set of signals (step 2A), and the three-axis to generate a second set of signals. The method further includes the step of adapting an accelerometer, for example an electronic compass (step 2B).

The first set of signals generated by the 2-axis or 3-axis accelerometer (step 2A) corresponds to accelerations in the X, Y and Z axes and / or changes in accelerations A _x , A _y and A _z These are proportional to changes in the inertial attitude of the portable electronic device. Similarly, a second set of signals generated by the three-axis magnetic field sensors (step B), the change in magnetic field strength and / or magnetic field intensity of the X-axis, Y-axis and Z-axis M _x, M _y and M _z These are also proportional to changes in the inertial attitude of the portable electronic device.

  Subsequently, the first and second sets of signals are processed (step 3). This may include, but is not limited to, converting an analog signal to a digital signal using an A / D converter. Subsequently, the digital signal is pitch, yaw, roll, i.e. the attitude of the device and / or changes thereto, and angular velocity around the X, Y and / or Z axis and / or changes thereto. Can be processed, for example, through a processing unit (step 4).

  Subsequently, the calculated pitch, yaw, roll and / or angular rotation is performed on the portable electronic device or an input signal compatible with an application program executable by the portable electronic device. (Step 5). More specifically, the calculated pitch, yaw, roll, and / or angular rotation is converted into an input signal that changes operation on the application program.

  For example, for use in connection with 3D image manipulation, acceleration and magnetic field strength are first calculated, followed by movement and movement of the 3D image along the X, Y, and / or Z axis, and / or the X axis, It can be adapted to describe the rotation of a 3D image around the Y and / or Z axis. Thus, if the portable electronic device is rotated about one or more of its own inertial axes, some or front of acceleration and magnetic field strength may change in pitch, yaw, roll, and / or angular rotation It will be a change that translates into a change in. When these changes are converted and input to an application program executed on the portable electronic device, the 3D image is moved in proportion to the input signal from the rotated portable electronic device.

  However, the application of the present invention is not limited to a portable device. Indeed, the present invention is applicable to any electronic device having a human machine interface, ie, a user interface, whether portable or not. For example, those skilled in the art will be able to use the pitch, yaw, and roll functions of the present invention with a mouse to generate an input signal to a personal computer, television, radio, DVD player, stereo system, or other multi-purpose. It can be adapted for, but not limited to, media devices and remote controls that generate signals to host devices such as electronic musical instruments (eg, electronic piano or electronic organ).

Claims (29)

Motion information including the dynamic object's attitude including the dynamic object's dynamic roll, dynamic pitch, and dynamic yaw measurements, and the dynamic object's angular velocity. A method of determining,
Providing a sensing device having a triaxial acceleration sensor and a triaxial magnetic field sensor;
Measuring at least one of the magnetic field strength and direction in three orthogonal or substantially orthogonal directions in the body reference coordinate system using the magnetic field sensor;
Measuring the total acceleration of the object in the body reference coordinate system using the acceleration sensor;
Combine at least one of the magnetic field strengths and directions and the total acceleration in the three orthogonal or nearly orthogonal directions and the kinematic model of the object attitude and angular velocity in a filter to calculate attitude and angular velocity. And
A method characterized by comprising. The filter is a first posture and angular velocity filter having gain, and a first posture and angular velocity filter having a state equation including a differential equation of posture quaternions and a kinematic differential equation of angular velocity. ,
The method according to claim 1. The kinematic differential equation of the angular velocity of the object is modeled as a first order Markov process.
The method according to claim 2. The first attitude and angular velocity filter uses pseudo roll, pseudo pitch and pseudo heading as measurement vectors, and uses the relation between attitude quaternion and roll / pitch / yaw as measurement equations,
The pseudo-roll and the pseudo-pitch are calculated using total acceleration, and the heading is calculated using magnetic field measurements converted from the main body reference coordinate system to a local level reference system.
The method according to claim 2. Further comprising updating the measurement of the state vector if the measurement vector is available;
The method according to claim 4. Further comprising time propagating the state vector between the measurement values of the filter,
6. The method of claim 5, wherein: Adjusting the gain of the first attitude and angular velocity filter to reduce measurement updates when the object is exposed to acceleration and magnetic field disturbances;
The method according to claim 4. The filter is a second attitude and angular velocity filter with gain, including a differential equation of attitude quaternion, a kinematic differential equation of angular velocity of the object, and an error model equation for the magnetic field sensor. A second attitude and angular velocity filter having an equation of state;
The method according to claim 1. The error of the magnetic field sensor is modeled as a constant value,
The method according to claim 8, wherein: The second attitude and angular velocity filter uses the measured value of the magnetic field from the triaxial magnetic field sensor and the measured value of acceleration from the acceleration sensor as the measurement vector of the filter, As a measurement model equation, a transformation equation of a magnetic field vector and an acceleration vector from the local navigation contact reference system to the main body reference coordinate system is used, and the magnetic field vector in the local navigation contact reference system is known.
The method according to claim 8, wherein: Further comprising updating the measurement of the state vector if the measurement vector is available;
The method according to claim 10. Further comprising time propagating the state vector between the measurement values of the filter,
The method according to claim 11. Adjusting the gain of the second attitude and angular velocity filter to reduce measurement updates when the object is exposed to acceleration and magnetic field disturbances;
The method according to claim 10. A global positioning system (GPS) receiver is used to determine the position of the object, the position including longitude, latitude and altitude;
The method according to claim 1. Further comprising using the longitude, latitude, and altitude in combination with a world magnetic model to provide a local magnetic field vector represented in a local navigation reference system.
15. The method of claim 14, wherein: The method further includes the step of automatically calibrating the three-axis magnetic field sensor by estimating an error source including a hard iron / soft iron strain item of the surrounding environment, wherein the estimation includes a local magnetic field vector having no error. Based on the principle that the length is constant,
The method according to claim 1. The method further includes the step of automatically calibrating the triaxial acceleration sensor by estimating a triaxial acceleration error source, wherein the estimation includes an error-free acceleration when the object is stationary. Based on the principle that the measured value is only gravitational acceleration,
The method according to claim 1. A system for determining movement information including posture and angular velocity of a dynamic object,
A magnetic field sensor adapted to measure in a body reference coordinate system at least one of the magnetic field strength and direction in three orthogonal or substantially orthogonal directions;
An acceleration sensor adapted to measure the overall acceleration of the object in the body reference coordinate system;
A processor adapted to calculate attitude and angular velocity by combining measurement data of total acceleration and measurement data of magnetic field and a kinematic model of attitude and angular velocity of the object in a filter;
With
A system characterized by that. The filter is a first posture and angular velocity filter having a gain, and has a system state equation including a differential equation of posture quaternions and a kinematic differential equation of angular velocity of the object. And an angular velocity filter,
The system of claim 18. The kinematic differential equation of the angular velocity of the object is modeled as a first order Markov process.
The system of claim 19. The first attitude and angular velocity filter uses a pseudo roll, a pseudo pitch, and a pseudo heading angle as the measurement vector of the filter, and uses the attitude quaternion and the roll, pitch, and yaw as the measurement equation of the filter. Using the relationship with the angle of
The pseudo-roll and the pseudo-pitch are calculated using total acceleration, and the pseudo-heading is calculated using measured values of the magnetic field converted from the main body reference coordinate system to the local level reference system.
The system of claim 19. The filter updates the state vector measurement if a measurement vector is available,
The system of claim 19. The filter causes time propagation of the state vector between the measured values of the filter;
The system of claim 22. The processor is configured to automatically calibrate the three-axis magnetic field sensor by estimating an error source including hard iron / soft iron distortion of the magnetic field from the surrounding environment of the three-axis magnetic field sensor. The estimation of the error source is based on the principle that the length of the local magnetic field vector without error is a constant value.
The system of claim 18. The filter adjusts the gain of the first attitude and angular velocity filter to reduce the update of the measurement when the object is exposed to acceleration or magnetic field disturbances;
The system of claim 22. The filter is a second posture and angular velocity filter having a gain, and includes a posture quaternion differential equation, a kinematic differential equation of the angular velocity of the object, and an error model equation of a three-axis magnetic field sensor. A second attitude and angular velocity filter having an included equation of state;
The system of claim 18. The second attitude and angular velocity filter uses a measurement value of the magnetic field from the triaxial magnetic field sensor and a measurement value from the triaxial acceleration sensor as a measurement vector of the filter, and uses a measurement model of the filter. The equation uses the transformation equation of the magnetic field vector and acceleration vector from the local navigation contact reference system to the main body reference coordinate system,
The magnetic field vector in the local navigation tangent reference system is known;
27. The system of claim 26. A Global Positioning System (GPS) receiver adapted to determine the position of the object,
The location includes longitude, latitude and altitude.
The system of claim 18. Means for providing a magnetic field vector in a local navigation reference system using the longitude, latitude and altitude in combination with a world magnetic model;
30. The system of claim 28.

Images