How to determine the transmitter location

Abstract

To determine a location of a transmitter in an environment having at least one receiver, the present disclosure provides for obtaining geometric information related to the environment to represent a 3D model, receiving signals at the receiver and measuring at least one angle of arrival "AoA" of the signals, calculating statistical properties of the AoA based on the measurements, using the 3D model and the statistical properties of the AoA to obtain paths with associated respective weights, and using the paths with associated weights to estimate the location of the transmitter.

Description

[Technical Field]

[0001] The present disclosure relates to a positioning method and a positioning device. [Background technology]

[0002] Standard positioning systems rely on so-called "line-of-sight" (LOS) measurements. In the example shown in Figure 1, where three base stations (BSs) estimate the angle of arrival (AoA) of a reference signal transmitted by a user equipment (UE), a typical triangulation method can be implemented as follows: The server, knowing the locations of the two BSs and the measured AoAs, finds the intersection of three rays originating from the BSs in the direction of the measured AoAs. The estimated location of the UE is this intersection.

[0003] However, the signal received at the BS may be "non line-of-sight" (NLOS). The signal may have been reflected before reaching the BS. In this case, the measured AoA indicates the direction of the last reflecting object, not the actual direction of the UE. As a result, the standard triangulation method described above is not sufficient for positioning operations.

[0004] One solution to address this problem is to use knowledge of the environment to refine the positioning operation. For example, in an offline phase, a database of measurements corresponding to each possible position of the UE can be created (at the BS). Then, in an online phase, measurements are performed and the most similar measurement in the database is searched for. Once the most similar measurement is found, the position corresponding to this measurement is the estimated position. This class of methods is called "fingerprinting". The database can be established via actual measurements or using a digital twin of the environment combined with ray tracing.

[0005] Alternatively, rather than using a database, online ray tracing can be considered: given the measured AoAs, ray tracing is performed in the direction of these AoAs. The intersections of the rays then become the estimated positions. This technique is called "inverse ray tracing" and its principle is shown in Figure 2. Alternatively, geometric equations can be considered instead of traditional ray tracing.

[0006] Of course, measurements are not perfectly accurate, and the measured AoA may be corrupted by some noise. Some rays may be emitted within an interval centered on the measured AoA. The width considered for this interval is twice the estimated standard deviation of the estimated AoA values. A least-squares solution assuming independent and identically distributed Gaussian noise for the estimated AoA can also be used.

[0007] In all these known techniques, the statistics of the measurements are not fully taken into account in conventional ray tracing aided positioning.

[0008] Performance may not be optimal because useful information is not taken into account. Summary of the Invention [Problem to be solved by the invention]

[0009] The present disclosure aims to improve this situation. [Means for solving the problem]

[0010] Therefore, the present disclosure provides a method for determining a location of a transmitter in an environment comprising at least one receiver, the method comprising: Obtaining geometric information related to the environment to represent a 3D model; receiving a signal at a receiver and measuring at least one angle of arrival "AoA" of the signal; Based on the measurements, Calculating statistical properties of the AoA; and using statistical properties of the 3D model and the AoA to obtain paths with associated respective weights; estimating the location of the transmitter using the paths with associated weights; We propose a method including:

[0011] The above mentioned environments and signals can relate to any kind of waves, i.e. radio frequency waves, or sound waves, or light waves, etc. Therefore, measurements can be made by radio frequency antennas, or any 2D or 3D optical sensor (e.g., a coupled camera), or any microphone array.

[0012] In one embodiment, obtaining paths with associated weights includes: obtaining samples of the AoA of the signal based on a distribution defined by a statistical property of the AoA; calculating a path using the 3D model, the path starting from the receiver at an angle defined by the sample; Associating the same weight to all paths; may include:

[0013] Alternatively, obtaining the paths with associated weights may be done by: calculating a path using the 3D model, the path starting from the receiver and extending at an angle within a predefined range; calculating each weight associated with a path based on statistical characteristics of the AoA; may include:

[0014] The predetermined range can be 360 ​​degrees (e.g., discretized in 5 degree increments) when there is no prior knowledge of the possible wave arrivals, or it can be 180 degrees when the sensor is attached to a supporting surface, or it can be 90 degrees (when the sensor is typically placed in a corner of a room), or something else.

[0015] In one embodiment, the method may include estimating a location of the transmitter, the estimating comprising: At least two AoA measurements are performed, and for each AoA measurement: Obtaining paths with respective weights associated therewith; position k , k=1,...,K, position X k The area that includes

number

number

number

[0016] In this embodiment, the calculated probability is

number

[0017] In an alternative embodiment, the calculated probability is: region

number

number

[0018] In this alternative embodiment, the above values ​​are k and paths and regions

number

[0019] Alternatively, the above values ​​can be expressed as X k and an X on the path k The weight of the path can depend on the distance between the orthogonal projection of .

[0020] Also, determining the estimated transmitter location based on the calculated probability may be performed by: Each position k , the value β from the calculated probabilities associated with each AoA. k and Value β k Position X where the maximum is reached k By finding the transmitter's position,

number

[0021] Alternatively, determining an estimated transmitter location based on the calculated probability may include: Each position k , the value β from the calculated probabilities associated with each AoA. k and Transmitter location

number

number

[0022] The present disclosure is also directed to a computer program comprising instructions, which when executed by a processor cause the computer program to perform the method defined above.

[0023] The present disclosure is also directed to a device for determining the location of a transmitter in an environment comprising a plurality of receivers each having a known location, the device comprising an interface for receiving measurements made by the receivers and a processor for implementing the method defined above.

[0024] Further details are provided in the remainder of this specification with reference to the accompanying drawings. [Brief explanation of the drawings]

[0025] [Figure 1] FIG. 1 illustrates the principle of positioning based on angle of arrival measurements in the line of sight (LOS) case. [Figure 2] FIG. 1 illustrates the use of ray tracing to improve positioning. [Figure 3] 1 shows an example of an environment with three receivers Rx (for example three base stations BS in a radio frequency environment). [Figure 4] 1 illustrates a transmitter Tx of a location target that transmits a signal (eg, a radio frequency signal). [Figure 5] FIG. 2 illustrates steps taken by a receiver to measure angle of arrival. [Figure 6] FIG. 1 illustrates the use of the environment geometry to obtain paths, each path having a different weight depending on the statistical properties of its corresponding AoA. [Figure 7] FIG. 1 illustrates estimated transmitter positioning. [Figure 8]FIG. 1 shows an example of an embodiment of a device such as a receiver Rx that executes the method presented above. [Figure 9] FIG. 1 illustrates an example of the main steps of a method according to the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0026] In the following description, we consider a scenario in which a transmitter (i.e., Tx), in this example a user equipment (UE), transmits positioning signals in a radio frequency environment to one or several receivers (Rx), such as base stations. The goal of the base station is to determine the position of the UE with the aid of a 3D model of the environment (e.g., via ray tracing).

[0027] Therefore, the present invention aims to improve positioning in wireless systems by utilizing not only the geometric properties of the environment (e.g., by ray tracing simulation) but also the statistical properties of the measurements. As shown in Figure 9, the method can include the following steps:

[0028] In step S1 of Figure 9, measurements are made by at least one Rx receiving the signal transmitted by Tx, more specifically measuring the angle of arrival (AoA) of the signal.

[0029] Additionally, 3D scene data (i.e., a "3D model") is obtained that characterizes (e.g., by ray tracing) the geometric properties of the considered environment.

[0030] In step S2, statistical properties of the AoA are calculated from the measurements.

[0031] In step S3, paths with associated weights are calculated using the statistical properties of the AoA and the 3D model. One of the following techniques can be considered: The exit angles of the paths are generated based on the statistical properties of the corresponding AoA described above. The 3D scene characterizes the remaining aspects of each path. The weights assigned to each path are the same for all paths. Or, ● The exit angles of the paths are generated uniformly. Then, the 3D scene characterizes the remaining aspects of each path, and the weights assigned to each path are calculated according to the statistical characteristics of the corresponding AoA.

[0032] In step S4, the probability of the transmitter's location is calculated based on the weighted paths obtained.

[0033] Next, in step S5, the location of the transmitter is estimated.

[0034] The above calculations are detailed below.

[0035] X denotes a random variable representing the transmitter location, Θ denotes a random variable representing the AoA (Θ varies in the range [0,2π] in two dimensions and in the range [0,π] × [-π,π] in three dimensions), and Y denotes a random variable representing the measurement. The statistical model of the AoA measurement can be expressed as the following equation:

number

[0036] Regarding these notations, for simplicity, p(θ|y) is shorthand for P(Θ=θ|Y=y), and similarly, p(x|y) is shorthand for p(X=x|Y=y). The variable y here relates to the measurement, and an instance of θ relates to one of the AoA samples mentioned above.

[0037] In the considered problem, at least two AoA measurements are required to estimate the transmitter's position. Let n measurements y = [y1, y2, ..., y n] and a 3D model (denoted as ψ), the goal is to obtain p(x|y,ψ). For simplicity, we assume there is one measurement per receiver and there are n receivers.

[0038] In the following, a ray is defined as a path characterized by the 3D scene, starting at an angle θ from a given position. Here, the starting position is the receiver's position (which is completely known). Let "i" refer to the ith receiver, where i=1, 2, ..., n. Based on this, we will use the notation path(θ,i) to refer to the ray or path starting at an angle θ from the ith receiver. Regrouping the group of paths from n receivers, we obtain the following equation:

number

[0039] Of course, there are many such groups path(θ) because a single receiver i typically emits multiple rays at different angles θ. As a result, there are many paths(θ,i). The set of possible angles θ = [θ1,θ i ,...,θ n ] is denoted by C. In that case, the probability p(x|y,ψ) is expressed as

number

[0040] The second term p(θ|y) in equation (3) corresponds to the statistics of the AoA based on the measurements. Thus, the quantity p(θ|y) relates to the statistical properties of the calculated AoA described above. When they are independent,

number

[0041] Given a 3D model ψ and an angle θ, the path path(θ) can be obtained. In this step, a ray tracing algorithm or manual calculation can be used. The first term in equation (3) can be rewritten as:

number

[0042] In words, the probability in equation (4) indicates the probability that the position x is the location of the transmitter given the path in the environment. A first option, which leads to a simple implementation, is to define the position as a square (or a cube in the 3D case)

number

number

[0043] The second option provides a more sophisticated statistical model of p(x|path(θ)) and is shown below:

number

number

number

[0044] As a result, to calculate equation (3) via ray tracing, one can proceed with the following steps (these steps can usually be defined in an algorithm of a computer program): 1. For each of the n receivers, uniformly discretize C = [0, 2π) and obtain l rays (i.e., one ray for each θ∈C). Each receiver's probability p(θ i |y i ) 2. Emit n × l rays in the discretized directions. 3. Different areas of the scene

number

number

number

number

number

[0045] An alternative embodiment for calculating equation (3) is via Monte Carlo sampling. Typically, a ray is calculated by i |y i ) can be directly sampled (i.e., selected) according to . As a result, the estimation algorithm above can be modified (especially in items 1 and 3.b) as follows: 1. For each of the n BSs, p(θ i |y i ) and sample l rays according to 2. Shoot n × l rays in the sampled directions. 3. Different areas of the scene

number

number

number

[0046] Therefore, to generalize the principles of these two algorithms, we can introduce the concept of a "weighted path." A weighted path is a set of measurements (e.g., p(θ|y i The pair {path(θ),weight} is generated according to the path distribution calculated from (θ), where weight refers to the amount by which the combination of rays is weighted in step 3b of the first algorithm.

[0047] In step 4 of both methods above, the estimated position x M is expressed as follows: k x k It can also be calculated by marginalizing

number

[0048] The distribution of paths can be used to generate the AoA θ (as step 1 of the second algorithm) and / or the weights (as step 3b of the first algorithm).

[0049] In this respect, a generalization of the principles presented above can be defined as follows:

[0050] In a 3D model characterizing a path between a transmitter Tx and at least one receiver Rx, the location of Tx is given by: A. Taking measurements on Rx and B. Calculating the distribution of light rays from the measurements; C. Generating weighted rays according to the calculated distribution; It can be calculated by:

[0051] In a first embodiment, step C above can be performed as follows: i , the interval C=[0,2π) is uniformly discretized, and for each measurement, i rays (one ray for each θ∈C). The selected ray is then emitted.

[0052] An estimate of the Tx's location can then be made as follows: I.Area

number

number

number

[0053] Alternatively, step C above can be performed as follows. That is, n measurements y i For each of p(θ|y i ) according to i rays are sampled. Then, all sampled rays are emitted.

[0054] An estimate of the Tx's location can then be made as follows: ● Area

number

[0055] Of course, the above method can use multiple receivers to make measurements. Typically, for a particular AoA measurement, p(θ i |y i ) can be refined using several "sub-measurements" (for example, made by a panel of antennas at the receiver). In this case, the measurement y i is y i ={y i 1 ,y i 2,...} Alternatively, a single receiver can be used, which can be mobile to make measurements according to different AoAs.

[0056] In practice, it is not always possible to perfectly model the digital twin of a scene. Usually, the positions of objects and their features are not known exactly and / or the ray tracing algorithm may ignore some physical aspects. In these cases, more complex statistical models need to be established to take into account the errors in the 3D model.

[0057] Digital twin with errors

number

number

number

[0058] Below are examples of the various situations discussed above.

[0059] Typically, when considering a simple environment with, for example, only one obstacle and four walls, and three receivers are deployed as shown in the example of Figure 3 (the positions of these Rx are completely known), this environment is fully modeled and a 3D model is obtained, including the geometric properties of the walls and obstacles, such as their positions, orientations, and dimensions. It may be necessary to locate a transmitter with an unknown position in this environment. This transmitter Tx, shown in the example of Figure 4, transmits a radio signal.

[0060] Each receiver Rx receives this radio signal and measures the angle of arrival (AoA) based on the received signal. Because obstacles in the environment may block the direct path between the first receiver BS1 and the transmitter UE, the AoA measured at Rx BS1 is along the direction of the upper wall, while Rx BS2 and Rx BS3 measure the AoA of the direct path between their BSs and the Tx UE (as shown in Figure 5). However, these measured AoAs may be in error (due to noise and / or interference, etc.).

[0061] Typically, in this situation, according to the embodiment presented above, each receiver can calculate the statistical characteristics of its AoA based on its own receiver configuration. As shown in FIG. 8, the receiver Rx configuration affecting the measurement can be, for example, an antenna system AS coupled to a processing circuit including a processor PROC and a memory MEM through an interface IN, as presented above. The memory stores at least computer program instructions according to the present disclosure. The algorithm starts with an estimate of the AoA, as presented above, and then proceeds to estimate the statistical characteristics of the measurement according to a distribution density function or particle distribution with associated parameters, etc. Next, using a 3D model of the environment and the measured AoA together with the associated statistical characteristics, it is possible to trace back the path the signal takes to reach the receiver Rx. Taking receiver BS1 as an example, assume that Rx BS1 obtains an AoA with the following statistics: ○ AoA=43°, 40% (i.e., p(θ=43°|y i )=0.4), ○ AoA=44 degrees, 20% ○ AoA=45 degrees, 10% ○ AoA=46 degrees, 10% ○ AoA=47 degrees, 10% ○ AoA>47 degrees, 5%, AoA<43 degrees, 5%.

[0062] Ray tracing simulation can be used in the 3D model for these five rays, each starting from Rx BS1 at one of the angles defined above, as shown in Figure 6.

[0063] Combining the paths from Rx BS1, Rx BS2 and Rx BS3 with their associated weights, the estimated location of the Tx UE can be calculated as shown in FIG.

Claims

1. 1. A method for determining a location of a transmitter in an environment comprising at least one receiver, the method comprising: obtaining geometric information related to the environment to represent a 3D model; receiving a signal at said receiver and measuring at least one angle of arrival "AoA" of said signal; Based on the measured values, calculating statistical characteristics of the angles of arrival; using the 3D model and statistical characteristics of the angles of arrival to obtain paths each with an associated weight; estimating a location of the transmitter using the paths with associated weights; and A method comprising:

2. Obtaining the paths with associated weights includes: obtaining samples of angles of arrival of the signals based on a distribution defined by a statistical characteristic of the angles of arrival; calculating a path using the 3D model, the path starting from the receiver at an angle defined by the sample; Associating the same weight to all paths; The method of claim 1 , comprising:

3. Obtaining the paths with associated weights includes: calculating the path using the 3D model, the path starting from the receiver and extending at an angle within a predefined range; calculating a respective weight associated with the path based on statistical characteristics of the angles of arrival; The method of claim 1 , comprising:

4. estimating the location of the transmitter; The estimating step comprises: measuring at least two of the angles of arrival, and for each of the angles of arrival measurements: obtaining paths with respective associated weights; Position X ｋ , k = 1 , . . . , K, the position X ｋ The area that includes [Equation 1] and The area [Equation 2] Collecting all paths that cross The position X for the angle of arrival measurement under consideration ｋ The probability associated with the region [Equation 3] based on the paths traversing the path and the respective weights associated with those paths; determining an estimated location of the transmitter based on the calculated probability associated with each angle-of-arrival measurement; The method according to any one of claims 1 to 3, comprising:

5. The calculated probability is [Equation 4] 5. The method of claim 4, wherein the weights are given by the product of all weights associated with paths traversing the vector.

6. The calculated probability is: The area [Equation 5] calculating a value associated with each path that traverses X ｋ and the path; The area [Equation 6] calculating all products of values ​​associated with paths traversing The method of claim 4, wherein the calculated value is:

7. The value is the position X ｋ and the path and the area [Equation 7] 7. The method of claim 6, wherein the weight of the path depends on the distance between the intersection of the first and second paths.

8. The value is the position X ｋ and the position X on the path ｋ 7. The method of claim 6, wherein the weight of the path depends on the distance between the orthogonal projection of the vector x and the vector y.

9. Determining an estimated location of the transmitter based on the calculated probability includes: Each said position ｋ , the value β ｋ and The value β ｋ Position X where the maximum is reached ｋ by determining the location of the transmitter [Equation 8] and estimating The method according to any one of claims 4 to 8, comprising:

10. Determining an estimated location of the transmitter based on the calculated probability includes: Each said position ｋ , the value β ｋ and the location of the transmitter [Equation 9] of, [Equation 10] and The method according to any one of claims 4 to 8, comprising:

11. A computer program comprising instructions which, when executed by a processor, cause the computer program to perform the method of any one of claims 1 to 10.

12. A device for determining the location of a transmitter in an environment comprising a plurality of receivers, each of whose location is known, said device comprising an interface for receiving values ​​of measurements made by said receivers, and a processor for implementing the method of any one of claims 1 to 10.

Images