DETERMINATION OF MOVEMENT PATTERNS OF A SIGNAL SOURCE

Abstract

According to the present invention, a method for determining the movement patterns of a low-energy signal source in a monitored area is provided, wherein in the monitored area a virtually bounded part of a vertical surface is defined as a gate by a one-dimensional antenna array, the low-energy signal source emits a signal with a constant frequency, all antennas of the antenna array sequentially sample the signal for a sequence of a certain number of periodically offset sampling times, and the movement of the low-energy signal source with respect to the gate is tracked by monitoring, based on calculated phase shifts, whether the low-energy signal source crosses the gate and consequently in which direction the low-energy signal source crosses the gate.whether the low-energy signal source moves near the gate without crossing it, and consequently on which side of the gate the low-energy signal source moves, and whether the low-energy signal source is near the gate without any movement being clearly detected.

Description

[0001] In facilities such as warehouses, factories, or similar locations with many stored items, it is crucial to be able to find stored items easily and quickly. Factories often comprise several rooms with different production stations and / or assembly lines, which may contain various machines for different processing steps. Components are typically transported to other rooms multiple times during a production process.

[0002] Alternatively, workshops, for example, might store many spare parts that need to be located when needed. If such spare parts are kept in a hall with numerous shelves or rooms, finding stored items can be time-consuming, for example, if an employee is looking for something they didn't put away themselves. In large factories, it can also be important to know in which rooms or stations the relevant components are located.

[0003] In warehouses, such as those used in production logistics, real-time tracking systems are a particularly effective solution for inventory management and asset tracking. By tracking the real-time location of products or product pallets within the facility, inventory counts can be automated, and misplaced products or materials can be located without delay. More efficient inventory management can reduce the need to order additional items to replace lost ones. Knowing the location of products and materials reduces the time spent searching for items, shortens production cycles, and allows employees to use their time more productively. All of this leads to significant savings and increased profits for the company.

[0004] To avoid a time-consuming search, this application discloses a method, a device, and a system for determining movement patterns, according to which an object is equipped with a signal source and the movement pattern of the signal source is tracked. This enables the real-time tracking of flows and movement patterns of, for example, components or stored objects. <Technischer Hintergrund>

[0005] Recently, the greatest successes in solving the problem of object location have been achieved using technology for measuring and calculating the angle of arrival (AoA) of unmodulated sinusoidal signals from BLE sources placed on these objects. Known methods for determining the angle of arrival of electromagnetic waves by analyzing a large number of IQ-sampled signals (e.g., the MUSIC method) use the power of the incoming signals as an informative parameter. Therefore, the accuracy of AoA determination decreases in environments with signal reflections from other objects, despite the use of computationally intensive algorithms. Inaccurate determination of the directions to the tracked objects often leads to unacceptable errors in determining their positions or movement patterns.The method proposed in this invention for determining the angle of incidence involves simple and computationally efficient statistical processing of a representative sample from a plurality of angles of incidence determined during received electromagnetic waves. German patent application DE 10 2011 103 213 A1 discloses a method and a detection device that distinguish between signals emitted by signal transmitters moving through a detection path and other signals. In this process, the signals are received by at least one receiver and selected and compared according to time intervals, signal transmitters, and receivers. German patent application DE 10 2009 057 349 A1 describes a differentiation method that, in the case of an RFID gate, separates the signals of RFID tags moving through the gate from interference signals based on signal strength differences and their dynamics.German patent DE 103 54 872 A1 discloses a device for detecting the direction of a target, which calculates the direction on the basis of phase differences of received radio waves and for this purpose comprises a direction calculation device, a range determination device and a direction correction device. Summary of the invention

[0006] According to the present invention, a method for determining the motion patterns of a low-energy signal source in a monitored area is provided.

[0007] According to the procedure - in the monitored area, a virtually bounded part of a vertical surface is defined as a gate by at least one one-dimensional antenna array, and - the motion pattern of the low-energy signal source relative to the gate is defined by the tendency to change an angle of incidence from +90° to -90° or from -90° to +90° with a transition through 0° at time t0, and - the movement of the low-energy signal source relative to the gate is tracked according to the motion scheme by means of at least one one-dimensional antenna array, by tracking, whether the low-energy signal source crosses the gate and consequently in which direction the low-energy signal source crosses the gate, whether the low-energy signal source moves near the gate without crossing it, and consequently on which side of the gate the low-energy signal source moves, and whether the low-energy signal source is near the gate without any movement being clearly detected, whereby - the procedure includes the following: Selecting representative samples from a plurality of calculated angles of arrival, which are calculated from signals sampled at different sampling times, wherein the representative sample comprises calculated angles of arrival based on phases received at pairs of two adjacent antennas of the at least one one-dimensional antenna array,

[0008] Calculating one statistical value for each representative sample as a statistically calculated angle of incidence, Selecting a time t0 in a period with high received signal strength by calculating a correlation coefficient P between a model of an ideal signal and a real signal for various possible times t of crossing the gate in the period with high received signal strength, where time t0 is chosen as the time t with the highest value of the correlation coefficient P. Selecting a first subsequence of statistically calculated angles of incidence before time t0 and a second subsequence of statistically calculated angles of incidence after time t0, by selecting two subsequences that are as large as possible, provided that both subsequences are of equal size, Determine, based on the statistically calculated angles of incidence of the first subsequence and the statistically calculated angles of incidence of the second subsequence, whether the low-energy signal source has crossed the virtual gate during the sequence. If it is determined that the low-energy signal source crossed the virtual gate during a sequence comprising the first subsequence and the second subsequence, determine the direction in which the low-energy signal source crossed the virtual gate by determining a probable motion pattern based on a tendency for the angle of incidence to change from +90° to -90° with a transition through 0° at time t0, and If it is determined that the low-energy signal source did not cross the virtual gate during the sequence, determine whether the low-energy signal source was near the virtual gate or on the side of the virtual gate with the positive sign of the signal's angle of incidence or on the side of the virtual gate with the negative sign of the signal's angle of incidence throughout the entire sequence, based on the statistically calculated angles of incidence of the entire sequence.

[0009] This makes it possible to track the movement patterns of a signal source and, accordingly, to quickly locate an object equipped with a signal source when needed. Furthermore, the accuracy of the tracked movement patterns can be improved.

[0010] According to an example of the invention, the angle of arrival of the signal from the low-energy signal source can be calculated using the angle-of-arrival (AOA) method. In such a case, the low-energy signal source emits an unmodulated sinusoidal signal (continuous tone), and all antennas of the antenna array sequentially sample the signal for a sequence of a specific number of periodically offset sampling times. Furthermore, in such a case, the method comprises calculating the phase of the sampled signal for each antenna of the antenna array for the sequence of a specific number of periodically offset sampling times, calculating the phase shift of signals received at two adjacent antennas, and calculating an angle of arrival of the signal based on the calculated phase shift and the distance between two antennas.

[0011] Calculating the angle of arrival according to the Angle-of-Arrival (AOA) method makes it possible to implement the present invention with simple and therefore also cost-effective signal sources such as a Bluetooth Low Energy (BLE) tag.

[0012] Furthermore, in the procedure, if it is determined that the low-energy signal source has not crossed the virtual gate during the sequence, a difference can be calculated between the number of statistically calculated positive angles of arrival and the number of statistically calculated negative angles of arrival, and based on the sign of the difference, it can be determined on which side of the virtual gate the low-energy signal source is located.

[0013] Thus, movement patterns of the signal source can also be tracked even if the signal source does not pass through a virtual gate.

[0014] If the magnitude of the difference is less than a predetermined value, it can be determined that the low-energy signal source was located near the virtual gate throughout the entire sequence.

[0015] Thus, movement patterns can also be tracked if the signal source does not cross a virtual gate, or hardly moves or does not move at all.

[0016] Furthermore, the correlation coefficient P can be calculated in the procedure based on a difference between the number of statistically calculated angles of incidence before time t0 with a positive sign and the number of statistically calculated angles of incidence before time t0 with a negative sign, and based on a difference between the number of statistically calculated angles of incidence after time t0 with a negative sign and the number of statistically calculated angles of incidence after time t0 with a positive sign.

[0017] This makes it easy, or requires minimal computational effort, to track the direction in which the signal source crossed the virtual gate. This also makes it possible to improve the accuracy of the tracked motion patterns, or to correctly determine the tracked motion patterns despite some erroneous samples.

[0018] In the monitored area, several one-dimensional antenna arrays, each with a plurality of antennas at a constant distance from each other, can be installed, on the basis of which a plurality of virtual gates are defined, and it can be stored for the low-energy signal source in which virtual space limited by virtual gates the low-energy signal source is located.

[0019] This makes it easy to find objects even if they are spread over very large areas or areas with many buildings or rooms.

[0020] Preferably, a calculated phase shift is corrected and the angle of incidence of the sampled signal is calculated based on the corrected phase shift.

[0021] This allows movement patterns to be reliably determined and tracked even in the presence of many interfering signals.

[0022] Furthermore, according to the present invention, a device for determining the movement patterns of a low-energy signal source in a monitored area is also provided.

[0023] This is - in the monitored area, a virtually limited part of a vertical surface is defined as a gate by at least one one-dimensional antenna array, - a motion scheme of the low-energy signal source relative to the gate is defined by the tendency of the change of an angle of incidence from +90° to -90° or from -90° to +90° with transition through 0° at time t0, and - the device is configured to track a movement of the low-energy signal source with respect to the gate according to the motion scheme by means of the at least one one-dimensional antenna array, by the device being configured to track, whether the low-energy signal source crosses the gate and consequently in which direction the low-energy signal source crosses the gate, whether the low-energy signal source moves near the gate without crossing it, and consequently on which side of the gate the low-energy signal source moves, and whether the low-energy signal source is near the gate without any movement being clearly detected, whereby the device is set up to: to select representative samples from a plurality of calculated angles of arrival, which are calculated from signals sampled at different sampling times, wherein the representative sample comprises angles of arrival calculated based on phases received at pairs of two adjacent antennas of the at least one one-dimensional antenna array, to calculate a statistical value for each representative sample as a statistically calculated angle of incidence, to select a time t0 in a period with high received signal strength by calculating a correlation coefficient P between a model of an ideal signal and a real signal for various possible times t of crossing the gate in the period with high received signal strength, wherein the device is configured to select as time t0 the time t with the highest value of the correlation coefficient P, to select a first subsequence of statistically calculated angles of incidence before time t0 and a second subsequence of statistically calculated angles of incidence after time t0, by selecting two subsequences that are as large as possible, provided that both subsequences are of equal size, to determine whether the low-energy signal source has crossed the virtual gate during the sequence, based on the statistically calculated angles of arrival of the first subsequence and the statistically calculated angles of arrival of the second subsequence, If it is determined that the low-energy signal source crossed the virtual gate during a sequence comprising the first subsequence and the second subsequence, to determine in which direction the low-energy signal source crossed the virtual gate by determining a probable motion pattern based on a tendency of a change in the angle of incidence from +90° to -90° with a transition through 0° at time t0, and wherein If it is determined that the low-energy signal source did not cross the virtual gate during the sequence, the device is further configured to determine whether the low-energy signal source was located near the virtual gate or on the side of the virtual gate with the positive sign of the angle of incidence of the signal or on the side of the virtual gate with the negative sign of the angle of incidence of the signal during the entire sequence, based on the statistically calculated angles of incidence of the entire sequence.

[0024] This makes it possible to track the movement patterns of a signal source and, accordingly, to quickly locate an object equipped with a signal source when needed. Furthermore, the accuracy of the tracked movement patterns can be improved.

[0025] According to an example of the invention, the angle of arrival of the signal from the low-energy signal source can be calculated using the angle-of-arrival (AOA) method. In such a case, the low-energy signal source transmits an unmodulated sinusoidal signal, and all antennas of the antenna array sequentially sample the signal for a sequence of a specific number of periodically offset sampling times. Furthermore, in such a case, the device is configured to calculate, for the sequence of a specific number of periodically offset sampling times, the phase shift of signals received at two adjacent antennas, and an angle of arrival of the signal based on the calculated phase shift and the distance between two antennas.

[0026] Calculating the angle of arrival according to the Angle-of-Arrival (AOA) method makes it possible to implement the present invention with simple and therefore also cost-effective signal sources such as a Bluetooth Low Energy (BLE) tag.

[0027] If it is determined that the low-energy signal source has not crossed the virtual gate during the sequence, the device is further configured to calculate a difference between the number of statistically calculated positive angles of incidence and the number of statistically calculated negative angles of incidence, and to determine, based on the sign of the difference, on which side of the virtual gate the low-energy signal source is located.

[0028] Thus, movement patterns of the signal source can also be tracked if the signal source does not cross a virtual gate or moves very little or not at all.

[0029] If the magnitude of the difference is less than a predetermined value, the device is set up to determine that the low-energy signal source was in close proximity to the virtual gate throughout the entire sequence.

[0030] This allows movement patterns to be tracked even if the signal source does not pass through a virtual gate.

[0031] Preferably, the device is configured to calculate the correlation coefficient P based on a difference between the number of statistically calculated angles of incidence before time t0 with a positive sign and the number of statistically calculated angles of incidence before time t0 with a negative sign, and based on a difference between the number of statistically calculated angles of incidence after time t0 with a negative sign and the number of statistically calculated angles of incidence after time t0 with a positive sign.

[0032] This improves the accuracy of the tracked motion patterns. Furthermore, it makes it easy to track the direction in which the signal source crossed the virtual gate.

[0033] The representative samples from a plurality of calculated angles of arrival can be limited to calculated angles of arrival that are calculated from signals with the highest received signal strength.

[0034] This can improve the accuracy of the tracked movement patterns.

[0035] Preferably, the device is configured to correct a calculated phase shift and to calculate the angle of incidence of the sampled signal based on the corrected phase shift.

[0036] This allows movement patterns to be reliably determined and tracked even in the presence of many interfering signals.

[0037] Furthermore, according to the present invention, a system with a plurality of devices for determining movement patterns of a low-energy signal source in a monitored area is provided.

[0038] This makes it easy to find objects even if they are spread over very large areas or many rooms.

[0039] Furthermore, according to the present invention, a computer program is provided with instructions which, when executed by a device, cause the device to perform a method according to the present invention.

[0040] Furthermore, according to the present invention, a non-transitory computer-readable storage medium is provided, containing instructions which, when executed by a device, cause the device to execute a method according to the present invention. Brief description of the characters Fig. Figure 1 illustrates an example of an antenna array and a virtual gate according to the present invention; Fig. 2 illustrates QI sampling; Fig. Figure 3 illustrates a calculation of the angle of arrival of a signal using the angle-of-arrival method; Fig. 4 illustrates an application of the present invention in a warehouse; Fig. Figure 5 illustrates an application of the present invention in a production facility with several production or storage halls. Example(s) of implementation

[0041] According to one embodiment, a system or facility in which objects are stored is referred to as a monitored area. In this monitored area, one or more virtual gates are defined according to the present invention.

[0042] The following is an example in which the angle of arrival of a signal source is determined and calculated using the Angle-of-Arrival (AOA) method. However, the present invention is not limited to determining the angle of arrival using the Angle-of-Arrival (AOA) method, and the angle of arrival can be determined and calculated using any other method.

[0043] As in Fig. As illustrated in Figure 1, the virtual gate is defined based on a one-dimensional antenna array, i.e., a linear antenna array. As shown in Figure 1, the virtual gate is defined based on a one-dimensional antenna array, i.e., a linear antenna array. Fig. As shown in Figure 1, the linear antenna array comprises a plurality of antennas arranged in a row. In the figure shown in Fig. In the example shown, the majority of antennas is 4. However, the majority of antennas is not limited to 4 and can be any other number. The majority of antennas are arranged horizontally in a row, for example, in a first direction that is in Fig. The virtual gate, defined by the array of antennas, is a vertical plane that is (essentially) perpendicular to the array of antennas, or rather, the array of antennas intersects the plane of the virtual gate.

[0044] For example, the array of antennas is arranged in a first direction, for instance parallel to the X-axis. The virtual gate is then defined in a vertical plane perpendicular to the first direction. That is, in which Fig. In the illustrated example 1, the virtual gate is defined in or parallel to the yz plane.

[0045] In Fig. Figure 1 also shows a point P on the axis of the linear antenna array. If a low-energy signal source is located at this point P, the signal is assumed to have an angle of incidence of 0°. The time at which the low-energy signal source crosses the gate and is, for example, at point P, is denoted by t0.

[0046] Furthermore, two arrows are shown in the figure. If a low-energy signal source is located on an arrow, or if an arrow points to a low-energy signal source, an angle of incidence of -90° or +90° is assumed, respectively. Which direction (side of the gate) corresponds to a positive angle of incidence and which direction (side of the gate) corresponds to a negative angle of incidence can be chosen arbitrarily, for example, when installing the gate. That is to say, the present invention is not limited to the one described in the figure. Fig. The sign convention shown is limited.

[0047] Furthermore, according to one embodiment, objects whose movement patterns are tracked and determined are equipped with Bluetooth Low Energy (BLE) tags. A BLE tag is an example of a low-energy signal source. However, according to the present invention, the object to be tracked can also be equipped with a transmitter or signal source according to a different standard.

[0048] These BLE tags emit a continuous tone, i.e., an unmodulated sinusoidal signal (with a defined error tolerance). This continuous tone corresponds, for example, to UHF radio signal packets of a known frequency F, such as 250 kHz (with a defined error tolerance), and can be transmitted over a randomly changing frequency channel.

[0049] When the signal of the BLE tag is received at the antenna array, it is possible to calculate the angle of arrival (as an azimuth angle) of the signal relative to the antenna array using the Angle-of-Arrival (AOA) method. < Calculation of the angle of arrival using the Angle-of-Arrival (AOA) method >

[0050] The Angle-of-Arrival (AOA) method is described below with reference to Fig. 2. The received signal is sampled sequentially at all antennas. Sequential sampling means that each antenna samples the received signal in succession, for example, first antenna A1, then antenna A2, then antenna A3, then antenna A4, and then starting again with antenna A1, and so on. A sample value of a signal typically contains the amplitude and phase of the sampled signal, measured, for example, according to IQ sampling.

[0051] In IQ sampling, the sampled signals of the received packets (UHF radio signal packets) are treated as sine waves and converted into IQ coordinates. Digital in-phase (I) and quadrature (Q) samples (IQ samples) are generated at a frequency of 4F within each period of the sine wave signals of each packet. An example of an IQ sample is shown in Fig. 2 shown.

[0052] Preferably, after receiving the first v ≥ 2 periods of the signal at frequency F, a cyclic switching of the n antennas of a one-dimensional array is performed, whereby the signal samples (sample slot) alternate with pauses in reception (switch slot) at equal time intervals of 1 / 4F. This enables synchronization.

[0053] For each of the IQ samples, the phase φ is determined. r The unmodulated sinusoidal signals received by the antennas, where r=1,2,3,..., are interpreted as reference and working samples. The phase φ is then... rDepending on the quadrant of a Cartesian coordinate system in which the IQ sample is located, it is calculated as follows: if I>0 (1st or 4th quadrant of the Cartesian coordinate system), then: φr=arctan Q / I, If I<0, Q>0 (2nd quadrant of the Cartesian coordinate system), then: φr=π+ arctan Q / I, If I<0, Q<0 (3rd quadrant of the Cartesian coordinate system), then: φr=−π+ arctan Q / I, where arctan stands for the arctangent.

[0054] Next, the phase shift Δφ, i.e., the difference between the phases measured at two adjacent antennas, is calculated. The phase shift Δφ is calculated such that it has a value between -n and +π. In particular, the phase shift Δφ can be calculated as follows: if(φr+1−φr)>π, then Δφ=φr+1−φr−2π, if(φr+1−φr)<−π, then Δφ=φr+1−φr+2π, if −π<(φr+1−φr)<π, then Δφ=φr+1−φr,

[0055] Thus, the calculated phase shift is mapped to a value range between -180° and +180°.

[0056] Preferably, in the first v ≥ 2 periods of the received signal, a phase shift of the radio frequency F of each received signal packet is calculated by averaging the difference between the phase values ​​of the ideal (-90°) and the actual sequence using the formula: Shift=[−90⋅4v−∑14v(φv+1−φv)] / 4v degrees / smpl

[0057] Grad / smpl describes a so-called drift by which the received or sampled phase shifts per received sample signal, i.e., per sampling period of an antenna; in other words, by how many degrees the shift or drift progresses between two successively sampled signals.

[0058] The phase correction of the signals from the 1st, 2nd, ..., nth antenna is performed in each of the q switching cycles (the cyclic switching of the antennas) of each packet according to the calculated shift or phase shift as per the following formula: corrφm,k=(4v−1)⋅Shift+r⋅Shift, where v≥2, m=1,2, ..., n, k=1,2,... ,qr=0,1,2,... n·q-1

[0059] A switching cycle q can also be referred to as a sampling point.

[0060] Subsequently, differences of the (n-1) pairs of corrected phase values ​​of the radio signals of the neighboring antennas A2-A1, A3-A2, ..., An-An-1, are expressed as corrφ according to the formulas. m,k calculated: ψi,k=corrφi,k,(Ai+1)−corrφi,k(Ai)−π, if 0 <corrφi,k(Ai)< π, ψi,k=corrφi,k,(Ai+1)−corrφi,k(Ai)−π, if 0 >corrφi,k(Ai)>−π, where i=1, 2, ..., (n-1), k=1, 2,..., q, - π < Ψ i,k < π

[0061] In the following, the phase difference, preferably the corrected phase difference or phase offset, between two adjacent antennas is denoted by ψ.

[0062] Correcting the calculated phases based on the calculated shift makes it possible to accurately determine the angle of incidence of a signal, even in the presence of many other interfering signals.

[0063] The following describes the calculation of the signal's angle of incidence θ using... Fig. 3 explained. For the sake of simplicity, in Fig. Figure 3 shows only two antennas, but preferably the antenna array includes more than two antennas.

[0064] In Fig. Figure 3 illustrates a scenario in which a transmitter, i.e., a signal source such as the BLE tag, transmits a signal. Antenna A1 of the antenna array is located farther from the transmitter than antenna A2. Consequently, signals with different phases are received at antennas A1 and A2. The dashed line connecting the transmitter to antenna A1 indicates the direction in which the signal received by antenna A1 propagates. In a real-world application, the distance between the antenna array and the transmitter is many times greater than the distance d between any two antennas. Therefore, for trigonometric purposes, a line connecting the transmitter to antenna A1 and a line connecting the transmitter to antenna A2 can be assumed to be equal to, or parallel to, the antennas.

[0065] The distance between two adjacent antennas is denoted by d. If a line from the signal transmitter to the antenna array is not exactly perpendicular to the first direction in which the antennas of the linear antenna array are arranged in a row, the distance from the transmitter to each antenna will be different. This difference in distance results in different measured phases at the different antennas when the signal is sampled. The difference in distance corresponding to the phase shift is calculated using ψ * λ / 2π. It can be assumed that for a phase shift between 0 and +π, the corresponding angle of incidence of the signal will be between 0° and 180°.

[0066] This means the angle of incidence AoA is calculated as follows: - if θ <= 90, then AoA = θ, - if θ > 90 then AoA = 90 - θ.

[0067] This serves to adapt the angle of incidence θ to the gate model, so that a zero line passes through the center of the antenna array, and the angles on one side of the gate or antenna array are 0 to +90 degrees, and on the other side are -90 to 0 degrees.

[0068] Thus, the in Fig. Calculate the 3 plotted angles of arrival θ (AoA) in azimuth using the following formula: θ=arcsin((ψ*λ) / (2πd)); where −90°<θ<90°,k=1,2,…,q,i=1,2,…,(n−1), λ - the wavelength of the received high-frequency oscillations of the signals is, d - represents the distance between adjacent antennas of a one-dimensional antenna array.

[0069] This calculation is performed repeatedly in each of the q = 2N / n antenna switching cycles during the duration of each received signal packet and for multiple antenna pairs. For example, in each of q switching cycles, n antennas are switched sequentially for the duration of each packet to calculate q·(n-1) incidence angle of the BLE signals of the day. In the example with n=4 antennas in an antenna array, there are n-1=3 antenna pairs for which a phase shift is calculated for each of q switching cycles, and based on this, the incidence angle of the signal is determined.

[0070] Alternatively, any other method can be used to calculate the position of the signal source relative to the antenna array, i.e., relative to the virtual gate. < Determination of movement patterns based on calculated angles of incidence >

[0071] The following describes how, based on the calculated angles of incidence of the signal, movement patterns of the signal source can be detected and determined.

[0072] For the following description, it is assumed that a positive sign of the signal's angle of incidence θ indicates that the signal source is located on one side of the virtual gate, whereas a negative sign of the signal's angle of incidence θ indicates that the signal source is located on the other side of the virtual gate. At a time when the signal source is traversing the virtual gate, i.e., is inside the virtual gate, the angle of incidence is 0°.

[0073] According to the present invention, a movement pattern of the signal source is detected by tracking a movement of the low-energy signal source with respect to the gate, by tracking, • whether the low-energy signal source crosses the gate and consequently ◯ in which direction the low-energy signal source crosses the gate, i.e., from the side with a positive sign of the signal's angle of incidence to the side with a negative sign of the signal's angle of incidence, or vice versa, • whether the low-energy signal source moves near the gate without crossing it and consequently ◯ on which side of the gate the low-energy signal source is moving, i.e., on the side with the positive sign of the signal's angle of incidence or on the side with the negative sign of the signal's angle of incidence, and • whether the low-energy signal source is near the gate without any movement being clearly detected.

[0074] Such a determination of the movement patterns of a signal source makes it possible to determine or identify, at any given time, on which side of a gate a corresponding signal source is located. In particular, probable movement patterns when crossing the gate area are determined based on the tendency of the change in the angle of incidence (AoA) from +90° to -90° or from -90° to +90° with a transition through 0° at time t0.

[0075] To trace the signal source, a representative sample is recorded. For this representative sample, the received signal packets of each (n-1) neighboring antenna pair of a one-dimensional antenna array, including their RSSI value, are accumulated or collected.

[0076] From this representative sample, a value and the sign (+ / -) of a statistical value such as an average median value of the angle-of-arrival values ​​(mAoA) of the (n-1) sequences of each received signal packet are then calculated for each sampling time as a statistically calculated angle of arrival.

[0077] An average median value is generated or calculated through the following steps: - After all antennas have sequentially sampled the received signal, a median value is formed for each antenna from a specific number of phases successively sampled by the antenna, - then the (n-1 for n antennas) differences (phase shift) of the median values ​​of the phases of two adjacent antennas are calculated, - the average of the n-1 differences of the median values ​​of the phases, i.e. the average of the n-1 calculated phase offsets, then yields the average median value.

[0078] That is, the angle of incidence of a signal is calculated by forming a median value for each antenna of at least one one-dimensional antenna array from a plurality of phases of signals received at different times, calculating a difference between the median values ​​for each pair of two adjacent antennas, and calculating the angle of incidence as the average of the differences of all pairs of two adjacent antennas.

[0079] Furthermore, a gate can also be defined by a plurality of one-dimensional antenna arrays. That is, a gate is formed by at least one one-dimensional antenna array. If the gate is defined by a plurality of one-dimensional antenna arrays, the angle of incidence is calculated as the average of the differences of all pairs of two adjacent antennas from all of the plurality of one-dimensional antenna arrays. In other words, in such a case, corresponding average median values ​​are calculated for all pairs of all adjacent antennas from the plurality of antenna arrays.

[0080] Accordingly, when using multiple (for example, 2 or 3) one-dimensional antenna arrays, the averaged median values ​​of the arrival angles of each received AoA packet are combined, ordered according to the time of their arrival, and processed as a sequence of signals from a single one-dimensional antenna array.

[0081] By using such average median values, the present invention becomes more robust and less susceptible to measurement errors caused by, for example, interference.

[0082] Preferably, in the sequence of received AoA packets, a plurality (for example, 3 or 5, but not limited thereto, preferably an odd number) with the highest RSSI values ​​is identified and selected for each sampling time, and their indices are mapped to the sequence of average median AoA values ​​in the order of their arrival. A range between the first and last indices in the order of their arrival can also be highlighted. For each index of the selected section of the sequence of AoA values, a correlation coefficient P between actual and ideal AoA strings of a defined length—i.e., between a model of an ideal signal and a real signal, where the string represents a crossing of the gate area—is then calculated for various time points t using the following formulas: P1=((first.plus−first.minus)+(second.minus−second.plus)) / L P2=−P1, P=max(P1 or P2), −1≤P1≤1, 0≤P≤1, where L is the length of the acquisition window (for example, 32 mean median values), and first.plus, first.minus, second.minus, second.plus indicate the number of mean median values ​​of the AoA (angle of incidence) values ​​with plus or minus signs in the first or second half of the acquisition window, i.e. - first. plus is the number of average median values ​​of the incidence angles with a positive incidence angle in the first half of the detection window, - first. minus is the number of average median values ​​of the angles of incidence with a negative angle of incidence in the first half of the detection window. - second.plus is the number of average median values ​​of the incidence angles with a positive incidence angle in the second half of the detection window, - second.minus is the number of average median values ​​of the angles of incidence with a negative angle of incidence in the second half of the detection window.

[0083] The time t for which the value P is highest is chosen as the time t0 of gate crossing, i.e., the time of crossing the virtual gate. That is, time t0 is the time at which (based on the calculated angles of incidence) it is most likely that the signal source has crossed the virtual gate.

[0084] This means that, based on the distribution of plus and minus signs of the average median values ​​AoA relative to t0, possible movement patterns of the marked objects in relation to the gate are calculated, and the time t0 with the highest probability of crossing the gate is selected.

[0085] The region described above with the highest signal strength, i.e., with the highest RSSI values, typically coincides with the period when the signal source is closest to the antenna array. Therefore, the period between the first and last sampling times selected with the highest RSSI values ​​is the period during which the signal source is most likely to have passed through the gate.

[0086] Consequently, it is advantageous to choose t0 between the first and the last selected sampling time with a high RSSI value. This can be done by calculating the value P for all time points t from the first sampling time with the highest RSSI values ​​to the last sampling time with the highest RSSI values, where t is a possible time of gate traversal and t0 is set to the value t for which the value P is maximal.

[0087] The values ​​P, P1, and P2 are then calculated for the selected time point t0, preferably for two equally sized, ideally large, subsequences of mean median values ​​before and after time t0. This means that time t0 does not necessarily lie in the middle of a representative sample with multiple mean median values. For example, 20 mean median values ​​may be available for the period before t0, while only 13 mean median values ​​are available for the period after t0. In such a case, the last 13 mean median values ​​before t0 and the 13 available mean median values ​​after t0 could be used as the data set for calculating the values ​​P, P1, and P2. That is, two subsequences that are as large as possible are selected, provided that both subsequences are of equal size.

[0088] P represents a correlation coefficient between the calculated angles of incidence and the probability that the signal source has passed through the virtual gate. That is, the larger the value of the coefficient P, the higher the probability that the signal source has passed through the virtual gate.

[0089] Consequently, the magnitude of the coefficient P can be used to determine whether the signal source has passed through the virtual gate or not. For example, if the value of the coefficient P is below a defined threshold (for example, 0.1), a decision can be made as to whether the tag, i.e., the signal source, is near the virtual gate (presence scheme) or has moved past one side of the gate without passing through it.

[0090] If the value of the coefficient P is above the specified threshold (for example, 0.1), it can be determined that the signal source has passed through the virtual gate.

[0091] Once it has been determined that the signal source has passed through the virtual gate, the calculated angle of incidence can be determined based on the signs or on the basis of the value P1 calculated as described above.

[0092] To determine the direction of passage through the virtual gate, medians of the mean AoA values, i.e., angles of arrival, of the signals in the packet sequence are calculated on the intervals before time t0 and after time t0.

[0093] If the sign (of the median value) of the angle of incidence is positive in the period before time t0 and negative in the period after time t0, the tag or signal source has crossed the virtual gate according to a "plus-minus" scheme; otherwise, according to a "minus-plus" scheme.

[0094] Alternatively, the direction of crossing the gate can be determined by assuming that if P1 is greater than zero, the signal source has crossed the virtual gate according to the "plus-minus" scheme, and if P1 is negative, it can be determined that the signal source has crossed the virtual gate according to the "minus-plus" scheme.

[0095] If the sign of the two corresponding median values ​​of the angle of incidence is the same in both intervals, before and after t0, it can be determined that the signal source did not cross the virtual gate after all.

[0096] If, for example, it has already been determined based on the coefficient P that the signal source has not crossed the virtual gate, the side of the gate on which the signal source is located is then determined.

[0097] To determine the side of the virtual gate where the signal source is located, the number of sign (n+) and sign (n-) values ​​of the mean angle of arrival (AoA) in the entire received packet sequence are calculated and averaged. Here, n+ is the number of calculated angles of arrival during the entire interval with a positive sign, and n- is the number of calculated angles of arrival during the entire interval with a negative sign.

[0098] If n+ is greater than n-, it is determined that the signal source is located on the side of the gate corresponding to an incidence angle θ of the signal greater than zero, and otherwise it is determined that the signal source is located on the side of the gate corresponding to an incidence angle θ of the signal less than zero.

[0099] Alternatively, a correlation coefficient -1 ≤ P ≤ 1 can be calculated for the day movement patterns “+-” and “-+”, according to the following formulas: P(plus)=(n+−n−) / (n++n−) P(minus)=−P(plus)

[0100] The larger of the two correlation coefficients P(minus) and P(plus) determines the day-movement pattern “plus” or “minus”.

[0101] If the values ​​of P(minus) and P(plus) are approximately equal, or if their magnitude is less than a threshold, a decision is made regarding the presence scheme. That is, if neither side of the gate can be clearly determined as the location, it is determined that the signal source is located in or very close to the virtual gate. <Examples of applications for determining movement patterns>

[0102] The motion pattern determination described above can be applied, for example, in large warehouses or factory complexes. For this purpose, several two-dimensional antenna arrays, i.e., linear antenna rows, are mounted on the ceiling of a hall, and the position of an object with a signal source, such as a BLE tag, is calculated using, for example, triangulation. Depending on the size of the hall, a corresponding number of antenna arrays must be installed.

[0103] With reference to Fig. Section 4 describes an example in which a hall is divided into four quadrants and, based on the determination of movement patterns described above, it can be determined in which quadrant a signal source, i.e., a sought-after object, is located.

[0104] It should be taken into account that in the example in Fig. 4. For the sake of simplicity, only four antenna arrays are used, but the invention is of course also applicable to examples with significantly more antenna arrays and correspondingly more areas. For the sake of clarity, in Fig. 4 For the antenna arrays for each gate G1 to G4, only two antennas are shown, whereby it can be understood from the preceding description that more than two antennas may be installed in an antenna array for each gate.

[0105] In the Fig. In the example shown in section 4, it can be determined from which quadrant a signal source moves to another quadrant when the signal source crosses a virtual gate.

[0106] If a signal source is near a gate or is moving past a gate on one side, the quadrant in which the signal source is currently located can also be determined.

[0107] For example, in Fig. 4. Assume that the antenna array at the top center defines a virtual gate G1. For gate G1, it is assumed that the right side of gate G1 is the "plus side" and the left side is the "minus side". If it is determined that a signal source traverses gate G1 in a minus-plus pattern, it can be determined that after traversing gate G1, the signal source will be located in the upper right quadrant. Fig. 4. Alternatively, it can also be determined that the signal source, after passing through gate G1, is located in the upper right quadrant. Fig. 4 is located when it is determined that the signal source has not crossed gate G1 and is on the plus side for the entire received packet sequence.

[0108] Additionally, a gate can be installed at the entrances to the warehouse, especially if there are multiple entrances, so that each signal source is assigned an area upon entering the warehouse. When the signal source then passes through another virtual gate in the warehouse, the area assigned to that signal source is updated.

[0109] Another application example is in Fig. 5 shown. Fig. Figure 5 illustrates an application of the present invention in a production facility with several production or storage halls. In this context, Fig. 5 Two halls are depicted (in this case a warehouse and a production hall). In the Fig. In the illustrated example 5, it is assumed that both halls have two entrances and two exits. Two possible routes to move from one hall to the other are shown with arrows.

[0110] Furthermore, in Fig.Figure 5 shows four antenna arrays or gates (G1 to G4). Thus, a gate is attached to each input and output. This allows any signal source moving through an input or output to be detected, and its corresponding movement pattern—the direction in which the signal source traverses the gate or input / output—to be determined. Such an application of the present invention makes it possible to track all parts equipped with a signal source and to determine, without significant effort and in real time, the precise location of a particular part or signal source.

[0111] In another example, antenna arrays can be installed in front of, after, and between rows of shelves, so that for each signal source it can be determined in which shelf or in which row of shelves a particular object or signal source is located.

[0112] In another example, a production site such as a factory might consist of several rooms and / or buildings. In such a case, gates equipped with antenna arrays can be installed at various passageways, doors, gates, and the like. These gates can then be used as checkpoints, allowing tracking when a signal source attached to an object or container is moved, for example, from one room to another.

[0113] Furthermore, a large number (e.g., 2, 3, 4, 5, or even more) of gates can be connected in series (e.g., at a checkpoint). This means that if almost all gates in the group detect a signal crossing, it can be assumed with a high degree of certainty that this detection is correct. However, if only one gate in the group detects a signal crossing, it can be assumed that this is, or could be, a measurement error.

[0114] According to the present invention, a computer-readable storage medium can also be provided on which instructions are stored which, when executed by a device, cause the device to perform the calculations and determinations of movement patterns of a signal source described above.

[0115] According to the present invention, a method for determining the movement patterns of a low-energy signal source in a monitored area is provided, wherein in the monitored area a virtually bounded part of a vertical surface is defined as a gate by a one-dimensional antenna array, the low-energy signal source emits an unmodulated sinusoidal signal, all antennas of the antenna array sequentially sample the signal for a sequence of a certain number of periodically offset sampling times, and the movement of the low-energy signal source with respect to the gate is tracked by monitoring, based on calculated phase shifts, whether the low-energy signal source crosses the gate and consequently in which direction the low-energy signal source crosses the gate.whether the low-energy signal source moves near the gate without crossing it, and consequently on which side of the gate the low-energy signal source is moving, and whether the low-energy signal source is near the gate without any movement being clearly detected.

Claims

[1] Method for determining motion patterns of a low-energy signal source in a monitored area, wherein - in the monitored area a virtually bounded part of a vertical surface is defined as a gate (G1, G2, G3, G4) by at least one one-dimensional antenna array (A1-An), and - the motion scheme of the low-energy signal source relative to the gate (G1, G2, G3, G4) is defined by the tendency of the change of an incidence angle (θ) from +90° to -90° or from -90° to +90° with transition through 0° at time t0, and - a movement of the low-energy signal source with respect to the gate (G1, G2, G3, G4) according to the motion scheme by means of at least one one-dimensional antenna array (A1-An) is tracked, by tracking, whether the low-energy signal source crosses the gate (G1, G2, G3, G4) and consequently in which direction the low-energy signal source crosses the gate (G1, G2, G3, G4), whether the low-energy signal source moves near the gate (G1, G2, G3, G4) without crossing the gate (G1, G2, G3, G4) and consequently on which side of the gate (G1, G2, G3, G4) the low-energy signal source moves, and whether the low-energy signal source is near the gate (G1, G2, G3, G4) without any movement being clearly detected, whereby - the procedure includes the following: Selecting representative samples from a plurality of calculated angles of arrival (θ) calculated from signals sampled at different sampling times, wherein the representative sample comprises calculated angles of arrival (θ) based on phases received at pairs of two adjacent antennas (An-An-1) of the at least one one-dimensional antenna array (A1-An), Calculating one statistical value for each representative sample as a statistically calculated angle of incidence, Selecting a time t0 in a period with high received signal strength by calculating a correlation coefficient P between a model of an ideal signal and a real signal for different possible times t of crossing the gate (G1, G2, G3, G4) in the period with high received signal strength, where time t0 is chosen as the time t with the highest value of the correlation coefficient P. Selecting a first subsequence of statistically calculated angles of incidence before time t0 and a second subsequence of statistically calculated angles of incidence after time t0, by selecting two subsequences that are as large as possible, provided that both subsequences are of equal size, Determine, based on the statistically calculated angles of incidence of the first subsequence and the statistically calculated angles of incidence of the second subsequence, whether the low-energy signal source has crossed the virtual gate (G1, G2, G3, G4) during the sequence. If it is determined that the low-energy signal source crossed the virtual gate (G1, G2, G3, G4) during a sequence comprising the first subsequence and the second subsequence, determine the direction in which the low-energy signal source crossed the virtual gate (G1, G2, G3, G4) by determining a probable motion pattern based on a tendency of a change in the angle of incidence (θ) from +90° to -90° with a transition through 0° at time t0, and If it is determined that the low-energy signal source did not cross the virtual gate (G1, G2, G3, G4) during the sequence, determine whether the low-energy signal source was located near the virtual gate (G1, G2, G3, G4) or on the side of the virtual gate (G1, G2, G3, G4) with a positive sign of the angle of incidence (θ) of the signal or on the side of the virtual gate (G1, G2, G3, G4) with a negative sign of the angle of incidence (θ) of the signal during the entire sequence, based on the statistically calculated angles of incidence of the entire sequence. [2] Method according to claim 1, wherein if it is determined that the low-energy signal source has not crossed the virtual gate (G1, G2, G3, G4) during the sequence, a difference is calculated between the number of statistically calculated angles of incidence with a positive sign and the number of statistically calculated angles of incidence with a negative sign, and The low-energy signal source is located on which side of the virtual gate (G1, G2, G3, G4) is determined based on the sign of the difference. [3] Method according to claim 2, wherein if the magnitude of the difference is less than a predetermined value, it is determined that the low-energy signal source has been in the vicinity of the virtual gate (G1, G2, G3, G4) throughout the entire sequence. [4] Method according to one of claims 1 to 3, wherein the correlation coefficient P is calculated based on a difference between the number of statistically calculated angles of incidence before time t0 with a positive sign and the number of statistically calculated angles of incidence before time t0 with a negative sign and based on a difference between the number of statistically calculated angles of incidence after time t0 with a negative sign and the number of statistically calculated angles of incidence after time t0 with a positive sign. [5] Method according to any one of the preceding claims, wherein in the monitored area several one-dimensional antenna arrays (A1-An) are installed, each with a plurality of antennas at a constant distance (d) from each other, on the basis of which a plurality of virtual gates (G1, G2, G3, G4) is defined, and for the low-energy signal source, in which the low-energy signal source is located in a virtual space limited by virtual gates (G1, G2, G3, G4). [6] Method according to any of the preceding claims, wherein a calculated phase shift is corrected and the angle of incidence (θ) of the sampled signal is calculated on the basis of the corrected phase shift. [7] Device for determining the motion patterns of a low-energy signal source in a monitored area, wherein - in the monitored area a virtually bounded part of a vertical surface is defined as a gate (G1, G2, G3, G4) by at least one one-dimensional antenna array (A1-An), - a motion scheme of the low-energy signal source relative to the gate (G1, G2, G3, G4) is defined by the tendency of the change of an incidence angle (θ) from +90° to -90° or from -90° to +90° with transition through 0° at time t0, and - the device is configured to track a movement of the low-energy signal source with respect to the gate (G1, G2, G3, G4) according to the motion scheme by means of the at least one one-dimensional antenna array (A1-An), by the device being configured to track, whether the low-energy signal source crosses the gate (G1, G2, G3, G4) and consequently in which direction the low-energy signal source crosses the gate (G1, G2, G3, G4), whether the low-energy signal source moves near the gate (G1, G2, G3, G4) without crossing the gate (G1, G2, G3, G4) and consequently on which side of the gate (G1, G2, G3, G4) the low-energy signal source moves, and whether the low-energy signal source is near the gate (G1, G2, G3, G4) without any movement being clearly detected, whereby the device is set up to: to select representative samples from a plurality of calculated angles of arrival, which are calculated from signals sampled at different sampling times, wherein the representative sample comprises angles of arrival calculated based on phases received at pairs of two adjacent antennas (An - An-1) of the at least one one-dimensional antenna array (A1-An), to calculate a statistical value for each representative sample as a statistically calculated angle of incidence, to select a time t0 in a period with high received signal strength by calculating a correlation coefficient P between a model of an ideal signal and a real signal for various possible times t of crossing the gate (G1, G2, G3, G4) in the period with high received signal strength, wherein the device is configured to select as time t0 the time t with the highest value of the correlation coefficient P, to select a first subsequence of statistically calculated angles of incidence before time t0 and a second subsequence of statistically calculated angles of incidence after time t0, by selecting two subsequences that are as large as possible, provided that both subsequences are of equal size, to determine whether the low-energy signal source has crossed the virtual gate (G1, G2, G3, G4) during the sequence, based on the statistically calculated angles of arrival of the first subsequence and the statistically calculated angles of arrival of the second subsequence, If it is determined that the low-energy signal source has crossed the virtual gate (G1, G2, G3, G4) during a sequence comprising the first subsequence and the second subsequence, to determine in which direction the low-energy signal source crossed the virtual gate (G1, G2, G3, G4) by determining a probable motion pattern based on a tendency of a change in the angle of incidence (θ) from +90° to -90° with a transition through 0° at time t0, and wherein If it is determined that the low-energy signal source did not cross the virtual gate (G1, G2, G3, G4) during the sequence, the device is further configured to determine whether the low-energy signal source was located near the virtual gate (G1, G2, G3, G4) or on the side of the virtual gate (G1, G2, G3, G4) with a positive sign of the angle of incidence (θ) of the signal or on the side of the virtual gate (G1, G2, G3, G4) with a negative sign of the angle of incidence (θ) of the signal during the entire sequence, based on the statistically calculated angles of incidence of the entire sequence. [8] Device according to claim 7, wherein if it is determined that the low-energy signal source has not crossed the virtual gate (G1, G2, G3, G4) during the sequence, the device is further set up to to calculate the difference between the number of statistically calculated angles of incidence with a positive sign and the number of statistically calculated angles of incidence with a negative sign, and to determine, based on the sign of the difference, on which side of the virtual gate (G1, G2, G3, G4) the low-energy signal source is located. [9] Device according to claim 8, wherein if the magnitude of the difference is less than a predetermined value, the device is configured to determine that the low-energy signal source has been in the vicinity of the virtual gate (G1, G2, G3, G4) throughout the entire sequence. [10] Device according to one of claims 7 to 9, wherein the device is configured to calculate the correlation coefficient P based on a difference between the number of statistically calculated angles of incidence before time t0 with a positive sign and the number of statistically calculated angles of incidence before time t0 with a negative sign and based on a difference between the number of statistically calculated angles of incidence after time t0 with a negative sign and the number of statistically calculated angles of incidence after time t0 with a positive sign. [11] Device according to any one of claims 7 to 10, wherein the device is configured to correct a calculated phase shift and to calculate the angle of incidence (θ) of the sampled signal on the basis of the corrected phase shift. [12] System comprising a plurality of devices according to any one of claims 7 to 11. [13] Computer program with instructions which, when executed by a device, cause the device to execute the method according to any one of claims 1 to 6. [14] Non-transitory computer-readable storage medium, with instructions stored thereon which, when executed by a device, cause the device to execute the method according to any one of claims 1 to 6.

Images