Gate adaptive radar data association method and device

Abstract

The invention discloses a gate adaptive radar data association method and device. The method comprises the following steps: acquiring one-step predicted position information of a plurality of targets and echo information of a plurality of sensors; generating a first position Euclidean distance feature tensor between the targets and a second position Euclidean distance feature tensor between the targets and echoes according to the one-step prediction position information and the echo information; determining a gate threshold parameter of the target based on the first position Euclidean distance feature tensor and a position Mahalanobis distance feature tensor between the target and the echo; determining an association result of each target in each sensor according to the position mahalanobis distance feature tensor between the target and the echo, the second position Euclidean distance feature tensor and a wave gate threshold parameter; according to the method, the environment of the current tracking scene can be better perceived under the condition of not depending on prior information through self-adaptive adjustment of the gate threshold parameter, the calculation burden is relieved for subsequent data association, and meanwhile, the calculation difficulty is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of radar target tracking, and in particular relates to a gate adaptive radar data association method and device. Background Art

[0002] With the continuous advancement of science and technology, people are no longer satisfied with target tracking with only a single sensor. Compared with the single-sensor tracking system, the core of the multi-sensor tracking system is to use the measurement information of multiple sensors to improve the stability and observation accuracy of target tracking. To realize this advantage, we must first clarify the source of each measurement information, that is, to achieve data association.

[0003] However, in multi-target tracking, especially in dense clutter scenarios, the use of multi-sensor systems brings a larger amount of data, making the data association process more challenging. On the one hand, the large amount of data poses a challenge to the algorithm's computational complexity and speed. On the other hand, how to achieve high-precision data association from multiple sensors poses a challenge to the algorithm's ability to process complex information and logic.

[0004] Traditional data association algorithms rely on prior information such as clutter density and clutter distribution to achieve accurate association. When the prior information is inaccurate, the algorithm performance will be greatly affected. In practical applications, this information may be difficult to obtain or even unknowable. In addition, traditional algorithms are more suitable for scenes with scattered targets and sparse environmental clutter. When the clutter is dense and the target is close, the algorithm association accuracy is low, and it may also cause combinatorial explosion and surge in computational complexity. Summary of the invention

[0005] The object of the present invention is to provide a gate adaptive radar data association method and device to reduce the amount of calculation when clutter is dense and the target is approaching.

[0006] The present invention adopts the following technical solution: a gate adaptive radar data association method, comprising the following steps:

[0007] Obtain one-step predicted position information of multiple targets and echo information of multiple sensors;

[0008] Generate a first position Euclidean distance feature tensor between targets, a second position Euclidean distance feature tensor between targets and echoes, and a position Mahalanobis distance feature tensor between targets and echoes according to the one-step predicted position information and the echo information;

[0009] Determine the gate threshold parameter of the target based on the first position Euclidean distance feature tensor and the position Mahalanobis distance feature tensor between the target and the echo;

[0010] The association result of each target in each sensor is determined according to the position Mahalanobis distance feature tensor between the target and the echo, the second position Euclidean distance feature tensor and the gate threshold parameter.

[0011] Furthermore, the gate threshold parameters are obtained through a CNN-based adaptive gate network module, and the specific method includes:

[0012] Extract feature information of the first position Euclidean distance feature tensor to obtain inter-target feature information;

[0013] Extract the feature information of the position Mahalanobis distance feature tensor between the target and the echo to obtain the feature information between the target echoes;

[0014] splicing the feature information between targets and the feature information between target echoes to obtain splicing feature information;

[0015] The splicing feature information is classified to obtain the category corresponding to the gate threshold parameter of each target; wherein the category is the size category of the gate threshold parameter.

[0016] Furthermore, the feature information of the position Mahalanobis distance feature tensor between the target and the echo is extracted, including:

[0017] The position Mahalanobis distance feature tensor between the target and the echo is mapped to a high-dimensional space through the first fully connected layer;

[0018] The corresponding target echo feature vectors are extracted through multiple one-dimensional convolutional layers with different kernel sizes;

[0019] Concatenate the feature vectors between multiple target echoes;

[0020] The second fully connected layer is used to perform feature fusion on the concatenated target echo feature vectors to obtain the target echo feature information.

[0021] Further, extracting feature information of the first position Euclidean distance feature tensor includes:

[0022] The first position Euclidean distance feature tensor is mapped to a high-dimensional space through a third fully connected layer, a batch normalization layer, and a fourth fully connected layer connected in series;

[0023] The corresponding inter-target feature vectors are extracted through multiple one-dimensional convolutional layers with different kernel sizes;

[0024] Concatenate feature vectors between multiple targets;

[0025] The fifth fully connected layer performs feature fusion on the concatenated inter-target feature vectors to obtain inter-target feature information.

[0026] Further, determining the association result of each target in each sensor includes:

[0027] The association probability of each target with each echo in each sensor is determined using a Bi-LSTM-based multi-sensor multi-target data association network module.

[0028] Furthermore, the association probability between each target and each echo in each sensor is determined by using a multi-sensor multi-target data association network module based on Bi-LSTM, including:

[0029] Selecting candidate echo information for each target from the echo information of each sensor;

[0030] Reset all element values ​​in the second position Euclidean distance feature tensor that do not correspond to the candidate echo information;

[0031] The reset second position Euclidean distance feature tensor is input into the Bi-LSTM layer to obtain high-dimensional features;

[0032] Input the high-dimensional features into the sixth fully connected layer to obtain fused features;

[0033] Input the fused features into the multi-channel feature construction layer to obtain multi-channel features; wherein the number of channels in the multi-channel features is the number of sensors;

[0034] The multi-channel features are input into the connected multi-channel one-dimensional convolutional layers and multiple seventh fully connected layers to obtain the association probability matrix of each sensor for each target.

[0035] Furthermore, the association probability matrix of each target and each sensor includes multiple one-dimensional vectors, each of which represents the association probability of a target and all echo information in a sensor; wherein the one-dimensional vector also includes the association probability of a target and a missed detection situation.

[0036] Furthermore, determining the association probability of each target with each echo in each sensor using a multi-sensor multi-target data association network module based on Bi-LSTM also includes:

[0037] In each one-dimensional vector, the echo information or missed detection corresponding to the maximum association probability is selected as the association result of the corresponding target.

[0038] Further, when a target is associated with multiple echoes, one of the echoes is selected for association with the target;

[0039] When an echo is associated with multiple targets, select one of the targets for the echo to associate with.

[0040] Another technical solution of the present invention: a gate adaptive radar data association device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above method when executing the computer program.

[0041] The beneficial effects of the present invention are as follows: the present invention can more accurately filter out clutter by adaptively adjusting the gate threshold parameters, ensuring that the clutter that falls into the gate while the real measurement falls into the gate is as small as possible, and better perceive the environment of the current tracking scene without relying on prior information, thereby reducing the computational burden for subsequent data association and reducing the computational difficulty. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flow chart of a gate adaptive radar data association method according to an embodiment of the present invention;

[0043] Figure 2 FIG. 1 is a working process diagram of an adaptive wave gate network module based on CNN in an embodiment of the present invention;

[0044] Figure 3 This is a working process diagram of a multi-sensor multi-target data association network module based on Bi-LSTM in an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of a simulation scenario in a simulation verification embodiment of the present invention;

[0046] Figure 5 This is a schematic diagram of the correlation accuracy of each target when tracking by different methods in the simulation verification embodiment of the present invention;

[0047] Figure 6 This is a schematic diagram of the correlation accuracy of each target under different clutter densities and different methods in the simulation verification embodiment of the present invention. DETAILED DESCRIPTION

[0048] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Among traditional data association algorithms, the decision-based nearest neighbor algorithm (NNDA) is the simplest one, which directly associates the target with the measurement that is closest to its predicted position in the relevant gate. However, it is difficult to solve the single target tracking problem in the case of dense clutter and the multi-target tracking problem with similar target trajectories. The probability data association algorithm (PDA) based on the Bayesian formula and the joint probability algorithm (JPDA) proposed on the basis of the PDA algorithm obtain the measurement associated with the target by weighting the measurement that falls into the associated gate. This type of algorithm improves the accuracy of data association, but relies on prior information. Among them, PDA is suitable for single target or multi-target tracking with sparse targets, and JPDA can adapt to more complex situations, but when the number of targets and candidate echoes is large, the computational complexity is high and it is difficult to operate in real time.

[0050] Traditional data association algorithms rely on prior information such as clutter density and clutter distribution to achieve accurate association. When the prior information is inaccurate, the algorithm performance will be greatly affected. In practical applications, this information may be difficult to obtain or even unknowable. In addition, traditional algorithms are more suitable for scenes with scattered targets and sparse environmental clutter. When the clutter is dense and the target is close, the algorithm association accuracy is low, and it may also cause combinatorial explosion and surge in computational complexity.

[0051] The data association method of deep learning uses a recurrent neural network as its main structure. By inputting target attribute information and radar observation information at a single or multiple moments, or feature information constructed from the two into the network, a vector representing the probability of association between the target and all measurements is obtained, and then the association relationship is judged to complete the final association. These methods are still in the exploratory stage, not mature and systematic enough, and do not take into account the more complex and challenging actual environment faced by engineering applications. Their applicability to different scenarios needs to be strengthened.

[0052] There are also some disadvantages in introducing data-driven methods: First, these methods are designed for single sensors and cannot efficiently handle data association problems in multi-sensor tracking systems. Second, since all measurements are directly used for association, some measurements that obviously do not belong to the target are also used as candidate measurements, which adds unnecessary difficulty to the task. At the same time, when the amount of clutter is large, it will also bring huge waste of resources. In addition, most of these methods directly use the target state value, the measurement state value, and the Euclidean distance between the above two as network inputs, so that their generalization ability is still insufficient. When the tracking scene changes significantly (such as a change in the target state distribution), the performance will deteriorate.

[0053] In order to overcome the above shortcomings, the present invention designs an intelligent data association method with adaptive gates, which uses a CNN-based adaptive gate module to replace a constant gate or a rule-based adaptive gate to mine implicit potential features from the target and measurement data, so that while ensuring that the real measurement falls into the gate, the number of echoes falling into the gate is as small as possible. This gate that changes with scene factors is helpful in dealing with scenes with a lack of prior knowledge. At the same time, the association module based on Bi-LSTM is used to capture and learn the mutual influence that may occur when the target data is associated, that is, the competition between different targets for the same measurement, as well as the potential information contained in multi-sensor data, to achieve more accurate target-measurement association. This avoids the high computational burden and low timeliness introduced by classic algorithms such as JPDA when considering competition between targets.

[0054] In order to solve the problem of greatly increased data volume, association difficulty and complexity caused by multiple sensors, dense clutter and close targets, a data-driven method with adaptive gate network and data association network as the main components is used to complete the data association task. The processing process is as follows Figure 1 As shown, this is achieved through the following steps.

[0055] Specifically, the present invention discloses a gate adaptive radar data association method, comprising the following steps: obtaining one-step predicted position information of multiple targets and echo information of multiple sensors; generating a first position Euclidean distance feature tensor between targets, a second position Euclidean distance feature tensor between targets and echoes, and a position Mahalanobis distance feature tensor between targets and echoes according to the one-step predicted position information and the echo information; determining a gate threshold parameter of the target based on the first position Euclidean distance feature tensor and the position Mahalanobis distance feature tensor between targets and echoes; determining an association result of each target in each sensor according to the position Mahalanobis distance feature tensor between targets and echoes, the second position Euclidean distance feature tensor, and the gate threshold parameter.

[0056] The present invention can filter out clutter more accurately by adaptively adjusting the gate threshold parameters, ensuring that the clutter falling into the gate while the real measurement falls into the gate is as small as possible, and better perceive the environment of the current tracking scene without relying on prior information, thereby reducing the calculation burden for subsequent data association and reducing the calculation difficulty.

[0057] That is to say, this method first constructs the features of the echo information and target track information available in the current frame to obtain the input tensor of the network model. Specifically, the target-target position Euclidean distance feature tensor (i.e., the first position Euclidean distance feature tensor between targets), the target-echo position Euclidean distance feature tensor (i.e., the second position Euclidean distance feature tensor between the target and the echo), and the Mahalanobis distance feature between the target-echo position considering the standard deviation of the measurement noise (i.e., the position Mahalanobis distance feature tensor between the target and the echo) are constructed.

[0058] Next, the first position Euclidean distance feature tensor and the position Mahalanobis distance feature tensor between the target and the echo are input into the CNN-based adaptive wave gate network module to learn the threshold size that is suitable for the current scene. The working process of the CNN-based adaptive wave gate network module is as follows: Figure 2 shown.

[0059] First, the input feature data is normalized to the maximum and minimum values, and then the two input feature vectors are extracted respectively. That is, the feature information of the first position Euclidean distance feature tensor is extracted to obtain the inter-target feature information, and the feature information of the position Mahalanobis distance feature tensor between the target and the echo is extracted to obtain the inter-target echo feature information.

[0060] In order to obtain more effective feature mapping, different feature extraction networks are designed for the two inputs. For the position Mahalanobis distance feature tensor between the target and the echo, the feature is mapped to a high-dimensional space through the first fully connected layer, and then the corresponding target echo feature vector is extracted through multiple one-dimensional convolutional layers with different kernel sizes. Then, multiple target echo feature vectors are concatenated, and finally, the feature is fused through the second fully connected layer to obtain the feature information between target echoes.

[0061] For the first-position Euclidean distance feature tensor, the feature extraction network is transformed based on the aforementioned network. Specifically, the part before entering the convolution is changed from a single-layer fully connected layer structure to a third fully connected layer, a batch normalization layer, and a fourth fully connected layer structure.

[0062] That is to say, the first position Euclidean distance feature tensor is mapped to a high-dimensional space through the third fully connected layer, batch normalization layer and fourth fully connected layer connected in series; the corresponding inter-target feature vectors are extracted through multiple one-dimensional convolutional layers with different kernel sizes; the multiple inter-target feature vectors are spliced; and the spliced ​​inter-target feature vectors are feature fused through the fifth fully connected layer to obtain inter-target feature information.

[0063] After obtaining the target echo feature information and the target inter-echo feature information, the target echo feature information and the target echo feature information are spliced ​​to obtain the spliced ​​feature information; the spliced ​​feature information is then classified to obtain the category corresponding to the gate threshold parameter of each target; wherein the category is the size category of the gate threshold parameter.

[0064] That is to say, the two features extracted from the input are concatenated and input into the classifier composed of the fully connected layer to obtain the adaptive wave gate threshold value of each target. Here, solving the adaptive threshold is modeled as a classification problem, and the output adaptive threshold belongs to different threshold size levels such as small (such as 1 to 4), smaller (such as 4 to 9), larger (such as 9 to 16), and large (such as 16 to 25), and corresponds to different classification categories and specific numerical values ​​according to the actual application.

[0065] After obtaining the threshold size level of the adaptive wave gate, the second position Euclidean distance feature tensor, the position Mahalanobis distance feature tensor and the wave gate threshold parameters are input into the multi-sensor multi-target data association network module based on Bi-LSTM to solve the association probability matrix of each target and each sensor, that is, to determine the association probability of each target with each echo in each sensor.

[0066] The working process of the multi-sensor multi-target data association network module based on Bi-LSTM is as follows: Figure 3 As shown. First, by comparing the position Mahalanobis distance feature tensor and the gate threshold parameter of each target, gate filtering is performed to determine the candidate echo corresponding to each target in each sensor, that is, candidate echo information is selected for each target in the echo information of each sensor. For example, the position Mahalanobis distance feature tensor value between the target and the echo is 25, and the gate threshold parameter is 20. Since 25>20, the echo is eliminated.

[0067] Then, the maximum and minimum values ​​of the second position Euclidean distance feature tensor are normalized, and the normalized second position Euclidean distance feature tensor is obtained according to the result of filtering out clutter by the wave gate. The specific operation is: the element values ​​corresponding to the candidate echo in the normalized second position Euclidean distance feature tensor remain unchanged, and the other element values ​​are reset to 1, that is, the element values ​​in the second position Euclidean distance feature tensor that do not correspond to the candidate echo information are reset.

[0068] Next, the reset second-position Euclidean distance feature tensor is input into a deep learning network, which includes a Bi-LSTM layer, a fully connected layer, a multi-channel feature construction layer, a multi-channel one-dimensional convolutional layer, and multiple fully connected layers.

[0069] Among them, the Bi-LSTM layer considers the mutual influence when different target data are associated, mines the potential information contained in the multi-sensor data, and maps the low-dimensional features to high-dimensional features. In other words, the reset second-position Euclidean distance feature tensor is input into the Bi-LSTM layer to obtain high-dimensional features.

[0070] The fully connected layer fuses the different features extracted by the forward LSTM and backward LSTM in the Bi-LSTM layer, that is, the high-dimensional features are input into the sixth fully connected layer to obtain the fused features.

[0071] The multi-channel feature construction layer constructs the fused features into multi-channel features, and the number of channels is the number of multi-sensors; that is, the fused features are input into the multi-channel feature construction layer to obtain multi-channel features; wherein the number of channels in the multi-channel features is the number of sensors.

[0072] The multi-channel one-dimensional convolution layer is used to extract and compress features; finally, multiple fully connected layers map the features into the association probability matrix of each target and each sensor, and each one-dimensional vector represents the association probability distribution vector between a target track and an echo obtained by a sensor. In other words, the multi-channel features are input into the connected multi-channel one-dimensional convolution layer and multiple seventh fully connected layers to obtain the association probability matrix of each target and each sensor.

[0073] Furthermore, the association probability matrix of each target and each sensor includes multiple one-dimensional vectors, each of which represents the association probability of a target and all echo information in a sensor; wherein the one-dimensional vector also includes the association probability of a target and a missed detection situation.

[0074] Taking into account possible missed detections, when calculating the probability distribution vector, in addition to calculating a corresponding association probability for each echo, a corresponding probability must also be calculated for missed detections, that is, the probability that the target has no corresponding real measurement at the current moment and the current sensor.

[0075] Finally, the association matrix is ​​calculated by association relationship judgment to determine the association result of each target in each sensor. Intuitively, the echo or missed detection with the highest probability of association between each sensor and the target is selected as the association result of the target, that is, the echo information or missed detection corresponding to the maximum association probability in each one-dimensional vector is selected as the association result of the corresponding target.

[0076] However, this cannot be done simply when a matching conflict occurs (for example, one echo is associated with multiple targets). This is because in the multi-target tracking process, each target track is associated with at most one echo in the echo of each sensor at each moment, and each echo can only be associated with at most one target.

[0077] Therefore, when solving the correlation result, we must first determine whether the above constraints are met by directly selecting the echo with the maximum correlation probability as the correlation echo. If so, this is the final correlation result; otherwise, for the case where the constraints are not met, the echo correlation result is adjusted based on the principle of giving priority to assigning the echo with a large correlation probability to the target with a small target track number.

[0078] That is, when a target is associated with multiple echoes, one of the echoes is selected for association with the target; when an echo is associated with multiple targets, one of the targets is selected for association with the echo.

[0079] Simulation verification example:

[0080] Consider multi-source radar multi-target tracking in a two-dimensional rectangular coordinate system o-ξη. Here, the number of radars M = 3, and the sampling interval is 1s. After coordinate transformation, each radar (i.e., sensor) is located at the origin of the coordinate system, and the measurement matrix at time k in the corresponding measurement model is Where I represents the unit matrix, and the measurement noise covariance matrix R at time k is k =15 2 I2. At the same time, considering the dense clutter scene, the clutter is evenly distributed in the detection field of view, and the clutter density is λ = 10 -3 , and the number of clutter obeys the Poisson distribution with parameter λ.

[0081] The number of targets is T = 5, and all targets move in a uniform linear motion in different directions within the observation range. In order to better display the target tracking results and compare the tracking performance of different algorithms, all measurements are uniformly displayed using the Cartesian coordinate system through coordinate transformation. Figure 4 As shown, targets 1, 2, and 3 can be regarded as a formation, and targets 4 and 5 can be regarded as a formation. The targets in the formation always have similar and close motion trajectories. The two formations approach each other in 10 to 20 seconds. The initial positions and motion states of the targets are shown in Table 1.

[0082] Table 1 Example of target initial position and motion state setting

[0083] Goal 1 Goal 2 Goal 3 Goal 4 Goal 5 X / m 700 700 700 75 100 VX / (m / s) -20 -20 -20 20 20 Y / m 1110 1050 1000 1150 1150 VY / (m / s) 0 0 0 -7 -7

[0084] The data association accuracy is designed as the evaluation index. The data association accuracy is defined as follows: For target i, its association accuracy is calculated as follows:

[0085]

[0086] Where k represents the sampling time, represents the probability of association between target i and the correct echo point of radar j in the tth test of the corresponding algorithm, Represents the total number of tests (i.e. the number of simulation scenarios).

[0087] It is worth noting that for data association methods that directly select associated echoes from echoes, such as the data association method and NNDA algorithm disclosed in the present invention, it is believed that if the selected measurement is the target true measurement, then is 1, otherwise it is 0. For data association methods that use the combined measurement obtained by weighted combination of candidate echoes as the associated echo, such as the PDA algorithm and the JPDA algorithm, it is considered that is the weighted probability value corresponding to the target true measurement in the combined measurement.

[0088] In order to study the performance of the method of the present invention, it was experimentally compared with the classical data association method and the method using the network. The classical data association method selected NNDA, PDA, and JPDA as references. Since there is no public neural network method for multi-sensors, a method based on the proposed multi-sensor multi-target data association network module but without considering the gate filter (i.e., DANet) and a method based on the proposed multi-sensor multi-target data association network module that only considers the fixed gate filter (FGDANet) were selected as references. Here, Kalman filtering is used to achieve target track update, and the above methods are respectively abbreviated as NNDA, PDA, JPDA, DANet, FGDANet and AGDAN (the method proposed in the present invention). At the same time, in order to clearly compare the tracking results of different algorithms, no track termination measures were taken.

[0089] Figure 5 The data association accuracy of each target when using different methods is shown. Obviously, in the tracking process of each target, AGDAN has achieved the highest association accuracy (the horizontal and vertical axes in the figure represent k). For any target, the association accuracy of the three classical algorithms is difficult to reach more than 60%, while the association accuracy of the three intelligent methods is almost all above 70%. This shows that under experimental conditions, the intelligent method has better performance. This is because compared with the classical method, the intelligent method effectively integrates the information-assisted association of multiple sensors.

[0090] In addition, it can be found that as the two formations gradually approached and the distance between the targets decreased, the association accuracy of all targets fluctuated between 10 seconds and 20 seconds, especially targets 2, 4, and 5 showed a significant decline. In this process, the ADGAN algorithm was least affected, which reflects the effectiveness of the adaptive gate network module. The adaptive gate network module can capture the changes in the relative positions between targets, and adaptively change the size of the relevant gates in combination with the current tracking state, eliminate clutter interference as much as possible, and reduce the range of candidate echoes during association, making the data association accuracy less sensitive to environmental changes.

[0091] To further evaluate the algorithm performance, ADGAN is compared with other methods under different clutter density conditions. The average ADA (AADA) is selected as the evaluation indicator. AADA represents the average value of the correlation accuracy of the five targets during the whole tracking process. Figure 6 As shown, the clutter density is λ = 10 -5 and λ=5×10 -3 In almost all clutter density scenarios, the intelligent method is significantly better than the classical method. In the intelligent method, except for λ=5×10 -5 Except for this specific scenario, AGDAN has the highest accuracy; and in this specific scenario, AGDAN's performance is only slightly behind the best performance. In addition, when the scene clutter density changes, AGDAN's AADA changes the least, and its performance is always stable at a high level. In summary, these results and analysis strongly indicate that the proposed method has better correlation accuracy and better robustness.

[0092] In summary, the present invention designs an adaptive wave gate network module based on CNN, which takes the target-target position Euclidean distance feature tensor and the target-echo position Euclidean distance feature tensor as input to solve the adaptive wave gate threshold of each target in the current frame. The wave gate threshold changes adaptively according to the relative spatial position relationship between the targets and the spatial distribution of the sensor echo position. In this way, when filtering out clutter and determining the candidate echo through the wave gate, the clutter that falls into the wave gate can be as small as possible while ensuring that the real measurement falls into the wave gate. Without relying on prior information, the environment of the current tracking scene can be better perceived, which reduces the computational burden and difficulty of subsequent data association.

[0093] In addition, the present invention also designs a multi-sensor multi-target data association network module based on Bi-LSTM, which takes the target-echo position Euclidean distance feature tensor, the target-echo position Mahalanobis distance feature tensor and the adaptive wave gate threshold value as input to solve the association probability matrix between each target and each sensor echo. This module calculates the association probability by mining the mutual influence when different target data are associated and the potential information hidden in the multi-sensor echo data, thereby improving the accuracy of the association result. At the same time, when solving the association problem, the missed detection situation is taken into account, and the result of data association is improved.

[0094] That is to say, the threshold corresponding to each target is solved through the CNN-based wave gate adaptive network module. The threshold changes with the changes of the target and echo information in the scene, so as to reduce the inclusion of clutter points while including real measurements. The echo is filtered according to the wave gate threshold, and the filtered echo is sent to the data association network module based on Bi-LSTM, which comprehensively utilizes the detection information of each sensor to solve the association probability matrix.

[0095] It can be seen that the method of the present invention applies deep neural networks to the framework of classical data association methods. On the one hand, compared with traditional methods, it reduces the dependence on prior knowledge and avoids combinatorial explosion; on the other hand, compared with data-driven methods, it constrains the selection range of candidate echoes and improves the association accuracy. This method can achieve good tracking results while ensuring a high association accuracy rate and taking less time.

[0096] The present invention also discloses a gate adaptive radar data association device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor. The above method is implemented when the processor executes the computer program.

[0097] The present invention also discloses an embodiment that provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0098] The present invention also provides a computer program product. When the computer program product runs on a data storage device, the data storage device can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0099] If the integrated unit module is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device capable of carrying the computer program code to a storage device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electrical carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0100] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0101] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0102] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules or units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0103] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

Claims

1. A gate adaptive radar data association method, characterized in that: The following steps are involved: Obtain one-step predicted position information of multiple targets and echo information of multiple sensors; Generate a first position Euclidean distance feature tensor between targets, a second position Euclidean distance feature tensor between targets and echoes, and a position Mahalanobis distance feature tensor between targets and echoes according to the one-step predicted position information and echo information; Determine a gate threshold parameter of the target based on the first position Euclidean distance feature tensor and the position Mahalanobis distance feature tensor between the target and the echo; The association result of each target in each sensor is determined according to the position Mahalanobis distance feature tensor between the target and the echo, the second position Euclidean distance feature tensor and the gate threshold parameter.

2. A gate adaptive radar data association method as claimed in claim 1, characterized in that: The gate threshold parameter is obtained by a CNN-based adaptive gate network module, and the specific method includes: Extracting feature information of the Euclidean distance feature tensor of the first position to obtain inter-target feature information; Extract the feature information of the position Mahalanobis distance feature tensor between the target and the echo to obtain the feature information between the target echoes; splicing the inter-target feature information and the inter-target echo feature information to obtain spliced ​​feature information; The splicing feature information is classified to obtain a category corresponding to the gate threshold parameter of each target; wherein the category is a size category of the gate threshold parameter.

3. A gate adaptive radar data association method as claimed in claim 2, characterized in that: Extracting feature information of the position Mahalanobis distance feature tensor between the target and the echo includes: The position Mahalanobis distance feature tensor between the target and the echo is mapped to a high-dimensional space through the first fully connected layer; The corresponding target echo feature vectors are extracted through multiple one-dimensional convolutional layers with different kernel sizes; splicing the characteristic vectors between a plurality of the target echoes; The second fully connected layer performs feature fusion on the spliced ​​target echo feature vectors to obtain the target echo feature information.

4. A gate adaptive radar data association method as claimed in claim 3, characterized in that: Extracting feature information of the Euclidean distance feature tensor of the first position includes: Mapping the first position Euclidean distance feature tensor to a high-dimensional space through a third fully connected layer, a batch normalization layer, and a fourth fully connected layer connected in series; The corresponding inter-target feature vectors are extracted through multiple one-dimensional convolutional layers with different kernel sizes; splicing a plurality of said target inter-target feature vectors; The fifth fully connected layer performs feature fusion on the spliced ​​inter-object feature vectors to obtain the inter-object feature information.

5. A gate adaptive radar data association method as described in any one of claims 2 to 4, characterized in that: Determine the associated results for each target in each sensor including: The association probability of each target with each echo in each sensor is determined using a Bi-LSTM-based multi-sensor multi-target data association network module.

6. A gate adaptive radar data association method as claimed in claim 5, characterized in that: The association probability of each target with each echo in each sensor is determined by using the Bi-LSTM-based multi-sensor multi-target data association network module, including: Selecting candidate echo information for each target from the echo information of each sensor; Resetting all element values ​​in the second position Euclidean distance feature tensor that do not correspond to the candidate echo information; The reset second position Euclidean distance feature tensor is input into the Bi-LSTM layer to obtain high-dimensional features; Inputting the high-dimensional features into the sixth fully connected layer to obtain fused features; Input the fused features into a multi-channel feature construction layer to obtain a multi-channel feature; wherein the number of channels in the multi-channel feature is the number of sensors; The multi-channel features are input into the connected multi-channel one-dimensional convolutional layers and multiple seventh fully connected layers to obtain the association probability matrix of each sensor for each target.

7. A gate adaptive radar data association method as claimed in claim 6, characterized in that: The association probability matrix of each target and each sensor includes multiple one-dimensional vectors, each of which represents the association probability of a target and all echo information of a sensor; wherein the one-dimensional vector also includes the association probability of a target and a missed detection situation.

8. A gate adaptive radar data association method as claimed in claim 7, characterized in that: The association probability of each target with each echo in each sensor is determined by using a multi-sensor multi-target data association network module based on Bi-LSTM, and also includes: In each of the one-dimensional vectors, the echo information or missed detection corresponding to the maximum association probability is selected as the association result of the corresponding target.

9. A gate adaptive radar data association method as claimed in claim 8, characterized in that: When a target is associated with multiple echoes, one of the echoes is selected for association with the target; When an echo is associated with multiple targets, select one of the targets for the echo to associate with.

10. A gate adaptive radar data association device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 9 is implemented.

Images