Probabilistic ray tracing-assisted positioning

Abstract

This invention relates to a method and apparatus for determining a piece of data representing the location of a user equipment within an environment comprising a set of base stations. According to the invention, the method comprises the following steps: - obtaining (S01) a vector of N angles of reception of a signal emitted by the user equipment, each angle of the vector corresponding to the angle of reception of the signal at a given base station in the set of base stations; - obtaining (S02) N probability distributions, each probability distribution being obtained as a result of light emission from a 3D representation of the environment; - calculating (S03) at least one probability of the location of the user equipment, and said at least one probability being a function of the N probability distributions and the vector of the N angles.

Description

Technical Field

[0001] This disclosure relates to positioning methods and apparatus. This application claims priority to European Patent Application No. EP23306926.9, filed on November 8, 2023, the contents of which are incorporated herein by reference. Background Technology

[0002] Standard positioning systems rely on so-called " Sight distance ( line-of-sight LOS (Loss of Memory) measurement. As an example, such as... Figure 1 As shown, if three base stations (“BS”) estimate the angle of arrival (AoA) of a reference signal transmitted by a user equipment (“UE”) such as a telecommunications terminal, a conventional triangulation method can be implemented as follows: A server knowing the positions of two BSs and the measured AoA searches for the intersection of three light rays leaving the BSs in the direction of the measured AoA. The estimated position of the UE is this intersection.

[0003] However, the signal received at BS may be " Non-line-of-sight ( non line-of-sight (NLOS). The signal may have been reflected before reaching the BS. In this case, the measured AoA does not indicate the UE's true direction, but rather the direction of the last reflected object. Therefore, the standard triangulation described above is insufficient for positioning operations.

[0004] One solution to this problem is to use online ray tracing: given measured AoA, rays are traced along those AoA directions. The intersection of the rays is the estimated location. This method is called "reverse ray tracing," and its principle is as follows: Figure 2 As shown. Alternatively, geometric equations can be considered instead of traditional ray tracing.

[0005] Of course, the measurement may not be perfectly accurate, and the measured AoA may be affected by some noise. Several rays can be emitted within the interval surrounding the measured AoA. The width of the interval considered is twice the estimated standard deviation of the AoA value. Least squares solution, assuming iid. Gaussian noise on the estimated AoA can also be used.

[0006] However, among all these previously known techniques, traditional ray tracing-assisted localization does not adequately consider the statistical data of measurements, and its performance cannot be optimal. To address this, a method for determining the transmitter's location has been previously proposed, in which the statistical properties of the AoA are calculated, and a 3D model of the environment is used to obtain some weighted paths, which are then used to determine the transmitter's location.

[0007] However, while relatively effective, this previously proposed method requires emitting a considerable amount of light during the online Monte Carlo light emission phase within a 3D model of the environment. This implies that a large amount of resources must be available, and the base station must possess these resources, which is typically not the case. Summary of the Invention

[0008] This disclosure aims to improve this situation.

[0009] Therefore, this disclosure proposes a method for estimating the location of a transmitter in an environment including at least one receiver. More specifically, a method is disclosed for determining a piece of data representing the location of a user equipment in an environment including a set of base stations, the method comprising the steps of: - Obtain a vector of N angles representing the reception of the signal emitted by the user equipment, each angle of the vector corresponding to the reception angle of the signal at a given base station in the set of base stations. - Obtain N probability distributions, each derived from the results of light emission in a 3D representation of the environment. - Calculate at least one probability of the location of the user equipment, and the at least one probability is a function of the vector of the N probability distributions and the N angles.

[0010] According to this disclosure, the step of obtaining the N probability distributions includes loading a set of probability distributions from a data structure.

[0011] According to this disclosure, the step of obtaining the N probability distributions further includes selecting a probability distribution set for each receiving angle of the signal from the probability distribution set.

[0012] According to this disclosure, the step of obtaining the N probability distributions includes calculating a set of probability distributions for each base station based on the received virtual angle.

[0013] According to this disclosure, the step of calculating the probability distribution density set for each base station based on the virtual reception angle includes at least one iteration of the following steps: - Determine a virtual reception angle for the signal from multiple possible reception angles. - Calculate a probability distribution based on the virtual receiving angle. - Store the probability distribution associated with the virtual angle in a data structure.

[0014] According to this disclosure, the step of calculating the probability distribution based on the virtual receiving angle includes: - At least one ray is emitted in a 3D representation of the environment using a Monte Carlo algorithm, wherein the AoD of the at least one ray is sampled according to a virtual angle. - By emitting at least one ray in a 3D representation of the environment, a map of points is constructed, wherein the at least one ray passes through a plane at a predetermined height in the 3D representation of the environment at the point. - Calculate the parameter distribution of the points in the graph, and pass the probability distribution associated with the virtual 3D angle.

[0015] According to this disclosure, the step of calculating the position of the user equipment as a function of the vectors of the N probability distributions and the N angles includes: - The global probability distribution is calculated by multiplying the N probability distributions together. - Determine the maximum value representing the most likely location of the user equipment based on the global probability distribution.

[0016] According to this disclosure, the steps to obtain N probability distribution sets include: for each angle of a vector of N angles, - At least one ray is emitted in a 3D representation of the environment using a Monte Carlo algorithm, wherein the AoD of the at least one ray is sampled according to the angle. - A map of points is constructed using the at least one ray emitted in the 3D representation of the environment, wherein the at least one ray passes through a plane at a predetermined height in the 3D representation of the environment at the point. - Calculate the parameter distribution of the points in the graph, and pass the probability distribution associated with the angle.

[0017] This disclosure also relates to a computer program that includes instructions that, when executed by a processor, cause the above-described method to be implemented.

[0018] This disclosure also relates to an apparatus for determining the location of a transmitter in an environment comprising a plurality of receivers, each receiver having a known location, the apparatus comprising an interface for receiving measurements performed by the receivers and a processor for implementing the method described above. Attached Figure Description

[0019] More details are presented in the following instructions with reference to the accompanying drawings, in which: - Figure 1 This illustrates the positioning principle based on the measurement of the angle of arrival in the case of line of sight (LOS). - Figure 2 This demonstrates the use of ray tracing to enhance positioning. - Figure 3 An example of an environment including three receivers Rx (e.g., three base stations BS in a radio frequency environment) and obstacles is shown. - Figure 4The transmitter Tx, which is to be located and transmits signals (e.g., radio frequency signals), is shown. - Figure 5 The steps for the receiver to perform an angle of arrival measurement are shown. - Figure 6 The diagram illustrates how paths are obtained using the geometry of the environment, with each path having a different weight based on its corresponding AoA statistical property. - Figure 7 The estimated transmitter location is shown. - Figure 8 The steps of the positioning method according to this disclosure are shown. - Figure 9 The diagram shows a spot plot obtained when a ray is emitted from a BS given AoA and given error statistics. The ellipse represents the fitted GMM (Gaussian mixture model) on the spot plot. - Figure 10 The cumulative density function of the positioning error is shown for both cases with fewer and more emitted rays. - Figure 11 An example of an implementation of an apparatus for performing the above method is shown. Detailed Implementation

[0020] The following description considers a scenario where a transmitter (or Tx) (such as a user equipment (UE) in a radio frequency environment in this example) sends location signals to one or more receivers (Rx) (such as a base station). The base station aims to locate the UE with the aid of a 3D model of the environment (e.g., via ray tracing). The geometric properties of the environment are then intended to be utilized (e.g., simulated via ray tracing). Statistical properties of measurements can also be used to enhance positioning in wireless systems.

[0021] As previously mentioned, the former approach allows for partial resolution of the user equipment (UE) location problem by using a model of the environment and identifying the UE's location through reverse ray tracing. These methods are useful in situations where the UE cannot transmit its own location (e.g., closed or partially closed environments, lack of GPS signals, etc.). More specifically, these methods use uplink angle of arrival (AoA) measurements of signals transmitted by the UE and employ a three-dimensional digital twin of the environment, as detailed below. Rays are emitted in a Monte Carlo manner based on AoA statistics, enabling the generation of a point map for each BS. These point maps are then combined, representing the intersections of the rays with the XY plane at a given UE height.

[0022] Figure 3 , Figure 4 and Figure 5The principle of AoA measurement is illustrated in a schematic scenario. Typically, when considering a simple environment, such as a single obstacle and four walls, ... Figure 3 The example shown deploys three receivers (the locations of these Rxes are known), thus allowing for perfect environment modeling to obtain a 3D model containing geometric properties such as location, orientation, and wall and obstacle dimensions. It may be necessary to locate transmitters whose locations are unknown in the environment. These transmitters, Txes, transmit radio signals, such as... Figure 4 As shown in the example, each receiver Rx receives a radio signal and measures the angle of arrival (AoA) based on the received signal. Because obstacles in the environment 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 BS3 measure the AoA of their direct paths with the Tx UE (as shown in the example). Figure 5 (As shown). However, the measured AoA may be incorrect (due to noise and / or interference, etc.). Using the AoA and a 3D representation of the environment, Monte Carlo ray tracing is performed to determine the UE's position, such as... Figure 6 and Figure 7 As shown. Taking into account error, the angle of arrival (AoA) becomes the angle of departure (AoD) of the ray. For example, a ray tracing simulation can be used for a 3D model with five rays, where each ray starts at Rx BS1, and some angles are defined as a function of error, such as... Figure 6 As shown. By combining the path with the association weights from Rx BS1, Rx BS2, and Rx BS3, the estimated location of Tx UE can be calculated, as follows. Figure 7 As shown. For example, considering a predetermined altitude, the intersection of the light ray and the plane at that altitude provides a point map for each BS. The point maps of each BS are combined, and this allows for obtaining a combined point map, the location of the UE being determined by the coordinates of the combined point map containing the most significant number of points from each BS.

[0023] While providing fairly good results, this method has proven to require emitting a large amount of light based on AoA statistics to obtain good accuracy in the actual position of the user equipment. This may require increased power or other resource consumption depending on operational implementation conditions, which could make it undesirable or unavailable.

[0024] Therefore, the inventors proposed fitting a parameter distribution (such as a Gaussian mixture model (GMM)) to each point map of each base station. For a given XY plane, the parameter distribution is a two-dimensional multivariate distribution. The probability density function (also known as pdf) obtained for each BS is multiplied, making it possible to calculate the location probability of the UE. This approach produces an algorithm robust to reduced ray emission, meaning fewer rays are emitted while maintaining the same level of results, or allowing for limited computational resources to be used, as disclosed below.

[0025] Therefore, for a given user equipment, combined with Figure 8 The proposed method described includes the following steps: - Obtain (S01) the vector of N angles of the received signal transmitted by the user equipment. Each angle of the vector corresponds to the angle at which the signal is received at a given base station in a set of base stations; - Obtain (S02)N probability distributions, each probability distribution being obtained from the result of light emission in the 3D representation of the environment; - Calculate (S03) at least one probability of the location of the user equipment, and use it as a function of the vector of the N probability distributions and the N angles.

[0026] Depending on the situation, the angles can be two-dimensional or three-dimensional. The N angles of the vector can be obtained from N corresponding base stations. In one variation, a base station may be able to capture two or more independent reception angles of a signal, and therefore may be able to transmit multiple reception angles of a single signal.

[0027] These probability distributions (sets) based on parameter distributions can be fitted and stored offline, thus avoiding light emission during the online phase and reducing the amount of resources (power, time) required to locate the user device. In fact, during the online phase, for a given measurement AoA, the corresponding parameters of the distribution are directly recovered from the stored table, and signal processing is then performed without emitting light. This significantly reduces the computational complexity of the positioning method.

[0028] Therefore, the proposed method can be used online ( That is, real time This means the parameter probability distribution is calculated in real time (i.e., after obtaining the AoA measurement from the user equipment). The proposed method can also be implemented partially offline, meaning the parameter probability distribution is calculated and stored offline based on a predetermined set of parameters (the number of emitted rays, the selected AoA, and a preset error for the AoA). The storage can be accessed from the device implementing the online portion of the method (e.g., one of the base stations, a central station), and the parameter probability distribution, along with the current AoA measurement of the user equipment signal, can be used online to locate the user equipment.

[0029] Regardless of whether the method is performed entirely online or partially online, it includes the following: In at least one example, fitting a parametric probability distribution, such as a Gaussian mixture model (GMM), to the point plot obtained for each BS. Figure 9 A dot plot and a corresponding example of a fitted GMM with four clusters are shown. The plot illustrates the results of ray emission in a Monte Carlo manner for a given base station (BS), with predetermined AoA measurements and predetermined AoA error statistics (provided as parameters). Ellipses represent the four clusters, and the distance between ellipses within a given cluster represents the dispersion of the distribution. Within a given cluster, the closer the distance, the more accurate the point's location. The probability of the UE being at the location of the measurement performed for each BS is then obtained by multiplying by the probability density function (pdf) for each BS.

[0030] The disclosed method can be implemented in two different modes: In the first mode, the light rays are emitted in the online phase, and the distribution is also calculated in the online phase: the position probability is obtained by multiplying the probability density functions of each BS. This provides the advantage of being more robust to a reduced number of emitted light rays in the online phase.

[0031] The second mode consists of an offline phase and an online phase: First, for each BS and for the set of discrete AoA, rays are emitted to obtain a point map for each discrete AoA. A parameter distribution is fitted and the parameters are stored. Then, in the online phase, AoA is measured for each BS, and the corresponding distribution parameters in the stored table are recovered. These distributions are then used as in the first implementation mode. This provides the advantage of not requiring ray emission in the online phase.

[0032] The implementation methods of the disclosed method will now be fully described.

[0033] For simplicity, a general description of the positioning device based on the reverse ray tracing positioning method is provided. In this context, the altitude of the UE to be located is known or estimated. According to the general method, given a measured AoA at a BS... The process involves emitting light rays from the BS along this direction, where the measured AoA is converted into the ray's departure angle (AoD), and the intersection of the projected ray with the plane provides a dot plot for a given BS. This process is repeated for all BSs, and the intersection of the rays (on the dot plot) is the estimated position of the UE.

[0034] If the AoA measurement is noisy, the following steps can be taken: For each BS, sample the angle of the emitted light ray based on the statistics of the AoA measurement error. This method is called the Monte Carlo method. In short, the result of this light emission is a set of points on the xy-plane at the UE's height. These points correspond to the locations where the emitted light ray crosses the xy-plane, providing a point map. As explained earlier, this method requires emitting a large number of rays to obtain good results.

[0035] As explained earlier, in order to avoid having to fire a large number of points, it is proposed to group the points on the graph by clusters and assign probabilities to these points in practice.

[0036] The following paragraphs provide a precise explanation of how the proposed method is implemented.

[0037] Statistical modeling and symbolic representation

[0038] Let X be a random variable representing the location of the UE. Let AoA be the real random variable representing the signal. Let be a random variable representing the BS measurement of AoA.

[0039] Regarding symbols, for ,use Similarly, for ,use .distributed Statistics representing uplink AoA measurement errors, for example, .

[0040] set up To measure the number of times, such that .

[0041] vector This includes measurements performed by all BSs. Therefore, for simplicity, we assume there is one measurement per BS. BS.

[0042] In the problem under consideration, the objective is to calculate (That is, considering the probability that position is x when y is y).

[0043] Probability distribution fitting

[0044] To address accuracy issues when dealing with a limited number of rays, it is proposed to fit a 2D (two-dimensional) probability distribution onto a point map obtained when rays are emitted from a single base in a Monte Carlo manner. Recall that parametric probability density estimation involves selecting a common distribution and estimating the parameters of the density function based on a sample of data. Nonparametric probability density estimation involves estimating the probability distribution as a histogram via techniques such as binning.

[0045] The technique disclosed herein involves fitting a distribution to the graph obtained from each BS (e.g., in the case of three BSs, one BS has three distributions, one for each BS). Using a continuous parameter distribution allows for probability values ​​for each location in the scene.

[0046] We define As a measure of generation The uplink signal from location The probability of transmission. Therefore, for each BS The result of the fitting step is the distribution. .

[0047] Therefore, this distribution serves as The expression for the function (an AoA statistic of a BS) is: = (1) The above terminology emphasizes that two special cases should be considered.

[0048] 1- If a ray of light passes through the xy plane several times, consider focusing on the term. Above. If it corresponds to AoD The light is only in position If a point traverses the plane xy once, then for all other points... , and If the light is in two positions and If it passes through the plane xy, then And for all other , Therefore, generally speaking, the intersection position probability for .

[0049] 2- One position Passed by several rays: In this case, as shown by the integral in the equation above, the probability corresponding to each ray is... They should be simply added together.

[0050] As mentioned above, the goal is to fit the probability distribution. The point set obtained when rays are emitted in a Monte Carlo manner. Since AoD is based on... Since sampling is performed, the above equation (1) becomes: , (2) This means probability It can be estimated in the following way: for crossing a given position The number of rays (i.e., in the diagram) The number of points at each point is counted, and a weighting factor is applied if the light ray crosses the xy plane multiple times. For simplicity, assume that the light ray passes through the xy plane only once.

[0051] One possibility for performing a fit is to use a Gaussian mixture model (GMM), but other distributions can also be considered. As mentioned earlier, Figure 9 An example of fitting a GMM to a set of points is shown. The aforementioned weighting can be easily considered in the GMM fitting algorithm if needed. A ray of light passes through the xy plane several times (”).

[0052] For each BS, this distribution fitting process is repeated. Therefore, we have Distribution A distribution for in Each measurement performed at each location in each BS .

[0053] Localization using fitted distribution

[0054] As can be seen in the following paragraphs, using the product of the fitted density functions does indeed provide an estimate of the location probability, as shown in equation (3.5).

[0055] General methods of positioning include using relative to Marginalization to calculate probability As shown below:

[0056] the term It is an indicator function because it only applies when... A ray of light in The term is equal to 1 only when there is an intersection.

[0057] Alternatively, the location may have:

[0058] Measurements in different BSs are independent, therefore:

[0059] Under the conditional independent likelihood (CIL) condition, the following can be solved:

[0060] Similarly,

[0061] Regrouping the above expressions, equation (3) can be derived as:

[0062] in, It is a location The prior probability. Without any prior information, the prior probability can be initialized uniformly, or it can be selected to follow an appropriate distribution. Then, each individual posterior probability can be calculated using equation (2) and parametric distribution data fitting. If the prior probability If it is uniform, then we can obtain: (3.5) The substitution is performed according to (1). Since the product operation and the integration operation can be interchanged, it can be verified that the two expressions starting with angle and position respectively are actually equal.

[0063] Therefore, it has a distribution For all It can be calculated directly by multiplying the probability density functions. This allows us to ensure that the method of the present invention is accurate.

[0064] This leads to the two main implementation patterns described below. Once you have The localization problem can then be derived from maximum a posteriori (MAP) estimation, as shown below:

[0065] Utilization of fitted distribution in data fusion

[0066] If another set of data is available, data fusion can be used to achieve better positioning accuracy. Assuming that in addition to the calculated data mentioned above... In addition, there is extra data. Fusion is considered to be:

[0067] Assuming that data y and z are independent, therefore:

[0068] Under CIL conditions, the following conclusions are drawn:

[0069] Therefore, equation (4) simply becomes:

[0070] In the above formula, From the data The information extracted from; and From the data The information extracted from it.

[0071] Based on the above mathematical explanation, two implementation modes of the proposed method can be derived: the first mode is a fully online mode, which, unlike existing methods, requires emitting a limited number of rays. The second mode is a partially offline mode, where some computations are performed offline to prepare some results in a partially available online database (query database). In both cases, the amount of online resources used is limited, and the accuracy of the method is improved.

[0072] First Implementation Mode

[0073] In the first implementation mode (10, Figure 8 In this context, the localization estimation is implemented as follows.

[0074] AoA measurement error Estimated offline for each BS. Then, in the online phase.

[0075] S01. Uplink AoA Measurement Performed by each BS, thus passing a vector of three-dimensional angles ( () ), S02. Obtain the corresponding probability distribution set ( The steps include: S021. Each BS emits light in a Monte Carlo manner (or a centralized device emits light for each BS), where the AoD angle is determined according to... Perform sampling. S022. For each BS, construct a graph of the points through which the light rays pass in the xy plane. S023. For each graph, the parameter distribution (such as GMM) Calculated S03. The steps for calculating the location of the user equipment (S03) include: S031. Distribution It is calculated by multiplying the calculated PDFs of all BSs. S032. Apply p(x|y) to the target task (to obtain an estimate of the location), for example, the maximum value of the distribution gives the most likely estimate.

[0076] Figure 10 The positioning accuracy is shown for both low-number-of-rays (100) and high-number-of-rays (10000) emitted per BS. The cumulative density function of the positioning error is used. Regarding the GMM fitting algorithm, for example, a standard GMM fitting algorithm with multiple patterns (clustering) is implemented, maintaining the model with the lowest possible Akaike information standard value. It can be noted that there is no significant difference in UE positioning accuracy in both cases (100 rays vs. 10000 rays). The distribution fitting method is able to maintain performance while reducing the number of rays. Therefore, this method allows for a reduction in the resources required to obtain the UE's location while maintaining the same accuracy of the results.

[0077] Second implementation mode

[0078] While the first implementation is robust to the reduced number of rays, it still involves ray emission in the online phase and fitting the distribution. This still results in considerable computational complexity and potential latency, which can be reduced.

[0079] This problem can be solved using the following second implementation mode. In short, the main steps involve calculating and storing the AoA for each possible measurement in an offline phase. The distribution. Then, in the online phase, each BS measures the angle. Then the corresponding storage distribution It is restored. Then, step S03 of the first implementation mode is executed, including steps S031 and S032. This method also combines Figure 8 The method has been described, and includes: During the offline phase: AoA measurement error Estimated offline by each BS or system. Then, offline preparation of the probability distribution set is performed (S00): S001 In the case of 3D angles, for possible 3D angles Discretize Considering the location of the BS, this discretization step can be optimized based on possible impossible receiving angles: for example, it is best to disregard impossible receiving angles (e.g., behind the BS if it is fixed to a wall); the discretization step can include dividing intervals in degrees, such as dividing them in 180-degree intervals. Interval. Therefore, for complete Thus, 360 can be obtained 180 = 64,800 possible 3D angles.

[0080] S002 For each BS and each discretized angle It emits light in a Monte Carlo manner, and the AoD angle is based on... Sampling is performed to obtain a point map of light rays passing through the xy plane. S003 For each obtained point plot, fit the distribution For example, GMM, and store the parameters of the distribution in a data structure.

[0081] During the online phase: S01 Uplink AoA Measurement Performed by each BS, thus passing a vector of three-dimensional angles ( () ), S02 distribution The parameters are recovered from the data structure: Using vectors ( From the angle of the measured 3D angle, the corresponding pre-calculated distribution is recovered from the stored probability distribution set prepared in step S00. That is, the entry used to obtain a corresponding probability distribution for a given BS is the measured 3D angle for that BS. S03. The steps for calculating the location of the user equipment (S03) include: S031. Distribution It is calculated by multiplying the pdfs of all BSs. S032. The maximum value of the distribution gives the most likely estimate.

[0082] As mentioned above, the main advantage of this second implementation mode is that it avoids ray emission and distribution fitting in the online phase. However, it requires storing the distribution parameters for each possible angle and each possible BS. The estimate of the number of values ​​to be stored is based on several parameters.

[0083] set up This represents the number of BS. In the simulation example, .set up The number of values ​​used to discretize the possible 3D angles (consider a simpler case, i.e., 180). 180 = 32400). Let P be the number of parameters in the fitted distribution. For example, if a 2D GMM with 7 clusters is used, then 6 × 7 = 42 parameters are needed (where 6 represents the product of the four values ​​of the 2D mean and covariance). Therefore, the number of parameters to be stored is:

[0084] Based on the given numbers, the number of values ​​is: 5,443,200 (or 1,360,800 per BS).

[0085] Please note that the previous examples only considered 2D positioning issues.

[0086] structure.

[0087] For 3D localization problems (i.e., user equipment needs to be positioned in 3D space), the proposed method can be repeated over multiple XY planes. The goal is then to find the correct XY plane, which gives the Z coordinates. For example, in variants or features, an XY plane with a high probability of a (x, y) location relative to other locations can be selected (i.e., an XY plane that is highly likely to cluster and has low variance). Many techniques for selecting an XY plane from several available planes can be envisioned or derived from the proposed method.

[0088] Depending on another variant or feature, the proposed method can also be extended to the 3D localization problem, as shown below. The 3D set of possible locations of the UE can be divided into small cubes, and it is now possible to check whether a ray passes through the cube to construct an equivalent of the point map. In other words, instead of cutting the 3D environment into several planes, the environment is cut into a set of cubes of a predetermined size. The other steps of the method remain unchanged.

[0089] According to another feature or variation, in the case of a non-uniform antenna pattern / beamforming for the UE, its orientation can be advantageously considered. Therefore, when constructing a point map based on ray tracing, the AoA of the rays at intersections can be considered. In other words, a map is constructed for one orientation of the UE, where only points corresponding to rays arriving with the expected orientation are retained. Therefore, for measurement... Construct several graphs, one for each orientation of the UE. Then, for each possible orientation, implement at least one of the two implementation modes described above (online only / partially offline), while maintaining the model that produces the best probability. Note that only the fitted distributions corresponding to the same orientation should be combined. Combining distributions associated with different orientations is useless.

[0090] According to the above embodiments, each receiver can determine all or part of the location of the user equipment based on its own receiver configuration. Furthermore, the determination of the user equipment location can also be performed by a centralized device (which may be one of the receivers Rx), which obtains the necessary data from the receivers and calculates the location, as disclosed in this disclosure. Figure 11As shown and described above, the configuration of the device DV affecting the measurement can be an antenna system AS, which is connected, for example, via an interface IN to a processing circuit including a processor PROC and a memory MEM. Alternatively, as a supplement or variation, necessary data (e.g., AoA) can also be received via a communication link CL, allowing the device to perform the position determination of this disclosure. The memory stores at least the instructions of the computer program according to this disclosure.

Claims

1. A method for determining a piece of data representing the location of a user equipment within an environment comprising a set of base stations, the method comprising the steps of: - Obtain (S01) a vector of N angles of the received signal emitted by the user equipment. Each angle of the vector corresponds to the reception angle of the signal at a given base station in the set of base stations. - Obtain (S02)N probability distributions, each probability distribution being obtained from the result of light emission in the 3D representation of the environment. - Calculate (S03) at least one probability of the location of the user equipment, and the at least one probability is a function of the vector of the N probability distributions and the N angles.

2. The method according to claim 1, wherein, The step of obtaining the N probability distributions (S02) includes loading the probability distributions from the data structure.

3. The method according to claim 2, wherein, The step of obtaining (S02) the N probability distributions further includes selecting each angle of signal reception from the probability distributions. A step in a probability distribution.

4. The method according to claim 1, wherein, The step of obtaining the N probability distributions (S02) includes the step of calculating the probability distribution for each base station based on the received virtual angle.

5. The method according to claim 4, wherein, The step of calculating the probability distribution density set for each base station based on the received virtual angle includes at least one iteration of the following steps: - Determine a virtual reception angle for the signal from among multiple possible reception angles. - Calculate a probability distribution based on the virtual receiving angle. - Store the probability distribution associated with the virtual angle in a data structure.

6. The method according to claim 5, wherein, The steps for calculating the probability distribution based on the virtual receiving angle include: - At least one ray is emitted into the 3D representation of the environment using a Monte Carlo algorithm, wherein the AoD of the at least one ray is sampled according to the virtual angle. - Using the at least one ray emitted in the 3D representation of the environment, a map of points is constructed, wherein the at least one ray passes through a plane at a predetermined height in the 3D representation of the environment at the point. - Calculate the parameter distribution of the points in the graph of the points, and pass the probability distribution associated with the virtual angle.

7. The method according to claim 1, wherein, The step of calculating (S03) the position of the user equipment as a function of the vectors of the N probability distributions and the N angles includes: - The global probability distribution (S031) is calculated by multiplying the N probability distributions together. - Based on the global probability distribution, determine (S032) the maximum value representing the most likely location of the user equipment.

8. The method according to claim 1, wherein, The steps to obtain (S02) N probability distribution sets include: for the N angles Each angle of the vector , - At least one ray is emitted into the 3D representation of the environment using a Monte Carlo algorithm, wherein the AoD of the at least one ray is determined according to the angle. Perform sampling. - Using the at least one ray emitted in the 3D representation of the environment, a map of points is constructed, wherein the at least one ray passes through a plane at a predetermined height in the 3D representation of the environment at the point. - Calculate the parameter distribution of the points in the graph, and transfer it to the angle. The associated probability distribution.

9. A computer program comprising instructions that, when executed by a processor, cause the implementation of the method according to any one of the preceding claims.

10. An apparatus for determining the location of a transmitter in an environment comprising a plurality of receivers, each of the plurality of receivers having a known location, the apparatus comprising an interface for receiving measurements performed by the receivers and a processor for implementing the method according to any one of claims 1 to 8.

Images