WO2019219665A1 - Acoustic distance measurement device and system - Google Patents

Abstract

The invention relates to a technique providing high precision in hybrid RF and audio system, while enabling ultra low power, by means of sending the ultrasonic waves ahead in anticipation of an RF-based wake-up signal. Herewith, the low power nodes are kept awake for a considerable time. The invention further relates to hybrid distance measurements approached from a different perspective in comparison with the art, and thus now being well suited for energy-constrained nodes. An RF signal is used for synchronisation, whereas the distance measurements are merely based on acoustic chirps.

Description

ACOUSTIC DISTANCE MEASUREMENT DEVICE AND SYSTEM

Technical field

The invention relates to a technique providing high precision in hybrid RF and audio system, while enabling ultra low power. In particular, the invention relates to hybrid (signalling for) distance measurements being approached from a different perspective compared to the art, and being suitable for energy-constrained nodes.

Background of the invention

The last couple of years, the need for positioning functionality in many emerging low power nodes, sensor networks, and Internet-of-Things (loT) applications grew significantly. A main challenge herewith is to offer a high reliability and precision within the (very) constraint energy budget of battery powered devices. The usage of sound in accurate distance measurements has certain benefits over electromagnetic waves, used in visual, radio frequency (RF) and optical techniques. In these low power applications, the main asset is improved accuracy obtained at lower system frequency due to the lower propagation speed of mechanical waves. Moreover since the introduction of Micro-electromechanical Systems (MEMS), modules can nowadays be realized at low cost with a small form factor and can compete with other distance measurement systems on a power level.

Ultrasonic Positioning systems (UPS) most of the time use three techniques to obtain the time-of- flight (TOF) estimation. The simplest form is threshold detection. In this technique, the transmission and detection of the arriving acoustic wave is triggered by the event where a voltage level exceeds some predetermined threshold. The advantage of this technique is that it does not need any complex computations and can be achieved with simple circuitry. Flowever, as a measurement technique it is error-prone, being vulnerable to low signal-to-noise ratios, low sampling frequencies and inherent biases. The next TOF estimation technique is phase detection, in which the phase difference between the transmitted and the en7tire received sound signal is measured. Single- frequency based phase shift method cannot work practically, due to the limited distance mathematically implied by the maximum phase difference of 2/Pi. The limited range can be improved by using two or more frequencies in a transmitted signal. The last TOF technique, and used most in practice, is cross correlation. Here the transmitted and received sound signal are cross- correlated and produce a maximum value at the time delay. Since single tone cross correlation produce similar peaks at false delays, frequency-modulated signals are used. Linear chirps are a good option in these cases, since the inter-correlated signal is compressed, making it more easily to detect in lower signal-to-noise ratio (SNR) scenarios and improve the range resolution. Hybrid RF signalling techniques, combine the latter two, improving the accuracy and precision of the TOF method.

In hybrid RF/acoustic positioning two techniques can be distinguished. The first one is introduced by the so-called Active Bat Local Positioning System. Fixed beacons on the ceiling receive an ultrasonic signal coming from a mobile emitter. The position is then calculated by the system, which triggers the transmission of a new ultrasonic signal by the means of a radio link. The second technique is introduced by the so-called Cricket System. Here, the beacons send out an ultrasonic signal together with a radio pulse for synchronization. The mobile device calculates its own position, which ensures its privacy. Recent hybrid signalling studies mainly focus on obtaining high accuracy with as little infrastructure as needed. However, in the context of Wireless Sensor Networks (WSN), energy is the main constraint. There is little research done on positioning methods that strive for a low power solution, hence research is intensifying. The discussed systems perform some form of time-of-flight (TOF) or time difference of arrival (TDOA) method of the hybrid signals to calculate the distances.

By means of example, the fundamental bottleneck of current systems and techniques becomes clear in a commonly known beacon-based system with mobile nodes operating on a small battery.

In general, RF-based positioning systems consist of a higher-end, centralized fixed system that can be powered and can provide the necessary calculations. The lower-end of these systems consist of mobile nodes, limited in power, both on an energy and computational level. Synchronisation between the different mobile nodes is realised, for example throughout RF communication, whether structured by the centralized system or the nodes mutually. Indoor positioning technology exists comprising such a system and moreover could be based on a state-of-the-art sensor fusion approach that combines RF RSSI data, motion sensor data and floor plan information to calculate the most likely position, using a particle filter. This technology has a positioning error of less than lm in 80% of the cases. Although this performance is adequate in most applications, there are several applications the system fails to address because better accuracy and/or precision is required. Adapting the above RF-system with ultrasonic technology such that a hybrid form of positioning is obtained, promises to augment the ranging accuracy significantly.

A positioning system overview in accordance with the art is depicted in Figure 1. There are four types of units in this infrastructure: a cloud server, a base station, beacons and tags. At the bottom of this hierarchic structure are the tags. These advertise packets to the base stations to set up a connection and multicast their on-board inertial sensor data. Next are the beacons, picking up the data that is multicasted by the tags and measuring the RSSI values of the tag-multicasts. This information is then multicasted to the base station. The base station collects the data from the tags and beacons in its range, which in turn transmits all the info to the cloud server over Ethernet or another potentially wireless communication technology. At the top of this hierarchic structure, the cloud server calculates the position of the moving tags. The system enables indirect remote positioning and self-positioning of smartphones, which requires less infrastructure. In other words, the environment determines the position of the node. We envisioned that there are also situations e.g. referring to self-positioning wherein the node itself can determine the position and transfers it to the beacons.

Position updates are performed at regular intervals for example every second. To prolong the battery life, the tags wake up for only a fraction of the time for example 1 millisecond during this second, transmitting the necessary data over RF. For acoustic ranging, this 1 millisecond is a short awake window. In other words, the issue is not a matter of complexity but the fact that the time is too short for acoustic waves or signals to arrive, i.e. acoustic signals are (too) long propagating or traveling en route. A solution can be found in mobile active architecture in which the tag actively transmits the sound signal. This results in the straightforward scheme at the mobile nodes in which it simply wakes up and transmits a sound during that millisecond. Inter-node synchronisation is not required in this case, since the system knows at what time the node is awake based on the much faster RF signal. Random wake-up times perform some form of Time Division Multiplexing (TDM) enabling some nodes to be active at the same time in the same environment. However, on an energy level, this is not the best solution. From a hardware perspective, sending out sound waves that can travel larger distances costs more energy than receiving them. Besides, beacons should be actively listening to and processing the acoustic signals coming from several mobile nodes. When in the same room their number grows too large, the TDM is not sufficient and more complex spread spectrum or multiple access algorithms should be used. Therefore, a low-power answer to this problem is needed on mobile passive systems, where the beacons send out acoustic signals. It is further desirable to impose limiting the exposure to ultrasonic and sonic noises to avoid potential damaging of the auditory system.

In mobile passive systems, the wireless channel is not dependent on the number of tags. The lower in number beacons send out the sound signal. The beacons have a higher energy budget and the tags only have to receive the signal, potentially solving the energy problem at the tags. As a result, the multiplexing problem in this scenario is simpler. Yet, the problem still lies in the 1-millisecond awake time and the acoustic timing paired with it. As it is preferred to impose limiting the exposure to ultrasonic and sonic noises and herewith avoiding potential damaging of the auditory system, the beacons should not send sound signals permanently. The ultrasonic sounds or noises can be very silent. As the sound level of ambient ultrasonic noise is subdued, the broadcasted ultrasonic signals can be very silent. It is noted that the mobile nodes should sleep as long as possible, such that they can save as much energy as possible. An obvious solution is letting the mobile node wake up upon a specific acoustic event coming from the beacons. There are two methods to perform this: the first is based on the sound level of an acoustic signal, whereas the second one is a tone-specific wake-up system. The PMM-3738-VM1010-R is a commercially available MEMS microphone that alerts a processor of an acoustic event when a sound is above a level threshold in which the MEMS microphone is switched to its normal mode, enabling it to record the sound signal. In spite of the low power consumption while being active in 'Wake on sound' mode (5mA), this method has the disadvantage of a variable wake up time (< 100ms), introducing large distance measurement errors. In addition, the processor would have to wait too long for the microphone to become active, consuming too much power in a state waiting on audio. Another disadvantage is that the wake-up only depends on the sound level. Moreover, false detections generated by sound signals coming from other sources than the beacons, could activate the processor, waking it up and consuming unnecessary power.

In answer to this latter problem tone specific wake-up systems can be referred to. As an example, a full custom CMOS ASIC can be designed to detect low-frequency periodicity in acoustic signals, a characteristic feature of sound generated by vehicle engines. However, for this particular example, the microphone should be awake the whole time, consuming too much energy. This can be partially solved by using a pulse width modulation scheme to turn the components on and off. Yet, the larger wake-up times of the MEMS microphones make this an inefficient method for a 1-millisecond awake state and still some energy is lost due to the idle on state of the microphone.

Further referring to the art, particular reference is made to the following patent applications. Reference US2006/013070 A1 from HOLM et al discloses a system with ultrasound transmission on the identification tag side, making it power hungry.

Reference US2006/077759 Al from HOLM discloses an alternative system with ultrasound reception on the identification tag, which is also power hungry, while RF transmission is also foreseen but for the sole purpose of tag ID code transmission.

Similar to the first reference from HOLM, a further US2011/026363 Al from LAVACHE et al discloses a system with ultrasound transmission on the mobile component side, making it power hungry. A notion of turning off the transducers is provided for the sole purpose to establish a period of silence. Similar to the first reference from HOLM and third reference from LAVACHE et al, US2014/160880 Al from KING et al describes a mobile device with both a ultrasound transducer and ultrasound receiver, making this extremely power hungry, while RF transmission is also foreseen for the sole purpose to synchronize the time or clocks of the communicating devices to thereby determine time of flight based on difference between transmit and reception time.

Aim of the invention

The aim of the invention is to provide a high precision and accurate positioning system for a low power nodes environment. Summary of the invention

In a first aspect of the invention an arrangement is provided, for determining an estimate of the distance between a first device and a second device based on sound or ultrasound waves. This estimate is not an absolute measure estimate. In fact, the distance estimate is determined based on time, which is easier to be measured and can be defined with higher accuracy when the wave or signal is traveling or propagating (much) slower. The arrangement comprises this first device and a system, comprising itself of a plurality of devices, amongst which the second device is one of these. According to an embodiment, the first device is for instance a tag, the second device is (part of) a beacon, and the system is a beacon system.

The first device comprises a sound or ultrasound wave-receiving component, adapted for receiving the (ultra-)sound waves. More in general, acoustic waves (sound or ultrasound) and corresponding acoustic (sound or ultrasound) wave-receiving components are referred. As an example of an acoustic wave or signal, a chirp signal can be provided wherein the frequency is varying in function of time. The first device also comprises an RF wave-receiving component specific to the first device, and hence this component is also called a first RF wave-receiving component or a device RF wave receiving component. Such component is adapted for receiving an RF wave, more in particular a so- called first RF wave, and providing a signal, e.g. an electrical signal or a logical signal, to the (ultra sound wave-receiving component to activate it, possibly temporarily meaning that the activation is from time to time, can be even for a considerable amount of time though not continuously. The first device further comprises a computing component, adapted for computing a signal, e.g. an electrical signal or a logical signal, carrying information of the distance, based on another signal obtained from the (ultra-)sound wave-receiving component. Moreover, the first device comprises an RF wave- transmitting component specific to the first device, and hence this component is also called a first RF wave-transmitting component or a device RF wave-transmitting component. Such component is adapted for transmitting the computed - possibly electrical or logical - signal as a second RF wave. According to an embodiment, the computing component of the first device is further adapted for performing the computation based on the signal from the (ultra-)sound wave-receiving component, being only active for a predetermined time duration. The computation may be based on cross correlation. According to an embodiment, the first device's computing component is further adapted for determining the estimate of the distance itself.

For sake of clarity the arrangement facilitates a method for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), the method comprising the steps of: (i) transmitting said (ultra-)sound waves (200) by said second device (20); (ii) (thereafter) transmitting a first RF wave (210) by said second device (20) based on a signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200); (iii) receiving said first RF wave (210) and (temporarily) activating an (ultra-)sound wave receiving component (30) within said first device (10); (iv) receiving said (ultra-)sound waves (200) by said (ultra-)sound wave receiving component (30). Therefore while preferentially (to avoid false detection) the transmitting a first RF wave (210) occurs after the (start of the) transmission of said (ultra-)sound waves (200), given the function of the first RF wave (210) to (temporarily) activating an (ultra-)sound wave receiving component (30), said transmitting a first RF wave (210) has to occur before the end of the (ultra-)sound waves (200) as received.

Preferably the first RF wave is sent at a time which is within the last 10% of the duration of the sound wave transmission. For example, if the transmission of (ultra)-sound waves by the second device lasts from a time t=0 to t=ti, then preferably the first RF wave is transmitted by the second device at a time which is between t=0.9ti and t=ti.

Within the invention one or more calculations or computations (such as but not limited to cross correlations) are or can be performed in relation to the determining said estimates of distance and based thereon positioning of devices. These one or more calculations or computations are or can be and are typically distributed over a plurality of places, such as the first device, the second device and any other device, such as base stations and/or connected cloud environments within the described arrangements. These calculations or computations can be hard coded within the used circuits and/or provided by computing components like programmable processors, FPGA's and the like. The system of the arrangement comprises the second device, a third device, a fourth device and a fifth device. The second device comprises an (ultra-)sound wave-transmitting component, adapted for transmitting the (ultra-)sound waves, whereas the second device itself is adapted for providing a signal, e.g. an electrical signal or a logical signal, to the third device, wherein this signal being indicative of the time of transmitting the (ultra-)sound waves. The third device comprises an RF wave-transmitting component specific to the system, and hence this component is also called a second RF wave-transmitting component or a system RF wave-transmitting component. Such component is adapted for transmitting the first RF wave, based on the signal as provided by the second device, and being received by the third device. The fourth device comprises an RF wave receiving component specific to the system, and hence this component is also called a second RF wave-receiving component or a system RF wave-receiving component. Such component is adapted for receiving the second RF wave. The fourth device is adapted for providing a signal, e.g. an electrical signal or a logical signal, to the fifth device. The fifth device comprises a computing component, adapted for determining - either directly or indirectly - the estimate of the distance based on the signal as provided by the fourth device, and being received by the fifth device, whereas this signal is carrying information obtained from the RF wave-receiving component specific to the system.

According to an embodiment, the arrangement comprises a plurality of first devices, being for example tags. In an embodiment, the second and third devices are combined as one single device, or even further combined with the fourth device as another single device. Flence, two, or respectively three devices are integrated in one single device. According to an embodiment, the arrangement comprises a plurality of such single devices being a combination of either two or three other devices, wherein such single devices are for instance beacons. The single devices can also be autonomous or independent, meaning that they function or operate on the energy or power received from a battery provided there within.

According to an embodiment, the arrangement is adapted for determining a position in an environment of one or more of the first devices, whereas the environment is provided with one or more of the second devices. The position determining can be computed or calculated from the obtained estimates of distance, of the one or more first devices relative to the second devices. In a second aspect of the invention a first device is provided suitable for use in an arrangement, for determining an estimate of the distance between this first device and a second device based on sound, ultrasound or acoustic waves. Again, such estimate is not an absolute measure estimate, but is determined based on time, which is easier to be measured and can be defined with higher accuracy when the wave (such as e.g. an acoustic wave in comparison with an RF wave) or signal is traveling or propagating (much) slower. According to an embodiment, the first device is for instance a tag, and the second device is (part of) a beacon.

The first device can be considered autonomous or independent, meaning that it can operate stand alone by means of a battery. While referring to tags as possible first devices, active tags are particularly to be mentioned as being autonomous or independent. Hence, according to an embodiment, the first comprises an energy storage device, such as a battery, suited for providing energy to the first device's components, e.g. wave-transmitting or wave-receiving components, or a computing component respectively. Also here, the first device comprises a sound or ultrasound wave-receiving component, adapted for receiving the (ultra-)sound waves. More in general, sound or acoustic waves and corresponding sound or acoustic wave-receiving components are referred. The first device also comprises an RF wave-receiving component specific to the first device, and hence this component is also called a first RF wave-receiving component or a device RF wave receiving component. Such component is adapted for receiving an RF wave, more in particular a so- called first RF wave, and providing a signal, e.g. an electrical signal or a logical signal, to the (ultra sound wave-receiving component to activate it, possibly temporarily meaning that the activation is from time to time, can be even for a considerable amount of time though not continuously. The first device further comprises a computing component, such as for example a processor, adapted for computing a signal, e.g. an electrical signal or a logical signal, carrying information of the distance, based on another signal obtained from the (ultra-)sound wave-receiving component. Moreover, the first device comprises an RF wave-transmitting component specific to the first device, and hence this component is also called a first RF wave-transmitting component or a device RF wave- transmitting component. Such component is adapted for transmitting the computed - possibly electrical or logical - signal as a second RF wave. According to an embodiment, the device RF wave receiving component is further adapted for providing a signal either to the computing component, and/or to the RF wave-transmitting component respectively, to possibly temporarily activate one of these components, or else activate both all or not from time to time, all or not in an alternating way or simultaneously.

The second device comprises an (ultra-)sound wave-transmitting component, adapted for transmitting the (ultra-)sound waves, whereas the second device itself is adapted for providing a signal, e.g. an electrical signal or a logical signal, to the third device, wherein this signal being indicative of the time of transmitting the (ultra-)sound waves.

In a third aspect of the invention, a single device is provided suitable for use in an arrangement, for determining an estimate of the distance between a first device and a second device based on sound, ultrasound or acoustic waves, wherein this single device comprises such second device, comprising an (ultra-)sound wave transmitting component, adapted for transmitting (ultra-)sound waves. The second device itself is adapted for providing a signal to a third device, wherein this signal being indicative of the time of transmitting (ultra-)sound waves. The third device mentioned comprises a RF wave-transmitting component, adapted for transmitting an RF wave, based on the signal received from the second device.

In a fourth aspect of the invention, another single device is provided comprising the single device in accordance with the third aspect, and further comprising a fourth device. The fourth device comprises an RF wave-receiving component adapted for receiving an RF wave. The fourth device is adapted for providing a signal, e.g. an electrical signal or a logical signal, to a fifth device. The fifth device is part of the arrangement, although not being part of the single device in accordance with the fourth aspect. Such fifth device is adapted for determining the estimate of the distance based on the signal received from the fourth device.

In a fifth aspect of the invention an arrangement is provided, for determining an estimate of the distance between a first device and a second device based on (ultra-)sound waves. The arrangement comprises a first device and a system, itself comprising a second, third, fourth and fifth device. The first device is adapted for receiving a first RF wave, receiving (ultra-)sound waves upon receiving the first RF wave, computing first type of information of or related to the distance, based on the (ultra- )sound waves, and transmitting the first type of information as a second RF wave. The (ultra-)sound waves will arrive at first device later than the first RF wave, whereas RF waves are traveling or propagating faster than (ultra-)sound or acoustic waves. The second device is adapted for transmitting the (ultra-)sound waves, providing second type of information being indicative of the time of transmitting said (ultra-)sound waves. The third device is adapted for transmitting the first RF wave, based on the second type of information being indicative of the time of transmitting said (ultra-)sound waves, as received from the second device. The fourth device is adapted for receiving the second RF wave, providing third type of information to a fifth device, being adapted for determining the estimate of the distance based on the provided third type of information thereto.

In a sixth aspect of the invention, an arrangement is provided for facilitating the determining of an estimate of the distance between a first device and a second device based on (ultra-)sound waves. The arrangement comprises a first device and a system, itself comprising a second, and a third device. The first device comprises an (ultra-)sound wave-receiving component, adapted for receiving (ultra-)sound waves. The first device further comprises a device RF wave-receiving component, adapted for receiving a first RF wave, and for providing a signal to the (ultra-)sound wave-receiving component to (temporarily) activate it. The first device also comprises a computing component, adapted for computing a signal, carrying information of the distance, based on a signal obtained from the (ultra-)sound wave-receiving component. The computing component is further capable to decide to take an action (by providing an action signal), based on the information carried. The system's second device comprises an (ultra-)sound wave-transmitting component, adapted for transmitting (ultra-)sound waves. The second device is adapted for providing a signal to a third device, wherein this signal is indicative of the time of transmitting (ultra-)sound waves. The third device comprises a system RF wave-transmitting component, adapted for transmitting the first RF wave, based on the signal. According to an embodiment, the carried information is the estimate of the distance itself and the action is defined as taking a security measure based thereon.

Brief description of the drawings

Figure 1 illustrates an embodiment of a setup used in a localisation system according to the art. Figure 2a illustrates an embodiment of a single beacon setup for periodic, synchronized audio broadcast in accordance with the invention. Figure 2b illustrates an embodiment of a ranging system setup, wherein the beacon periodically transmits a sound signal, and all ultra low-power nodes are woken up and synchronized on basis of the RF signal, in accordance with the invention.

Figure 3a illustrates an embodiment of a timing diagram of a single period of the audio broadcast in accordance with the invention. A chirp signal as a potential waveform is displayed at the transmitter and three mobile nodes. Figure 3b illustrates an embodiment of timing overview of the transmitter and three mobile nodes, wherein Mobile Node 1 and 2 are within the range of the beacon, whereas Mobile Node 3 is not, and the distances to the transmitter are calculated based on the received sound chirp, in accordance with the invention.

Figure 4 shows an embodiment of cross correlation results in the case of different awake times: 1 ps, 10 ps, 100 ps and 1ms in accordance with the invention.

Figure 5 shows an embodiment of a room overview in accordance with the invention. The green centrepiece is the sound source. Microphone 1 can be found on the lower left corner, microphone 6 is located left from the sound source.

Figure 6 shows an embodiment of the pre-processing, Monte Carlo simulations and post-processing in accordance with the invention. The actual Monte Carlo simulations are repeated 10 000 times. Figure 7 illustrates an embodiment of microphone 6 normal and Epanechnikov kernel distributions for a bandwidth of 30 kHz and 60 kHz at a SNR of 3dB, in accordance with the invention.

Figure 8 illustrates an embodiment of microphone 1 normal and Epanechnikov kernel distributions for a bandwidth of 10 kHz, 30 kHz and 60 kHz at a SNR of 6dB, in accordance with the invention. Figure 9 illustrates an embodiment of off-centred Monte Carlo simulations for four microphones in a high reverberant room, in accordance with the invention.

Figure 10 illustrates an embodiment of off-centred Monte Carlo simulations for four microphones in a low reverberant room, in accordance with the invention.

Figure 11-16 illustrate schematically embodiments of an arrangement, its devices and/or components in accordance with the invention.

Figure 17 illustrates an embodiment comparing the Peak Prominence Method with the Maximum Method, wherein the right distance peak is the first of a series of local maxima and being correctly appointed to by the Peak Prominence Method, in accordance with the invention. Detailed description of the invention

The invention provides a technique that realizes the high precision in hybrid RF and audio system, while enabling ultra low power. This is achieved through the concept of sending the ultrasonic waves ahead in anticipation of an RF-based wake-up signal. This concept allows resolution of the need of keeping nodes awake for a considerable time. Moreover, with the invention hybrid distance measurements can be approached from a different perspective, whereas current proposed systems in the art are not well suited for energy-constrained nodes. An RF signal is used for synchronisation, e.g. in providing a synchronised wake up, yet the distance measurements are merely based on acoustic signals having some parameter changing as a function of time. Flence, the time can be derived or estimated that the part of the signal being visible was on the way. As an example, a chirp signal can be used wherein the frequency is acting as the parameter changing in function of time

The invention builds on the insight that combining (i) a slow travelling wave (like but not limited to a (ultra-)sound wave or other type of pressure wave), which is suitable for accurate distance (position) determination with (ii) a fast travelling wave (like but not limited to a RF wave or even also a (ultra-)sound wave or other type of pressure wave if velocity dependence of the used media can be used) to awake devices in an appropriate manner for enabling capturing said slow travelling wave result in either energy efficient and/or security by design type of solutions, especially if such devices are capable to retrieve information in short time windows and/or capable to decide on actions, taking into account distance information. Given the above, preferably as most energy efficiently the fast travelling wave is transmitted after the slow travelling wave.

ULTRASONIC RANGING BASED ON PERIODIC SYNCHRONISED AUDIO SIGNALS

The invention relates to a progressive ultrasound broadcast solution.

Since there still needs to be research done for fast, reliable auditory wake up, another passive mobile system solution should be designed. An alternative to this problem can be obtained by periodically sending out sound signals at the beacons. To counter the imposed restricted audio broadcast and to guarantee that every node receives an audio signal, the nodes should be synchronised. Figure 2a shows the system setup in the case of a single beacon audio transmission. The starting time of the audio broadcast at the beacon (B) is registered by the system (To). Depending on the distance to the beacons, delayed versions of the sound signal are perceived. An example for three mobile nodes (Mi, /W2 and M_z) can be seen in Figure 3a. At a certain point in time (T_A) an RF signal is transmitted from the beacons, synchronously waking the mobile nodes and making them active for 1 millisecond. During this time, a part of the perceived audio is recorded, sampled and used for data transfer (indirect positioning) or for further local processing (self positioning).

A linear chirp is used as audio broadcast for two reasons. First, depending on the distances, other frequencies are received by the mobile nodes. This enables to perform distance measurements purely on the received frequency shift. Second, in processing the received signals, chirps have compressed inter-correlation signals and remain very well correlated over Doppler shifts. The gain in SNR and the improvement on the range resolutions depend on the compression ratio of the chirp, which is equal to its time bandwidth product. The power of the compressed correlation peak depends on the chirp bandwidth d/and the pulse duration t.

Another limitation of the pulse duration (r) is the maximum distance between a mobile node and a beacon, e.g. a 30ms signal limits this distance to 10.2m in an environment with a sound velocity of v_s = 340— . The key here is to wake the mobile nodes as close as possible to the end of the sound transmission. M_z in Figure 3a is too far away from the beacon to receive the sound signal (— > t)

V_S

and therefore incapable of calculating the distance to the beacon. This feature is an advantage in larger rooms with a higher number of beacons in it to cover the whole area. The mobile nodes will only receive the audio signal from neighbouring beacons, making positioning a lot simpler.

When comparing this latter solution to the other proposed systems, it has some clear advantages towards energy consumption at the mobile node side. As this is a main concern when exploiting the problem from an industrial point of view, this system will be further tested in a simulation environment.

The proposed hybrid RF/ultrasonic ranging system concept is also shown in Figure 2b. In its generic form, it consists of a single beacon (B) and one or more mobile nodes (M_x). A single RF signal simultaneously wakes up all the mobile nodes. A distance measurement is performed as follows: a beacon starts broadcasting an audio signal with a certain duration (r_tx) at starting time (To). At a given time (T_A), all mobile nodes wake up simultaneously for a short time (t_GC) and receive, depending on their distance to the beacon, a specific part of the delayed, distorted audio broadcast.

A timing diagram for three mobile nodes is also depicted in Figure 3b. The coverage range of the system is determined by the audio broadcast duration at the transmit side (¾) and the speed of sound, e.g. for a sound signal of 30ms and a speed of sound of v_s = 340 m/s, the maximum measurable distance between a mobile node and a beacon is 10.20 m. To achieve this coverage, the receivers should wake up as late as possible. T_A is determined based on the three timing parameters: T_A = (To + T_tx) - t_GC. M_z in Figure 2b, is too far away from the beacon to receive the sound signal when it wakes up since (—> T_A - To). It is therefore incapable of calculating its distance to the beacon.

V_S

This feature can be an advantage in larger rooms as the mobile nodes only receive the audio signal from neighbouring beacons, restricting the possible positioning areas.

The timing overview of Figure 3b illustrates that the proposed ranging concept differs from conventional hybrid RF/acoustic TDOA systems. As all mobile nodes wake up at the same time, the distance measurement data is comprised of the received audio signal (D/i for Mi, D/₂ for M2) at the wake-up time of the mobile nodes (T_A in all cases). This approach opens the opportunity to operate at significant less energy consumption due to the restricted awake time of the mobile node. It impacts, however, the accuracy of the measurements, as it limits the amount of received data available for the positioning purposes.

The sampled audio in the awake state can be used for local processing (self positioning) or can be transmitted to a central unit (indirect positioning). Two-dimensional positioning of the mobile node can be achieved by adding at least two more beacons to the system in Figure 2b. The acquired, relative distances to these beacons can be used in multilateration or other geometric models to find the position of the mobile nodes. Identification of the sound signal is crucial here, and multiple access techniques (FDMA, TDMA, CDMA, etc.) as known in the art, could give a solution to this problem.

The invention also relates to pulse compression and the use chirp. Opposite to the advantages on an energy level, the small awake time limits the sampled data, making (local or global) processing harder. The main question arises whether this 1 millisecond contains enough data to further process the recorded acoustic signal or sequence

A method that can be used to perform fast, low processing distance calculations with small data is cross correlation. In linear chirps, using cross correlation or self-correlation is a form of pulse compression. Such a linear chirp can be described as: if 0 < t < t (Eq.l)

otherwise

Where r is the pulse duration, A the amplitude as a rectangle function, /o is the carrier frequency and D/is the chirp bandwidth. The linear ramp can be seen in the instantaneous frequency equation:

Flere <p(f) is the phase of the chirped signal. The cross correlation between the transmitted and received signal can be achieved by convolving the received signal with the conjugated and time- reversed transmitted signal:

(s_c, s_c)(t) = f _¥ s ^* (— t') s_c(t - t') dt' (Eq.3)

Taking into account the autocorrelation function:

(s_c, ⁼ A²i A sine [Dί t L Q i .Tcf_Q t

sc) ( (Eq.4) With A a triangle function, with a value of O on l ⁰⁰ > 2]

°°] and linearly increases on £ 2’ ¾ where it has its maximum 1, and then decreases linearly on [0> !L Around the maximum, this function

behaves like a cardinal sine, with a -3dB width of around ~ D/. For common values of D/, is smaller than r, hence the name pulse compression.

The minimal range resolution can be computed as: c

P = ( Eq.5)

2 Af

With c the propagation speed of the medium. E.g. a 10 kHz signal has a range resolution of 1.7 cm. The pulse compression ratio is a measure how compressed the inter-correlated signal is in comparison to the original transmitted signal and can be found with the following equation:

t

t D/ (Eq.6)

The pulse compression ratio is generally greater than 1 and lies most of the times between 20 and 30.

SIMULATIONS

In the proposed system, only a fraction of the transmitted signal is cross-correlated with itself. The received sound snippets have similar rate of frequency change as the original signals, but since the awake time is smaller, the frequency bandwidth will lower as well. Since the bandwidth of the inter- correlated signal is inversely proportional to the frequency bandwidth, the sync function will be spread out, making it harder to determine the maximum value. In noisy environments, this influences the SNR in a negative way.

Inter-correlation simulations were performed to see whether the compression ratio is still small enough with a 1-millisecond interval. Based on the frequency response of ultrasonic MEMS, a linear chirp with a centre frequency /o = 45 kHz and a bandwidth of A/ = 30 kHz was chosen. The pulse duration r was set to 30ms, corresponding to a maximum distance of 10.2m. In these simulations, a certain delay was added to the original signal, imitating the propagation time, after which a rectangular filter was used to simulate the wake up time of the mobile node. Figure 4 shows the simulation results of four different wake up times, i.e. 1 ps, 10 ps, 100 ps and 1ms. Table I shows the corresponding chirp bandwidth, inter-correlation bandwidth, range resolution and compression ratio. It can be concluded both visually as numerically that the 1-millisecond awake time is large enough for sufficient pulse compression. The expected sync function, with small inter-correlation bandwidth can be seen as enabling to calculate the distance-improved precision.

INTERCORRELATION

BANDWITH, RANGE RESOLUTIONS AND COMPRESSION RATIO.

Awake Time 1 mus 10 mu s 100 mu s 1 ms

Frequency bandwith D/ (Hz) 1 10 100 1000

Intercorrelation bandwith t' (s) 1 0,1 0,01 0,001

Range resolution p (m) 170 17 1,7 0,17

Pulse compression ratio p le-06 0,0001 0,01 1

Based on these results, the next step is to test whether the excellent inter-correlation performances still stand when noise and reverberations are added.

A. Simulation Environment

Further simulations were performed in the Roomsim "Shoebox" simulation room. This MATLAB- based program simulates the geometrical acoustics of a perfect rectangular parallelepiped room volume using the image source model to produce an impulse response from each omnidirectional primary source to a directional receiver (single sensor, a sensor pair or simulated head). These impulse responses are then convolved with an audio file to produce the perceived acoustic signal at the sensor. Different absorption coefficients can be selected for each of the six surfaces, resembling different materials. By means of example, the sound source is positioned centrally in a 6 by 4m room as depicted in Figure 5. Twelve single sensor omnidirectional receivers are evenly distributed in the room, with a 30cm minimum distance to a wall. The cross correlation between the fully received linear chirp and the broadcasted signal is again the room impulse response. B. Monte Carlo Simulations

Monte Carlo simulations are performed to test the cross correlation efficiency in noisy and reverberant rooms. The absorption coefficients in Table II represent the three Monte Carlo simulation environments, i.e. a room with high, normal and low reverberation time. The simulation process is depicted in Figure 6. The first step is to generate the sound signal that is broadcasted at the transmitter. With this, several sound signals can be tested, such as the chirp bandwidth, the pulse duration and carrier frequency. Next, this signal is simulated in the Roomsim Shoebox model. The room impulse responses at the different receivers are calculated and in combination with the original sound signal, the received sound signal is generated. Step 3 is the first part of the Monte Carlo simulation: noise addition. White noise with a certain SNR is added to the received sound signals. The second part is applying a 1-millisecond rectangular window to the noisy signal, mimicking the wake up time of the mobile nodes. Next a cross correlation between the emitted sound signal and the truncated, received sound signal is performed. The last part of the Monte Carlo simulation is finding the maximum value, out of which the distance is calculated. These last steps are repeated for 10 000 times. The post-processing consists of fitting the acquired distances to a normal distribution and an Epanechnikov Kernel distribution. This latter was chosen next to the normal distribution due to its optimal performance in a mean square error sense. It shows a more optimally smoothed kernel density estimate in the case of non-Gaussian distributions.

TABLE II

SIMULATIONS ABSORPTION COEFFICIENTS OF THE THREE ROOMS .

Frequency (Hz) 125 250 500 100 2000 4000

High reverberent 0,1 0,05 0,06 0,07 0,09 0,08

Normal reverberen 0,5 0,7 0,6 0,7 0,7 0,5

Low reverberent 0,9 0,9 0,9 0,9 0,9 0,9

-

As generally known, matched filtering is best suited for detection in an Additive White Gaussian (AWG) noise. However, due to the frequency dependency of the absorption coefficients, the noise gets coloured. To test the influences of this noise, AWG was added to the linear chirps with different SNRs in a room with a high reverberation time. Different frequency bandwidths were used as well to test whether there is an influence on the systems accuracy and precision. Two microphones were initially tested: microphone 1, in the lower left corner, and microphone 6, next to the sound source. The real distances to the sound source are respectively 3.191m and 0.9m. Table III summarises the standard deviation and expectation for microphone 6. The mean values match the real distance and increasing the signal to noise ratio lowers the standard deviation, thus increasing the accuracy.

TABLE III

STANDARD DEVIATION AND MEAN VALUE FROM MONTE CARLO SIMULATIONS WITH MICROPHONE 6 WITH T = 30 ms, 10 000

ITERATIONS AND XQ = 0.9 m

SNR (dB) 20 10 6 3 1 0

30kHz - 60 kHz m 0,900306 0,90018 0,90118 0,90263 0,9028 0,90334

s 1,06476E-13 0,0040557 0,0099161 0,016329 0,022058 0,024395

30kHz - 60 kHz m 0,90035 0,90057 0,90066 0,90055 0,9007 0,9008

- 30 kHz s 0,00042272 0,0023424 0,0035029 0,0047411 0,0062396 0,0069358

40 kHz - 50 kHz m 0,90747 0,9089 0,90715 0,90772 0,91043 0,90658

s 0,0077837 0,022269 0,036666 0,05218 0,063986 0,070085

40kHz - 60kHz m 0,900299 0,89794 0,89537 0,89702 0,89802 0,89544

s 0.000155078 0,013072 0,019046 0,025322 0,031761 0,031807

The bandwidth dependency performs as expected. Increasing the bandwidth improves the accuracy and precision. As the microphone response limits the frequency bandwidth of the linear chirp, a triangle-function can double this, improving the accuracy even more. It can be concluded that the proposed system has sub-centimetre accuracy for distance calculations near the sound source. Results for a 3dB SNR with a 30 kHz and 60 kHz bandwidth signal are depicted in Figure 7. Oversmoothening in the Epanechnikov kernel distribution show the side lobes, which are respectively 11mm and 5mm away from the centre, and being conform to the wavelength:

0.01133m

z

The results of microphone 1 tell a different story (Table IV). Increasing the SNR still has a positive influence on the system's precision, however the accuracy is not optimal at maximum SNR. An extended table for a 30 kHz bandwidth signal can be found in Table V. At higher SNRs, the mean value and quantiles are situated around a larger distances than the actual span. A visual examination of the Epanechnikov kernel distributions with a 6dB SNR, shows a larger peak around this distance. Increasing the SNR even more, rises the amplitude of the secondary peak even more, sometimes fully eliminating the primary peak which leads to low standard deviations (and thus higher precision) but lower accuracy. Since the microphone is positioned in the corner of the room, this secondary peak can be related to reverberations. Since the sound source is on a fixed position and the absorption coefficients are low, standing waves and constructive interference can occur, resulting in false maximum detections. Further simulations should check what the influence is of the symmetric set-up.

TABLE IV

S TANDARD DEVIATION AND MEAN VALUE FROM MONTE CARLO SIMULATIONS WITH MICROPHONE 1 WITH T = 30 ms, 10 000

ITERATIONS AND XQ = 3.191 m

SNR (dB) 20 10 6 3 1 0

30kHz - 60 kHz m 3,7683 3,6943 3,6381 3,4716 2,9509 2,5267

s 0,014976 0,17277 0,24728 0,95149 1,8991 2,3825

30kHz - 60 kHz m 2,9762 0,73209 0,41248 0,14226 0,026839 -0,22143

- 30 kHz s 2,5419 4,3528 4,4193 4,4143 4,2816 4,2137

40 kHz - 50 kHz m 3,1963 3,1995 3,197 3,0504 2,6119 2,2311

s 0,02481 0,047134 0,13354 0,96918 1,8325 2,282

40kHz - 60kHz m 3,4109 3,4956 3,552 3,2893 2,6378 2,2318

s 0,035174 0,22968 0,3748 1,3531 2,2903 2,6578

The same phenomena can be observed when decreasing the frequency bandwidth. Lower bandwidths show better accuracy and precision, opposite to what is expected. The 10 kHz bandwidth, Epanechnikov kernel distribution of the 10 kHz bandwidth linear chirp in Figure 8 shows a clear maximum peak at the expected distance (f_o is still 45 kHz). At this limited bandwidth, the SNR behaviour is the same as in microphone 6. More false peak detection may occur with increased bandwidth. When comparing the 60 kHz bandwidth signal (30khz to 60 kHz and back to 30 kHz) to the 30 kHz, it can be seen that the standard deviation evolves entirely different at every SNR given, and is considerably larger at 60 kHz. The large sigma value for this latter is quite consistent and can be visually explained in Figure 8. The kernel distribution shows that there is a negative peak, a result of the cross correlation of a triangle chirp. If the wake up time is too short, the frequency shift is not diverse enough, making it possible that the mirrored frequency in the triangle chirp produces a maximum that is larger than its original frequency. In other words, the wake up time can be selected to best suit the detection of the acoustic signal. The wake up time in these simulations is not chosen at the middle of the triangle function broadcast, causing the centre of the mirrored peaks shifted relative to the cross correlation centre. The best results are obtained when putting this awake time in the middle of this extended bandwidth chirp.

TABLE V

EXTENDED MEASUREMENTS ON A 30 KHZ BROADCAST BANDWITH IN

MICROPHONE 1 .

SNR (dB) m s Median Variance Qo,025 Qo,25 Qo,5 Qo,75 Qo,975

20 3,77 0,01 3,77 0,00 3,73 3,77 3,77 3,77 3,78

10 3,69 0,17 3,77 0,03 3,22 3,73 3,77 3,77 3,78

6 3,64 0,25 3,74 0,06 3,19 3,69 3,74 3,77 3,81

3 3,47 0,95 3,73 0,91 1,83 3,27 3,73 3,78 3,85

1 2,95 1,90 3,71 3,88 -4,26 3,21 3,71 3,77 4,09

0 2,53 2,38 3,64 5,76 -4,94 3,10 3,64 3,77 4,11

C. Off centred simulations The previous simulations were performed in an environment where there is perfect symmetry in respect to the room in both the receivers and the sound source, something that will rarely happen in a real environment. This can cause unwanted interference, having a negative impact on the simulation results. In the next simulations, the source is put off centre (3.2 x 2.2 m) in different reflective rooms. Tests were performed to measure the effect of the absorption coefficients and symmetry at the same time. An extra corner microphone and one next to a wall were added to the simulations. The real distances to microphone 1, 4, 6 and 8 are: 3.467m, 3.140m, 1.118m and 2.508m. Only a linear chirp with the frequency bandwidth of 30 kHz is used. Figure 9 demonstrates the off centred Monte Carlo Simulations of the four microphones in a highly reflective room with a SNR of lOdB. Microphone 1 behaves the same as before, having a good precision and accuracy. The corner microphones have low accuracy and higher precision, again indicating a larger distance than the real distance. The microphone close to the wall has a high accuracy, but lower precision. The side lobes are double the distance to the wall away from the centre lobe, indicating that these are caused by reflections. It can be concluded that positioning the sound source off centre has minimal influence on the results.

If the same simulations are performed in a room with normal absorptive coefficients, it can be seen that the amplitude of the secondary peak becomes less dominant than the primary peak, resulting in more precise results, but at lower accuracy. Comparable results are booked for the wall microphone. The SNR behaviour in both cases is the same as earlier described. Increasing the absorption coefficients shows even better results. The results in the room with the highest absorption coefficients and SNR ratio (20dB) are displayed in Figure 10. In all four microphones, the normal distribution shows a mean value the same as the real distance with a small standard deviation.

It can be concluded that a 1-millisecond sample of the received signal gives an inter-correlation bandwidth small enough for improved distance measurements. Both precision and accuracy are lowering when the SNR ratio is decreased. The periodic, synchronised broadcast method performs well in situations where there is almost no reverberation. However, mobile nodes positioned close to walls and in corners suffer from these reverberant signals, calculating erroneous, larger distances. Bandwidth limitations can be a solution in these situations. Addition of white Gaussian noise, at certain SNRs, decreases both precision and accuracy.

Referring to Figure 11, an arrangement 1000 is provided, for determining an estimate 400 of the distance 410 between a first device 10 and a second device 20 based on sound or ultrasound waves 200. This estimate 400 is not an absolute measure estimate. In fact, the distance estimate is determined based on time, which is easier to be measured and can be defined with higher accuracy when the wave or signal is traveling or propagating (much) slower. The arrangement 1000 comprises this first device 10 and a system 70, comprising itself of a plurality of devices 20, 90, 110, 130, amongst which the second device 20 is one of these. According to an embodiment, the first device 10 is for instance a tag, the second device 20 is (part of) a beacon, and the system 70 is a beacon system.

The first device 10 comprises a sound or ultrasound wave-receiving component 30, adapted for receiving the (ultra-)sound waves 200. More in general, acoustic waves (sound or ultrasound) and corresponding acoustic (sound or ultrasound) wave-receiving components are referred to. As an example of an acoustic wave or signal, a chirp signal can be provided wherein the frequency is varying in function of time. The first device 10 also comprises an RF wave-receiving component 40 specific to the first device 10, and hence this component is also called a first RF wave-receiving component 40 or a device RF wave-receiving component 40. Such component 40 is adapted for receiving an RF wave 210, more in particular a so-called first RF wave 210, and providing a signal 300, e.g. an electrical signal or a logical signal, to the (ultra-)sound wave-receiving component 30 to activate it, possibly temporarily meaning that the activation is from time to time, can be even for a considerable amount of time though not continuously. The first device 10 further comprises a computing component 50, adapted for computing a signal 310, e.g. an electrical signal or a logical signal, carrying information of the distance 410, based on another signal 320 obtained from the (ultra-)sound wave-receiving component 30. Moreover, the first device 10 comprises an RF wave- transmitting component 60 specific to the first device 10, and hence this component is also called a first RF wave-transmitting component 60 or a device RF wave-transmitting component 60. Such component 60 is adapted for transmitting the computed - possibly electrical or logical - signal 310 as a second RF wave 220. According to an embodiment, the computing component 50 of the first device 10 is further adapted for performing the computation based on the signal 320 from the (ultra sound wave-receiving component 30, being only active for a predetermined time duration. The computation may be based on cross correlation. According to an embodiment, the first device's computing component 50 is further adapted for determining the estimate 400 of the distance 410 itself. The system 70 of the arrangement 1000 comprises the second device 20, a third device 90, a fourth device 110 and a fifth device 130. The second device 20 comprises an (ultra-)sound wave- transmitting component 80, adapted for transmitting the (ultra-)sound waves 200, whereas the second device 20 itself is adapted for providing a signal 330, e.g. an electrical signal or a logical signal, to the third device 90, wherein this signal 330 being indicative of the time of transmitting the (ultra-)sound waves 200. The third device 90 comprises an RF wave-transmitting component 100 specific to the system 70, and hence this component is also called a second RF wave-transmitting component 100 or a system RF wave-transmitting component 100. Such component 100 is adapted for transmitting the first RF wave 210, based on the signal 330 as provided by the second device 20, and being received by the third device 90. The fourth device 110 comprises an RF wave-receiving component 120 specific to the system 70, and hence this component is also called a second RF wave receiving component 120 or a system RF wave-receiving component 120. Such component 120 is adapted for receiving the second RF wave 220. The fourth device 110 is adapted for providing a signal 340, e.g. an electrical signal or a logical signal, to the fifth device 130. The fifth device 130 comprises a computing component 140, adapted for determining - either directly or indirectly - the estimate 400 of the distance 410 based on the signal 340 as provided by the fourth device 110, and being received by the fifth device 130, whereas this signal 340 is carrying information obtained from the RF wave-receiving component 120 specific to the system 70.

According to an embodiment, the arrangement 1000 comprises a plurality of first devices 10, being for example tags. In an embodiment, the second and third devices 20, 90 are combined as one single device 150 as illustrated in Figure 13, or even further combined with the fourth device 110 as another single device 160 as illustrated in Figure 14. Flence, two, or respectively three devices are integrated in one single device 150, 160. According to an embodiment, the arrangement 1000 comprises a plurality of such single devices 150, 160 being a combination of either two or three other devices, wherein such single devices 150, 160 are for instance beacons. The single devices 150, 160 can also be autonomous or independent, meaning that they function or operate on the energy or power received from a battery provided there within.

According to an embodiment, the arrangement 1000 is adapted for determining a position in an environment of one or more of the first devices 10, whereas the environment is provided with one or more of the second devices 20. The position determining can be computed or calculated from the obtained estimates 400 of distance 410, of the one or more first devices 10 relative to the second devices 20.

While referring to Figure 12, according to an embodiment, the device RF wave receiving component 40 of the first device 10 is further adapted for providing a signal 350, 360 either to the computing component 50, and/or to the RF wave-transmitting component 60 respectively, to possibly temporarily activate one of these components, or else activate both all or not from time to time, all or not in an alternating way or simultaneously. Energy 500 may be provided to each one of the first device's components 30, 40, 50, 60 by means of an energy storage device 170 such as a battery.

As shown in Figure 15, the invention also provides an arrangement 1000, for determining an estimate 400 of the distance 410 between a first device 10 and a second device 20 based on (ultra sound waves 200. The arrangement 1000 comprises a first device 10 and a system 70, itself comprising a second 20, third 90, fourth 110 and fifth device 130. The first device 10 is adapted for receiving a first RF wave 210, receiving (ultra-)sound waves 200 upon receiving the first RF wave 210, computing first type of information 600 of or related to the distance 410, based on the (ultra- Sound waves 200, and transmitting the first type of information 600 as a second RF wave 220. The (ultra-)sound waves 200 will arrive at first device 10 later than the first RF wave 210, whereas RF waves are traveling or propagating faster than (ultra-)sound or acoustic waves. The second device 20 is adapted for transmitting the (ultra-)sound waves 200, providing second type of information 610 being indicative of the time of transmitting said (ultra-)sound waves 200. The third device 90 is adapted for transmitting the first RF wave 210, based on the second type of information 610 being indicative of the time of transmitting said (ultra-)sound waves 200, as received from the second device 20. The fourth device 110 is adapted for receiving the second RF wave 220, providing third type of information 620 to a fifth device 130, being adapted for determining the estimate 400 of the distance 410 based on the provided third type of information 620 thereto.

The invention is also related to an arrangement 1000 for facilitating the determining of an estimate 400 of the distance 410 between a first device 10 and a second device 20 based on (ultra-)sound waves 200. As depicted in Figure 16, the first device 10 being part of the arrangement 1000 may comprise a computing component 50, adapted for computing a signal 310, carrying information of the distance 410, based on a signal 320 obtained from the (ultra-)sound wave-receiving component 30. The computing component 50 is further capable to decide to take an action (by providing an action signal 370), based on the information carried.

CONCLUSION

A solution for sound based distance measurements in energy-restricted systems can be a periodic and synchronised broadcast method. Monte Carlo simulations with this method show excellent results for received signals with high SNR in rooms with eminent absorption coefficients. The system's precision and accuracy are highly dependent on the detection of reverberant signals coming from nearby walls and objects. The cross correlation between the original signal and the received signals generate comparable results with earlier work using this algorithm. The received signal is a sampled, smaller part of the original signal (1ms) instead of the whole signal. It should be noted that this system performs well in situations with low reverberation. However, in real-world applications, reverberant signals will be present. Further focus on the relation between the bandwidth and the accuracy and precision in reverberant sensitive positions is also part of the invention. In addition, the invention covers filtering out the real distance peak instead of the falsely maximum peak caused by reflections when using triangle signals for improved bandwidths.

The invention setup as discussed above only uses one sound source, with additional white noise. For positioning purposes in a 2D environment, as well related to the invention, at least 3 sound sources should be used. This suggests that at the awake time, three signals are received at the same time. Hence, identifying which signal comes from which source is a main concern here. The multiple access techniques suitable for this problem and applicable chirp based techniques are likewise included with the invention.

Subsequent to the simulations being complete, at least in this basic mode of operation, a hardware implementation is provided with the invention, and herewith the system's behaviour is measured in the physical world. According to an embodiment of the invention, ultrasonic speakers may be set up that transmit periodically the linear chirp. At the receiver side, great care is recommended on the power consumption of the MEMS microphone, associated amplifier, and extra processing and sampling power. Real world restrictions like the processor's limited sampling frequency or the wake- up time of the MEMS microphone are considered with the invention, and effects thereof on the distance measurements.

Referring further to location-based services, and herewith secure indoor positioning, location verification techniques can be used to verify a user's position claim. While most focus on resistance against various attacks and in general only little attention is paid to the search of low power solutions, a simple low power location verification system can be provided. An ultrasound-RF time- of-flight indoor positioning technique can be applied to provide accurate information about a low power mobile node. Location verification is guaranteed by emitting a pseudo-noise code, which has to be reproduced by the mobile node to prove physical presence.

Connected devices are creating a smart world: smart homes, hospitals, and cities, smart cars, logistics and industries. The success of new applications employing wireless devices will heavily depend on the security of these embedded systems. The geographical position of persons and devices can be very relevant information in many applications requiring secure authentication over wireless links. The described invention can be applied to offer this position information for low power nodes at high precision. Thus, the invention provides (low power) security applications based on the described arrangement/methods. Generally speaking, the provided arrangement and/or methods according to the invention are suitable for use in either (low power) security applications such as secret key generation from relative localization (distance) information obtained using the arrangement and/or methods, and/or in so-called distance-bounding applications. Distance- bounding cryptographic protocols rely on the travel time of electromagnetic waves. In practice, the required ranging accuracy cannot be achieved with regular computing capabilities. Hybrid RF- acoustic signaling as proposed with the present invention offer significantly better precision/complexity and could hence be leveraged on to realize practical distance-bounding solutions. ENHANCED ACCURACY SOLUTIONS

Enhanced accuracy solutions are now described as further part of the invention. Section above showed that utilizing the maximum as a selection criteria for the distance measurements results in an adequate solution, but could be further optimized. In most cases, the correct maximum is the first received maximum of a series of local maxima, as the lower frequencies are not present yet due to later audio transmission. Three methods are now proposed to select this first local peak and enhance the system's accuracy: window functions, peak prominence and delta peak method. The sole condition of these methods is to keep the processing power as low as possible, as it is proportional to the energy consumption.

A. Method 1: Window Functions

In a first method, window functions are applied to the correlation results, in a way that the first peaks of the local maxima are increased relatively to the others. The most straightforward window function is a linearly decreasing function. The cumulative density function (CDF) can be plotted as a function of the distance error for example in case of the original, maximum method and e.g. in the case that four linear window functions are applied to the correlation data in the shoebox simulation environment with an absorption coefficient of (a = 0.3). The negative slope of such a CDF function cannot be larger than 1, as some correlation data will become negative and limit the maximum reachable distance. In general, applying a window function to the correlation results in an improved accuracy of the system. The lowest and maximum distance error values are obtained when the slope is -1. The logical, next step could be e.g. applying quadratic window functions of the type y = ±ax² + bx + c to check if it is better to give more or less weight to the early peaks. Out of the CDF plots according to simulations, it can be deduced that the positive quadratic function has an even better effect then the linear window function.

A faster initial decline which consequently increases the influence of the earlier peaks improves the accuracy. To test the limits of this fast, initial decline, exponential window functions can be added: y = a^x~b . A good measure of the decay in exponential function is the half-life time (To_.s). The exponential window should preferably not decay too fast, resulting in similar CDF as with a steep slope linear function (T = 1 ms). Choosing the half-life too large results in similar CDF plots as with the positive quadratic function. According to a simulation, the optimal exponential function is the one with a decay time of G0.5 = 3 ms, a tenth of the original broadcasted signal. When comparing the optimal exponential window function to the quadratic function, it becomes clear that low distance error value of this window performs better but the maximum distance errors are more profound. A choice between a more precise or more accurate system can be made here.

B. Method 2: Peak Prominence

The window function method gives significance to the earlier correlation data. This method works well for microphones close to the speaker but as the difference between the maxima decreases with larger distances, peaks earlier than the correct maximum are chosen, resulting in even larger errors. The second improved accuracy method aims to avoid this issue by searching the local maxima without modifying the pulse compression data. Prominence of a peak measures how much the peak stands out due to its intrinsic height and its location relative to other peaks. It can be calculated as follows: extend a horizontal line from a chosen peak to the left and the right until the line crosses a signal because there is a higher peak, or it reaches the left or right end of the signal. Find the minimum of the left and right interval. This point can be a valley or a signal endpoint. Calculate the prominence by taking the difference between the height of the peak and the higher of the two interval minima. A low, isolated peak can be more prominent than a higher member of a tall range, as shown with Figure 17. This technique is commonly used in topography, in which it represents the elevation of a mountain summit relative to the surrounding terrain, and serves as a criterion to define a separate peak.

The index used for distance calculations is for example selected by calculating the prominences of all correlation peaks, setting a prominence threshold, the peak prominence factor (PPF), and using the index of the first peak in the array of the prominences larger than the threshold. Both reflections and noise affect the prominence of signals. The influence of the reflections on the accuracy is minimal, as, in a line-of-sight scenario, the correlated maxima of these reflections are positioned later then the original sound signal. Noise on the other hand can reduce the prominence of the correlated peaks, lowering the distinctness of the local maxima and complicate the PPF determination. In the simulation environment, a clear exponential relationship can be found between the SNR and the optimal PPF. Determining the SNR requires unwanted, additional measurements or noise power estimation techniques, both requiring extra energy on processing and power level.

Setting a fixed prominence factor can resolve this problem. However, choosing the PPF too low can cause erroneous noise peaks to be included in the array; choosing it too high can exclude the real distance peaks, resulting in a similar effect as the maximum method. A passband of PPF values can be derived from distance error plots. According to a simulation of the PPF, it can be shown that the bandwidth is proportional to the SNR. Taking these bandwidths into account, a single value (e.g. PPF = 65) can be derived in which the peak prominence method operates adequately over the different SNR values.

C. Method 3: Delta Peak

The last method that improves the system's accuracy is the delta peak method. In this approach, the difference between two consecutive local maxima is calculated and the peak after the largest positive difference is chosen for further distance estimations. As in the peak prominence method, the correlated data is not altered and the advantage over this latter method is that it is computationally less complex. However, the energy efficiency gain impairs the accuracy. This can be seen for instance on the CDF of the original and adapted methods for a room with a = 0.3 and a SNR of 3dB. E.g. 63% of the delta peak distance calculation errors are smaller than 10 cm, in comparison to 56% with the maximum method, 68% with the quadratic window method and 73% with the peak prominence method. Additionally, the robustness against room characteristics is the lowest of all techniques. Similar to the maximum method, the large delta values due to positive interference imply the wrong index, lowering the system's accuracy. The delta-peak heatmap shows these outliers close to the corners and walls of the simulation environment.

Distance error simulations of all proposed methods when different levels of white noise are added show that the peak prominence approach has the highest accuracy, even at a very low SNR. Fixing the local peak threshold results in similar, negative SNR-dependencies to the other methods. Of the two remaining methods, the positive quadratic window function performs the best. The invention provides arrangements (1000) for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), via the computing component (50), which based on a signal (320) obtained from said (ultra-)sound wave receiving component (30), computes a signal (310), carrying information of said distance (410). In an embodiment of the invention this computation in said computing component (50), being based on cross correlation.

In a further embodiment thereof, said computation is based on determining a maximum in said signal (320) or in a signal derived therefrom, preferably the maximum of the cross correlation function.

In a further embodiment, said computation is based on determining a maximum in a signal derived from signal (320) by applying a window function, selected to relatively increase first peaks in said signal relative to the others. In a first further embodiment a linear window function is used. In an alternative further embodiment a quadratic window function is used. In yet another alternative further embodiment an exponential window function is used.

In an alternative embodiment thereof, said computation is based on evaluating different maxima or peaks in said signal (320) or in a signal derived therefrom, preferably of the cross correlation function, for instance by taking into account a maximum value and/or its location relative to other maxima and/or combining maxima or peak data (e.g. taking the difference of consecutive maxima). Combination with one of the windowing techniques is possible.

FINAL CONCLUSION

With the invention, a system is presented and demonstrated being able to perform cm-accurate, acoustic- based distance measurements with a restricted energy budget. Monte Carlo and acoustic shoebox simulations with 600 distributed microphones show centimeter accuracy for the synchronized wake-up and pulse compression method close to the sound sources. The accuracy of the basic system decreases near walls, due to interference of reflective signals, and in low receive SNR regions. Three methods are proposed to improve this accuracy with a limited processing power. The Peak Prominence Method improves the accuracy in the low SNR scenarios towards a factor 10 for low distance error values. Experimental validations with a self-developed ultrasonic receiver verify the enhanced accuracy methods and confirm the low-power acoustic reception and processing, e.g. with a power consumption of 2.0765 pW for a single measurement or over 37 years of operation on a CR2032 coin cell battery. This result is an agreement with analyses demonstrating comparable results with the used cross correlation method. In accordance with the invention, it is particularly noted that the received signal is a sampled, smaller part of the original signal (e.g. 1ms) instead of the whole signal. Moreover, with the invention calibration solutions are also included, to determine the wake-up time of the receiver hardware components. The latter could impact both precision and power consumption.

The invention provides an arrangement (1000) for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), the arrangement comprising a first device (10), comprising: an (ultra-)sound wave receiving component (30), adapted for receiving said (ultra-)sound waves (200); a device RF wave receiving component (40), adapted for receiving a first RF wave (210); providing a signal (300) to said (ultra-)sound wave receiving component (30) to temporarily activate it, hence the determined estimate is based on smaller (sampled) part of the transmitted (ultra-)sound waves (200). The invention also provides methods for calibrating the used arrangement, in particular the selection of the wake-up time of the (ultra-)sound wave receiving component (30). Alternatively also the parameters of the windowing functions and/or the peak or maxima evaluation methods outlined above are part of a calibration method.

The presented ranging method could be extended to perform positioning. The current system based on a single sound source may be used. For positioning purposes in a 2D environment, at least three sound sources have to be used, whereas a minimum of four are required in case of a 3D environment. This suggests that at the awake time, three signals are received at the same time. Needless to say, identifying which signal comes from which source is an important concern here. A more detailed investigation may reveal which multiple access technique is suited to address this challenge and which existing chirp based techniques are to be tested. The invention provides an arrangement being adapted for determining a position in an environment of one or more of said first devices (10), said environment being provided with one or more of said second devices (20), said position determining being computed from the obtained estimates (400) of distance (410), relative to said second devices (20), based on (ultra-)sound waves (200), whereby these (ultra-)sound waves (200) are provided and hence separately identifiable by use of a multiple access techniques such as FDMA, TDMA, CDMA.

Claims

Claims

1. An arrangement (1000) for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), the arrangement comprising:

• said first device (10), comprising:

o an (ultra-)sound wave receiving component (30), adapted for

^■ receiving said (ultra-)sound waves (200);

o a device RF wave receiving component (40), adapted for

^■ receiving a first RF wave (210);

^■ providing a signal (300) to said (ultra-)sound wave receiving component (30) to (temporarily) activate it;

o a computing component (50), adapted for

^■ computing a signal (310), carrying information of said distance (410), based on a signal (320) obtained from said (ultra-)sound wave receiving component (30);

o a device RF wave transmitting component (60), adapted for

^■ transmitting said computed signal (310) as a second RF wave (220);

• a system (70), comprising:

o said second device (20), comprising:

^■ an (ultra-)sound wave transmitting component (80), adapted for

- transmitting said (ultra-)sound waves (200);

said second device (20), adapted for

^■ providing a signal (330) to a third device (90), said signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200); o said third device (90), comprising:

^■ a system RF wave transmitting component (100), adapted for

- transmitting said first RF wave (210), based on said signal (330); o a fourth device (110), comprising: ^■ a system RF wave receiving component (120), adapted for

- receiving said second RF wave (220);

said fourth device being adapted for providing a signal (340) to a fifth device (130);

o said fifth device (130), comprising:

^■ a computing component (140), adapted for

- determining said estimate (400) of the distance (410) based on said signal (340).

2. The arrangement of claim 1, wherein said computing component (50), being further adapted for performing said computation based on said signal (320) from said (ultra-)sound wave receiving component (30) being only active for a predetermined time duration.

3. The arrangement of claim 1 or 2, wherein said computation in said computing component (50), being based on cross correlation.

4. The arrangement of claim 1 to 3, wherein said computing component (50), being further adapted for determining said estimate (400) of the distance (410) itself.

5. The arrangement of any of the previous claims, comprising of a plurality of said first devices

(10).

6. The arrangement of any of the previous claims, wherein said second device (20) and said third device (90) being combined as one single device (150), optionally being further combined with said fourth device (110) as another single device (160).

7. The arrangement of claim 6, comprising of a plurality of said single devices (150, 160).

8. The arrangement of claim 1, further being adapted for determining a position in an environment of one or more of said first devices (10), said environment being provided with one or more of said second devices (20), said position determining being computed from the obtained estimates (400) of distance (410), relative to said second devices (20).

9. A first device (10) suitable for use in an arrangement (1000) for determining an estimate (400) of the distance (410) between said first device (10) and a second device (20) based on (ultra sound waves (200), said first device (10) comprising:

o an (ultra-)sound wave receiving component (30), adapted for

^■ receiving said (ultra-)sound waves (200);

o a device RF wave receiving component (40), adapted for ^■ receiving a first RF wave (210);

^■ providing a signal (300) to said (ultra-)sound wave receiving component (30) to (temporarily) activate it;

o a computing component (50), adapted for

^■ computing a signal (310), carrying information of said distance (410), based on a signal (320) obtained from said (ultra-)sound wave receiving component (30);

o a device RF wave transmitting component (60), adapted for

^■ transmitting said computed signal (310) as a second RF wave (220); and wherein said second device (20) comprising:

o an (ultra-)sound wave transmitting component (80), adapted for

^■ transmitting said (ultra-)sound waves (200); said second device (20), adapted for

o providing a signal (330), said signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200).

10. The first device (10) of claim 9, further comprising an energy storage device (170) suited for providing energy (500) to said components (30, 40, 50, 60).

11. The first device (10) of claim 9 or 10, wherein the device RF wave receiving component (40), being further adapted for providing a signal (350, 360) to said computing component (50) and/or said device RF wave transmitting component (60) to (temporarily) activate it.

12. A single device (150) suitable for use in an arrangement (1000) for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), said single device (150) comprising:

said second device (20), comprising:

^■ an (ultra-)sound wave transmitting component (80), adapted for

- transmitting said (ultra-)sound waves (200);

said second device (20), adapted for

^■ providing a signal (330) to a third device (90), said signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200); said third device (90), comprising: ^■ a RF wave transmitting component (100), adapted for

- transmitting an RF wave (210), based on said signal (330).

13. A single device (160), comprising the single device (150) of claim 12, further comprising a fourth device (110), comprising:

^■ a RF wave receiving component (120), adapted for

- receiving an RF wave (220);

^■ said fourth device being adapted for providing a signal (340) to a fifth device (130), part of said arrangement, and said fifth device (130) being adapted for determining said estimate (400) of the distance (410) based on said signal (340).

14. An arrangement (1000) for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), the arrangement comprising:

• said first device (10), adapted for

o receiving a first RF wave (210);

o receiving said (ultra-)sound waves (200) upon receiving said first RF wave (210); o computing information (600) of or related to said distance (410), based on said (ultra-)sound wave (200);

o transmitting said information (600) as a second RF wave (220).

• a system (70), comprising:

o said second device (20), adapted for

o transmitting said (ultra-)sound waves (200);

o providing information (610) being indicative of the time of transmitting said (ultra-)sound waves (200);

o a third device (90), adapted for

o transmitting said first RF wave (210), based on said information (610) being indicative of the time of transmitting said (ultra-)sound waves (200);

o a fourth device (110), adapted for

• receiving said second RF wave (220);

• providing information (620) to a fifth device (130); o said fifth device (130), adapted for

- determining said estimate (400) of the distance (410) based on said provided information (620).

15. An arrangement (1000) for facilitating the determining of an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), the arrangement comprising:

• said first device (10), comprising:

o an (ultra-)sound wave receiving component (30), adapted for

^■ receiving said (ultra-)sound waves (200);

o a device RF wave receiving component (40), adapted for

^■ receiving a first RF wave (210);

^■ providing a signal (300) to said (ultra-)sound wave receiving component (30) to (temporarily) activate it;

o a computing component (50), adapted for

^■ computing a signal (310), carrying information of said distance (410), based on a signal (320) obtained from said (ultra-)sound wave receiving component (30), said computing component (50) further being capable to decide to take an action (by providing an action signal (370)), based on said information;

• a system (70), comprising:

o said second device (20), comprising:

^■ an (ultra-)sound wave transmitting component (80), adapted for

- transmitting said (ultra-)sound waves (200);

said second device (20), adapted for

^■ providing a signal (330) to a third device (90), said signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200); o said third device (90), comprising:

^■ a system RF wave transmitting component (100), adapted for

transmitting said first RF wave (210), based on said signal (330).

16. The arrangement of claim 15, wherein said first device (10) further comprises a device RF wave transmitting component (60), adapted for transmitting said computed signal (310) as a second RF wave (220); and said action signal (370) being transmitting said computed signal (310) as a second RF wave (220); and said system (70) further comprises a fourth device (110), comprising:

^■ a system RF wave receiving component (120), adapted for

- receiving said second RF wave (220);

said fourth device being adapted for providing a signal (340) to a fifth device (130);

o said fifth device (130), comprising:

^■ a computing component (140), adapted for determining said estimate (400) of the distance (410) based on said signal (340).

17. The arrangement of claim 15, wherein said carried information being said estimate (400) of the distance (410) itself and said action being taking a security measure based thereon.

18. The arrangement of claim 15 to 17, wherein said computing component (50), being further adapted for performing said computation based on said signal (320) from said (ultra-)sound wave receiving component (30) being only active for a predetermined time duration.

19. The arrangement of claim 15 to 18, wherein said computation in said computing component (50), being based on cross correlation.

20. The arrangement of claim 15 to 19, wherein said computing component (50), being further adapted for determining said estimate (400) of the distance (410) itself.

21. The arrangement of any of the previous claims, comprising of a plurality of said first devices

(10).

22. The arrangement of claim 16 , wherein said second device (20) and said third device (90) being combined as one single device (150), optionally being further combined with said fourth device (110) as another single device (160).

23. The arrangement of claim 22, comprising of a plurality of said single devices (150, 160).

24. The arrangement of claim 15 or 16, further being adapted for determining a position in an environment of one or more of said first devices (10), said environment being provided with one or more of said second devices (20), said position determining being computed from the obtained estimates (400) of distance (410), relative to said second devices (20).

25. A first device (10) suitable for use in an arrangement (1000) for determining an estimate (400) of the distance (410) between said first device (10) and a second device (20) based on (ultra sound waves (200), said first device (10) comprising:

o an (ultra-)sound wave receiving component (30), adapted for

^■ receiving said (ultra-)sound waves (200);

o a device RF wave receiving component (40), adapted for

^■ receiving a first RF wave (210);

^■ providing a signal (300) to said (ultra-)sound wave receiving component (30) to (temporarily) activate it;

o a computing component (50), adapted for

^■ computing a signal (310), carrying information of said distance (410), based on a signal (320) obtained from said (ultra-)sound wave receiving component (30); said computing component (50) further being capable to decide to take an action (by providing an action signal (370)), based on said information; and wherein said second device (20) comprising:

o an (ultra-)sound wave transmitting component (80), adapted for

^■ transmitting said (ultra-)sound waves (200); said second device (20), adapted for

^■ providing a signal (330), said signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200).

26. The first device (10) of claim 25, further comprising a device RF wave transmitting component (60), adapted for transmitting said computed signal (310) as a second RF wave (220), and said first device (10) further comprising an energy storage device (170) suited for providing energy (500) to said components (30, 40, 50, 60).

27. The first device (10) of claim 25 or 26, wherein the device RF wave receiving component (40), being further adapted for providing a signal (350, 360) to said computing component (50) and/or said device RF wave transmitting component (60) to (temporarily) activate it.

28. A method for determining an estimate (400) of the distance (410) between a first device (10) and a second device (20) based on (ultra-)sound waves (200), the method comprising the steps of: (i) transmitting said (ultra-)sound waves (200) by said second device (20); (ii) (thereafter) transmitting a first RF wave (210) by said second device (20) based on a signal (330) being indicative of the time of transmitting said (ultra-)sound waves (200); (iii) receiving said first RF wave (210) and (temporarily) activating an (ultra-)sound wave receiving component (30) within said first device (10); (iv) receiving said (ultra-)sound waves (200) by said (ultra-)sound wave receiving component (30); (v) computing within said first device (10) a signal (310), carrying information of said distance (410), based on a signal (320) obtained from said (ultra-)sound wave receiving component (30); (vi) transmitting said computed signal (310) as a second RF wave (220); (vii) receiving said second RF wave (220) and (viii) determining said estimate (400) of the distance (410) based on said second RF wave (220).

29. The method of claim 28, wherein the time of transmitting said first RF wave (210) being further being based on upon a (estimated) range within which the first and second devices are desired to be positioned.

30. The method of claim 28 or 29, wherein said transmitting a first RF wave (210) occurs before the end of the (ultra-)sound waves (200) as received, preferably the first RF wave is sent at a time which is within the last 10% of the duration of the sound wave.