Multi-target tracking method and device of millimeter wave radar and storage medium

Abstract

The application discloses a multi-target tracking method and device of a millimeter wave radar and a storage medium, relates to the technical field of radar signals, and discloses a multi-target tracking method and device of a millimeter wave radar and a storage medium, which comprises the following steps: performing multi-frame accumulation processing on radar signals through a sliding window, avoiding the missing of radar signals at individual moments caused by the flicker problem existing in the radar signals, simultaneously extracting effective point cloud data from a to-be-matched point cloud set through an association gate to group an initial confirmation matrix, thereby reducing the dimension of the initial confirmation matrix by partitioning the scene corresponding to each radar signal, obtaining a target confirmation matrix, and performing joint probability data association based on the target confirmation matrix, so that the number of interconnection matrices obtained by splitting the initial confirmation matrix is reduced, the change relationship between the number of joint events corresponding to the interconnection matrices and the to-be-matched point cloud data is changed into an approximately linear relationship, and the calculation resources of the joint probability data association algorithm are greatly reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of radar signals, in particular to a multi-target tracking method of a millimeter wave radar, a device and a storage medium. BACKGROUND

[0002] Target tracking is a core technology of radar signal processing, and plays a key role in intelligent transportation systems. It can analyze target information such as distance, speed, and azimuth angle detected by a millimeter wave radar to predict and update the motion state and trajectory of a vehicle, thereby realizing vehicle obstacle detection and avoidance, adaptive cruise, and other functions.

[0003] In the field of millimeter wave radar tracking, the most widely used algorithm is currently based on the Bayesian theory. This algorithm includes key components such as a motion model, track initiation, data association, and state filtering. Data association is a core problem in multi-target tracking algorithms. Classic data association algorithms include the Joint Probability Data Association (JPDA) algorithm. However, the number of joint events in the JPDA algorithm is an exponential function of all candidate measurements, which increases rapidly with the increase in measurement density, leading to an explosion in computational complexity and a huge memory burden. Therefore, the JPDA algorithm is difficult to apply in engineering. SUMMARY

[0004] The main purpose of the present application is to provide a multi-target tracking method of a millimeter wave radar, a device and a storage medium, which aims to solve the technical problem of excessive computational resources in the current multi-target tracking based on the joint probability data association algorithm.

[0005] To achieve the above purpose, the present application provides a multi-target tracking method of a millimeter wave radar, which comprises:

[0006] By setting multiple sliding windows, the radar signals accessed are subjected to multi-frame accumulation processing and density-based clustering processing to obtain a point cloud set to be matched. The target point cloud set at the previous time is traversed to calculate a predicted point cloud set.

[0007] According to the association relationship between the predicted point cloud set and the point cloud set to be matched, an association gate is established.

[0008] Based on the association gate, effective point cloud data in the point cloud set to be matched and first point cloud data corresponding to the effective point cloud data in the predicted point cloud set are obtained. An initial confirmation matrix is established according to the association relationship between the effective point cloud data and the first point cloud data.

[0009] obtaining a plurality of predicted point cloud indexes associated with each of the effective point cloud data in the initial confirmation matrix, grouping the effective point cloud data having the same predicted point cloud index and the predicted point cloud data pointed by the predicted point cloud index into a group to obtain a plurality of groups, and establishing a target confirmation matrix belonging to the group, and performing joint probability data association according to the target confirmation matrix of each group, wherein the predicted point cloud set comprises a plurality of predicted point cloud data.

[0010] In an embodiment, the step of obtaining the to-be-matched point cloud set by setting a plurality of sliding windows and performing multi-frame accumulation processing and density-based clustering processing on the accessed radar signals comprises:

[0011] The plurality of sliding windows are set, and the plurality of radar signals output by the millimeter wave radar in sequence are stored into the corresponding sliding windows, and when it is detected that a sliding window operation is completed, the radar signals stored on the plurality of sliding windows are extracted to obtain a plurality of point cloud data;

[0012] The plurality of point cloud data are subjected to density-based clustering processing to obtain the to-be-matched point cloud set.

[0013] In an embodiment, the millimeter wave radar comprises a plurality of target trackers, and one target tracker is used to track one point cloud data. The step of obtaining the predicted point cloud set by traversing the to-be-matched point cloud set comprises:

[0014] The target tracker is initialized;

[0015] After traversing the target tracks corresponding to each target point cloud data in the target point cloud set at the last time, the motion state prediction value of each target point cloud data at the current time and the state covariance matrix prediction value at the current time are calculated by using a uniform acceleration motion model to obtain the predicted point cloud set.

[0016] In an embodiment, the to-be-matched point cloud set comprises a plurality of to-be-matched point cloud data. The step of establishing an association gate according to the association relationship between the predicted point cloud set and the to-be-matched point cloud set comprises:

[0017] Each to-be-matched point cloud data is traversed, and the predicted point cloud data corresponding to the target point cloud data is traversed according to the to-be-matched point cloud data traversed to obtain the target predicted point cloud data corresponding to each target point cloud data;

[0018] The association gate is established on each target predicted point cloud data according to the to-be-matched point cloud data.

[0019] In an embodiment, the step of obtaining the valid point cloud data in the to-be-matched point cloud set based on the association gate comprises:

[0020] determining whether the target point cloud data falls within the association range of the association gate through the association gate;

[0021] if the target point cloud data falls within the association range, extracting the to-be-matched point cloud data corresponding to the target point cloud data falling within the association range as the valid point cloud data.

[0022] In an embodiment, the step of performing joint probability data association according to each target confirmation matrix of the groups comprises:

[0023] traversing each target confirmation matrix and splitting each target confirmation matrix to obtain a plurality of interconnection matrices, wherein each interconnection matrix represents a feasible joint event;

[0024] calculating the probability of the feasible joint event according to the interconnection matrix to obtain first probability data, and calculating the probability that the valid point cloud data included in the interconnection matrix belongs to the first point cloud data to obtain second probability data;

[0025] if the second probability data is greater than the second preset probability threshold according to the comparison relationship between the second probability data and the second preset probability threshold, marking the matching state of the valid point cloud data corresponding to the second probability data as a correct matching state, wherein the first point cloud data in the correct matching state is determined as first to-be-calculated point cloud data, and the valid point cloud data in the correct matching state is determined as second to-be-calculated point cloud data;

[0026] calculating the Kalman gain and the innovation vector of the first to-be-calculated point cloud data under the uniform acceleration motion model, updating the motion state prediction value and the state covariance matrix prediction value according to the Kalman gain and the innovation vector, to obtain the motion state and the posterior state covariance of the first to-be-calculated point cloud data.

[0027] In an embodiment, after the step of if the second probability data is greater than the second preset probability threshold, marking the matching state of the valid point cloud data corresponding to the second probability data as a correct matching state, comprises:

[0028] determining the valid point cloud data other than the second to-be-calculated point cloud data as to-be-judged point cloud data in an error matching state;

[0029] determining whether the to-be-judged point cloud data meets a preset creation condition;

[0030] If the to-be-judged point cloud data meets the preset creation condition, the to-be-judged point cloud data is determined as first next target point cloud data of a next target tracking process.

[0031] In an embodiment, after the step of performing joint probabilistic data association according to the target confirmation matrix, the method comprises:

[0032] The track and point cloud attribute corresponding to each first point cloud data are traversed, and the first point cloud data with a track state of a dead track and / or the first point cloud data with an attribute state of an abnormal attribute are extracted from each track and point cloud attribute, and then the first point cloud data with the track state of the dead track and / or the first point cloud data with the attribute state of the abnormal attribute are removed to obtain second next target point cloud data.

[0033] The first next target point cloud data and the second next target point cloud data are transmitted into a next target tracking process.

[0034] In addition, to achieve the above object, the present application further provides a multi-target tracking device of a millimeter wave radar, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the multi-target tracking method of the millimeter wave radar as described above.

[0035] In addition, to achieve the above object, the present application further provides a storage medium, which is a computer readable storage medium, and the storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the multi-target tracking method of the millimeter wave radar as described above.

[0036] The one or more technical solutions provided by the present application have at least the following technical effects:

[0037] The radar signal accessed is processed by multiple sliding windows for multi-frame accumulation and density-based clustering to obtain a point cloud set to be matched, and a predicted point cloud set is calculated by traversing the point cloud set to be matched; an association gate is established according to the association between the predicted point cloud set and the point cloud set to be matched; based on the association gate, effective point cloud data in the point cloud set to be matched and first point cloud data corresponding to the effective point cloud data in the predicted point cloud set are obtained, an initial confirmation matrix is established according to the association between the effective point cloud data and the first point cloud data; a plurality of predicted point cloud indexes associated with each effective point cloud data in the initial confirmation matrix are obtained, effective point cloud data with the same predicted point cloud index and predicted point cloud data pointed to by the predicted point cloud index are grouped into a group to obtain a plurality of groups, and a target confirmation matrix belonging to the group is established, and joint probability data association is performed according to the target confirmation matrix of each group, wherein the predicted point cloud set comprises a plurality of predicted point cloud data.

[0038] The present application processes the radar signal by setting a sliding window for multi-frame accumulation, compensates the position of the radar signal accessed at the previous time on the sliding window according to its motion state, avoids the missing of the radar signal at individual time caused by the flicker problem of the radar signal, extracts effective point cloud data from the point cloud set to be matched through the association gate, establishes an initial confirmation matrix through the association, groups the initial confirmation matrix, partitions the scene corresponding to each radar signal to reduce the dimension of the initial confirmation matrix, obtains a plurality of low-dimensional target confirmation matrices, and performs joint probability data association based on the low-dimensional target confirmation matrix, thereby reducing the number of interconnection matrices obtained by splitting the initial target confirmation matrix, changing the exponential growth relationship of the number of joint events corresponding to the interconnection matrix into an approximately linear relationship, and greatly reducing the computing resources of the joint probability data association algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0039] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the specification, serve to explain the principles of the application.

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0041] Figure 1 A flowchart is provided for the first embodiment of the multi-target tracking method of the millimeter wave radar of the present application;

[0042] Figure 2A flowchart of a general process of a multi-target tracking method of a millimeter wave radar of the present application is shown in FIG. 1.

[0043] Figure 3 A flowchart of a process provided for a second embodiment of a multi-target tracking method of a millimeter wave radar of the present application is shown in FIG. 2.

[0044] Figure 4 A flowchart of a process provided for a third embodiment of a multi-target tracking method of a millimeter wave radar of the present application is shown in FIG. 3.

[0045] Figure 5 A device structure diagram of a hardware running environment involved in a multi-target tracking method of a millimeter wave radar in an embodiment of the present application is shown in FIG. 4.

[0046] The object implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0047] It should be understood that the specific embodiments described herein are merely intended to explain the technical solutions of the present application, and are not intended to limit the present application.

[0048] In order to better understand the technical solutions of the present application, the following will be described in detail in combination with the drawings of the specification and the specific embodiments.

[0049] The main solution of the embodiment of the present application is: by setting multiple sliding windows, multi-frame accumulation processing and density-based clustering processing are performed on the accessed radar signals to obtain a to-be-matched point cloud set, and by traversing the to-be-matched point cloud set, a predicted point cloud set is calculated; according to the association relationship between the predicted point cloud set and the to-be-matched point cloud set, an association gate is established; based on the association gate, effective point cloud data in the to-be-matched point cloud set and first point cloud data corresponding to the effective point cloud data in the predicted point cloud set are obtained, and according to the association relationship between the effective point cloud data and the first point cloud data, an initial confirmation matrix is established; multiple predicted point cloud indexes associated with each effective point cloud data in the initial confirmation matrix are obtained, effective point cloud data and predicted point cloud data pointed to by the predicted point cloud index with the same predicted point cloud index are grouped into a group, multiple groups are obtained, and a target confirmation matrix belonging to the group is established, and joint probability data association is performed according to the target confirmation matrix of each group, wherein the predicted point cloud set includes a plurality of predicted point cloud data.

[0050] In the tracking field of millimeter wave radar, the most widely used algorithm is the multi-target tracking algorithm based on Bayesian theory, which includes key components such as motion model, track initiation, data association and state filtering. Among them, data association is the core problem in multi-target tracking algorithm. The classic data association algorithm includes joint probabilistic data association algorithm (JPDA) and the like. In the JPDA algorithm, the number of joint events is an exponential function of all candidate point cloud data, which increases rapidly with the increase of point cloud data density, resulting in an explosion of calculation and huge memory burden. Therefore, the JPDA algorithm is difficult to apply to engineering landing. At present, most of the research on multi-target tracking algorithm based on JPDA focuses on reducing the number of executable joint events by increasing the restriction of elliptical gate. The threshold condition needs to be set according to experience, and the relationship between the number of joint events and the number of point cloud data is not fundamentally changed.

[0051] The present application provides a solution. The radar signal is processed by setting a sliding window for multi-frame accumulation. The position of the radar signal accessed at the previous time is compensated on the sliding window according to its motion state, avoiding the missing of radar signal at individual time caused by the flicker problem of radar signal. At the same time, the effective point cloud data is extracted from the to-be-matched point cloud set by the association gate to group the initial confirmation matrix, so as to reduce the dimension of the initial confirmation matrix, obtain the target confirmation matrix, and perform joint probability data association based on the target confirmation matrix. The number of interconnection matrices obtained by splitting the initial confirmation matrix is reduced, the relationship between the number of joint events corresponding to the interconnection matrix and the number of to-be-matched point cloud data is changed from exponential growth to approximately linear relationship, and the calculation resources of the joint probability data association algorithm are greatly reduced.

[0052] It should be noted that the execution subject of the present embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a multi-target tracking device of millimeter wave radar, etc. In the following, the multi-target tracking device of millimeter wave radar is taken as an example to illustrate the present embodiment and each of the following embodiments.

[0053] Based on this, the present embodiment provides a multi-target tracking method of millimeter wave radar. Referring to Figure 1 and combining Figure 2 , Figure 1 is a flowchart of the first embodiment of the multi-target tracking method of millimeter wave radar of the present application.

[0054] In the present embodiment, the multi-target tracking method of millimeter wave radar includes steps S10-S40:

[0055] Step S10, by setting multiple sliding windows, the accessed radar signal is processed by multi-frame accumulation and density-based clustering, to obtain the point cloud set to be matched, and by traversing the target point cloud set of the last time, the predicted point cloud set is calculated.

[0056] Because the point cloud data corresponding to the radar signal transmitted by the millimeter wave radar has a flicker problem, which will cause the point cloud data at individual time to be missing, in the clutter environment, the missing of the point cloud data will reduce the robustness of multi-target tracking to some extent.

[0057] Therefore, in order to avoid the above problems, the embodiment proposes to first set multiple sliding windows, store the accessed radar information through the sliding window, and then perform multi-frame accumulation processing on the radar signal stored in the sliding window. When the radar signal from the millimeter wave radar is received at the current time, the radar information is stored in the front end of the sliding window. When entering the next time, and receiving new radar signals from the millimeter wave radar at the next time, the positions of the radar signals stored in the sliding window at the previous time of the next time are changed. Specifically, the position is compensated according to the time interval of the new radar signal and the corresponding motion state, so as to complete a sliding window operation.

[0058] The multi-frame radar signal corresponding to the completed sliding window operation is processed by density-based clustering, so as to cluster the point cloud data corresponding to the radar signals with similar characteristics, and obtain the point cloud set to be matched. The specific steps are shown in steps S13-S14. Then, by traversing the target point cloud set of the last time and calculating, the predicted point cloud set is obtained.

[0059] In a feasible implementation, step S10 includes steps S11-S12:

[0060] Step S11, initializing the target tracker;

[0061] It should be noted that the millimeter wave radar includes multiple target trackers, and one target tracker is used to track one radar signal (i.e. point cloud data). Hereinafter, one target tracker is described.

[0062] If it is judged that the target tracker has not been initialized, the state transition matrix (F) of the initialized target tracker, the covariance matrix (R) of the observation noise, and the covariance matrix (Q) of the process noise of the system state transition model are calculated.

[0063] Step S12, after traversing each target point cloud data corresponding to the target track in the target point cloud set of the last time, the motion state prediction value of each target point cloud data at the current time and the state covariance matrix prediction value at the current time are calculated by the uniform acceleration motion model, to obtain the predicted point cloud set (i.e.Figure 2 Step S6) in the method.

[0064] One target point cloud data corresponds to one target track of the vehicle, at this time, the target track corresponding to each target point cloud data in the target point cloud set obtained at the last time is traversed, and the x, y, v x , v y , a x and a y of the target track input to the Kalman filter are predicted based on the uniform acceleration motion model (i.e. Figure 2 Step S5) in the method, specifically based on expression ①.

[0065] X t (KK-1)=FX t (K-1)——Expression ①

[0066] In addition, the state covariance matrix of the target track input to the Kalman filter is predicted, specifically based on expression ②.

[0067] P t (KK-1)=FP t (K-1)F T +Q——Expression ②

[0068] Wherein, K represents the time, t represents the tth target track, X t (KK-1) represents the predicted value of the motion state of the Kth target track after information update at K-1 (i.e. the predicted value of the motion state at the current time), P t (KK-1) represents the predicted value of the state covariance matrix of the Kth target track after information update at K-1 (i.e. the predicted value of the state covariance matrix at the current time).

[0069] Step S20, according to the association between the predicted point cloud set and the to-be-matched point cloud set, an association gate is established (i.e. Figure 2 Step S7) in the method.

[0070] In order to screen the to-be-matched point cloud data included in the to-be-matched point cloud set, and obtain the effective point cloud data having association with any predicted point cloud data in the predicted point cloud set, the embodiment establishes an association gate, and judges the association between the predicted point cloud data and the to-be-matched point cloud data corresponding thereto through the association gate.

[0071] It should be noted that the point cloud data to be matched is the motion state point cloud data of a certain target actually acquired at a new moment, while the predicted point cloud data is the point cloud data of the target at the previous moment predicted based on the motion state. Therefore, the point cloud data to be matched for the same target at the same moment and the predicted point cloud data for the same target at the same moment have a corresponding relationship.

[0072] In one feasible implementation, step S20 includes steps S21 to S22:

[0073] Step S21: Traverse each of the point cloud data to be matched, and based on the traversed point cloud data to be matched, traverse the predicted point cloud data corresponding to the target point cloud data to obtain the target predicted point cloud data corresponding to each of the target point cloud data.

[0074] Step S22: Based on the point cloud data to be matched, establish the association gate on each of the target predicted point cloud data.

[0075] Specifically, by traversing each point cloud data in the set of point clouds to be matched, and during this traversal, traversing the predicted point cloud data corresponding to the traversed point cloud data, after obtaining the predicted point cloud data corresponding to each point cloud data to be matched, specific expressions ③ to ⑥ are used to establish association gates on the predicted point cloud data (i.e., Figure 2 Step S7 in the process.

[0076] Z t (KK-1)=HX t (KK-1)————Expression ③

[0077]

[0078]

[0079] S t (K)=HP t (K)H T +R————Expression ⑥

[0080] H represents the observation matrix, Z t (KK-1) represents the predicted value of the observation at time K after the information is updated at time K-1. Let z represent the prediction innovation vector between the j-th valid point cloud data and the t-th predicted point cloud data. j (K) represents the measurement value of the j-th valid point cloud data, γ is the association range of the association gate, and S t (K) is the observation error covariance matrix.

[0081] The observation matrix of the predicted point cloud data is input into expression ⑥, and the observation error covariance matrix of the predicted point cloud data in the uniform acceleration motion model is output. Meanwhile, according to the predicted point cloud data at K-1 time, the observation prediction value of the predicted point cloud data at K time is updated by using the uniform acceleration motion model through expression ③, and based on the observation prediction value, the predicted innovation vector of the predicted point cloud data and the to-be-matched point cloud data corresponding to the predicted point cloud data is obtained through expression ④. The predicted innovation vector and the corresponding observation error covariance matrix are substituted into expression ⑤ to obtain the judgment value d of the target point cloud data 2 . It is judged whether the judgment value is within the effective range γ. If the judgment value is within the effective range, it is indicated that the target point cloud data is matched with the to-be-matched point cloud data associated therewith.

[0082] In step S30, based on the association gate, the effective point cloud data in the to-be-matched point cloud set and the first point cloud data corresponding to the effective point cloud data in the predicted point cloud set are obtained, and the initial confirmation matrix is established according to the association relationship between the effective point cloud data and the first point cloud data (i.e., step S9 in Figure 2 ).

[0083] Through the association gate in step S20, if the judgment value corresponding to the to-be-matched point cloud data is within the effective range, the to-be-matched point cloud data is determined as the effective point cloud data. Since the predicted point cloud data is the predicted data of the to-be-matched point cloud data at a new time, the predicted point cloud data at the same time and the to-be-matched point cloud data have a corresponding relationship. At this time, the predicted point cloud data corresponding to the effective point cloud data is obtained, and the predicted point cloud data is determined as the first point cloud data. Then, the initial confirmation matrix is established according to the association relationship between the effective point cloud data and the first point cloud data.

[0084] In a feasible implementation, step S30 includes steps S31-S32.

[0085] In step S31, it is judged whether the target point cloud data falls within the association range of the association gate (i.e., step S8 in Figure 2 ).

[0086] In step S32, if the target point cloud data falls within the association range, the to-be-matched point cloud data corresponding to the target point cloud data falling within the association range is extracted as the effective point cloud data.

[0087] As can be seen from expressions ③ to ⑥, the association gate is used to judge whether the target point cloud data corresponding thereto is within the association range of the association gate, that is, whether the judgment value corresponding to the target point cloud data is within the effective range. If the to-be-matched point cloud data is within the association range, the to-be-matched point cloud data is determined as the effective point cloud data.

[0088] If the to-be-matched point cloud data is not within the association range, it indicates that the target motion state corresponding to the to-be-matched point cloud data does not match the motion state of the current target predicted point cloud data, and there is an abnormal motion state, so the to-be-matched point cloud data is determined as abnormal point cloud data and is not included in the subsequent process step execution calculation.

[0089] In step S40, a plurality of predicted point cloud indexes associated with each valid point cloud data in the initial confirmation matrix is obtained, valid point cloud data with the same predicted point cloud index and predicted point cloud data pointed by the predicted point cloud index are grouped into a group, a plurality of groups are obtained, and a target confirmation matrix belonging to the group is established (i.e. Figure 2 In step S10 in the foregoing embodiment, joint probability data association is performed according to the target confirmation matrix of each group, wherein the predicted point cloud set includes a plurality of predicted point cloud data.

[0090] First, according to each valid point cloud data in each initial confirmation matrix, a plurality of predicted point cloud indexes associated with each valid point cloud data is confirmed, then each valid point cloud data is traversed, if two valid point cloud data have a common predicted point cloud index, the two valid point cloud data are grouped into a group, the predicted point cloud index set of the group is the union of the common predicted point cloud indexes between the two valid point cloud data, then other valid point cloud data is traversed, whether there is valid point cloud data with the same predicted point cloud index as the predicted point cloud index set of the group is confirmed, if there is, the valid point cloud data is grouped into the group, and so on, until all valid point cloud data is grouped into a plurality of groups, the target confirmation matrix of each group is confirmed according to the plurality of groups obtained by grouping, and the valid point cloud data in each target confirmation matrix is connected with other valid point cloud data in the target confirmation matrix through the predicted point cloud index commonly associated therewith, Figure 2 One group in the foregoing embodiment is a group.

[0091] The expression ⑦ is explained in combination.

[0092]

[0093] wherein w jt represents the target confirmation matrix of j rows and t columns, n represents the group index to which the target confirmation matrix belongs, and the expression ⑦ judges whether the valid point cloud data in the target confirmation matrix has an association relationship with the first point cloud data, and specifically:

[0094] In the process of traversing each valid point cloud data, it is judged whether the first point cloud data corresponding to the traversed valid point cloud data belongs to the predicted point cloud index, if yes, the expression ⑦ outputs 1; if not, the expression ⑦ outputs 0.

[0095] The final grouped target confirmation matrix only establishes the association between the effective point cloud data within the group and the predicted point cloud data within the group, wherein the first list in the target confirmation matrix indicates that the effective point cloud data may come from clutter.

[0096] Based on the number of interconnection matrices obtained by splitting the grouped target confirmation matrix, the number of joint events corresponding to the interconnection matrices becomes approximately linear with the exponential growth of the number of point cloud data to be matched, which greatly reduces the computing resources of the joint probability data association algorithm, as described in steps S41-S44.

[0097] In a feasible implementation, step S40 includes steps S41-S44:

[0098] Step S41: Each target confirmation matrix is traversed and split to obtain a plurality of interconnection matrices (i.e., step S11 in Figure 2 , wherein each interconnection matrix represents a feasible joint event.

[0099] Each target confirmation matrix is traversed to split the target confirmation matrix to obtain an interconnection matrix, and each interconnection matrix represents a feasible joint event θ i , i = 1, 2,..., L, wherein L is the total number of feasible joint events.

[0100] Step S42: The probability of the feasible joint event is calculated based on the interconnection matrix to obtain first probability data, and the probability that the effective point cloud data included in the interconnection matrix belongs to the first point cloud data is calculated to obtain second probability data (i.e., step S12 in Figure 2 ).

[0101] The probability of the feasible joint event is calculated based on expressions ⑧-⑨.

[0102]

[0103]

[0104] , wherein T represents the number of effective point cloud data in the feasible joint event, Z K represents the motion state prediction value at time K, P D represents the probability of correct effective point cloud detection, FM represents the number of effective point cloud data in the executable joint event that does not match the predicted point cloud index, represents the innovation vector of the effective point cloud data and its associated first point cloud data, S tj represents the observation covariance matrix, M represents the dimension of the innovation vector, P n {θ iZ K} represents the first probability data.

[0105] After the first probability data is calculated based on the expressions ⑧-⑨, the second probability data is calculated based on the expression ⑩.

[0106]

[0107] wherein β jt represents the second probability data of the jth effective point cloud data belonging to the tth first point cloud data.

[0108] In step S43, according to the comparison relationship between the second probability data and the second preset probability threshold, if the second probability data is greater than the second preset probability threshold, the matching state of the effective point cloud data corresponding to the second probability data is marked as a correct matching state, wherein the first point cloud data in the correct matching state is determined as the first to-be-calculated point cloud data, and the effective point cloud data in the correct matching state is determined as the second to-be-calculated point cloud data.

[0109] In step S44, the Kalman gain and the innovation vector of the first to-be-calculated point cloud data under the uniform acceleration motion model are calculated, and the motion state prediction value and the state covariance matrix prediction value are updated according to the Kalman gain and the innovation vector, so as to obtain the motion state and the posterior state covariance of the first to-be-calculated point cloud data (i.e. Figure 2 Step S13 in the method for calculating the motion state of the point cloud data).

[0110] According to the calculated second probability data, if it is judged that the second probability data is greater than the second preset probability threshold, it is determined that the matching state of the effective point cloud data corresponding to the second probability data is a correct matching state, and then the effective point cloud data in the correct matching state is determined as the second to-be-calculated point cloud data, and the Kalman gain and the innovation vector of the first to-be-calculated point cloud data under the uniform acceleration motion model are calculated based on the expression and the expression .

[0111]

[0112] K t (K) = P(KK-1)H T (S t (K)) -1 Expression

[0113] wherein m represents the number of effective point clouds in the target confirmation matrix, V t represents the innovation vector of the tth first to-be-calculated point cloud data after calculating all the observation covariance matrices.t (K) represents the Kalman gain of the tth first to-be-calculated point cloud data.

[0114] At this time, each first point cloud data is traversed, and if the matching state of the traversed first point cloud data is a correct matching state, then the motion state prediction value and the state covariance matrix prediction value obtained based on step S12 are updated based on the expression and in combination with the Kalman gain and the innovation vector.

[0115] X t (K) = X t (KK-1) + K t (K) V t Expression

[0116]

[0117] wherein X t (K) is the updated next motion state prediction value at time K, P t (K) is the updated next state covariance matrix prediction value at time K.

[0118] In order to facilitate understanding, the following scenario is given in the embodiment.

[0119] In this scenario, there are N motor vehicle targets, and the target set T N = {1, 2,..., N}, and there are N to-be-matched point cloud data Z N = {1, 2,..., N} at the current time, it is assumed that each motor vehicle target is associated with a target point cloud data, and the motor vehicle target associated with each target point cloud data is different.

[0120] At this time, the traditional joint probability data association algorithm is as follows:

[0121] The dimension of the initial confirmation matrix: N x (N+1)

[0122] The number of interconnection matrices obtained by splitting the above initial confirmation matrix: 2 N

[0123] The dimension of the above interconnection matrix: N x (N+1)

[0124] The calculation resources occupied by the joint events corresponding to the interconnection matrix:

[0125]

[0126] The joint probability data association algorithm based on the embodiment is as follows:

[0127] Dimensions of the initial confirmation matrix: N x (N+1)

[0128] Number of target confirmation matrices based on the initial confirmation matrix split: N

[0129] Dimensions of the target confirmation matrix: 1 x 2

[0130] Number of interconnection matrices based on the target confirmation matrix split:

[0131] Dimensions of the interconnection matrix: 1 x 2

[0132] Computational resources occupied by the joint event corresponding to the interconnection matrix:

[0133]

[0134] The ratio between the computational resources obtained based on the traditional joint probability data association algorithm and the computational resources obtained based on the joint probability data association algorithm of the present embodiment is obtained as follows:

[0135]

[0136] According to the above ratio, when there is one maneuvering target vehicle, i.e. N = 1, the ratio M between the computational resources obtained based on the traditional joint probability data association algorithm and the computational resources obtained based on the joint probability data association algorithm of the present embodiment is 1, i.e. there is no difference between the computational resources occupied by the traditional joint probability data association algorithm and the computational resources occupied by the joint probability data association algorithm of the present embodiment. However, when there are ten maneuvering target vehicles, i.e. N = 10, the ratio M between the computational resources obtained based on the traditional joint probability data association algorithm and the computational resources obtained based on the joint probability data association algorithm of the present embodiment is 28160, i.e. the computational resources required by the joint probability data association algorithm of the present embodiment only need to occupy The computational resources of the joint probability data association algorithm are greatly reduced.

[0137] In this embodiment, the radar signals are processed by setting a sliding window for multi-frame accumulation, and the radar signals accessed at previous time are position compensated on the sliding window according to their motion state, so as to avoid the missing of radar signals at individual time caused by the flicker problem of radar signals. In addition, effective point cloud data is extracted from the to-be-matched point cloud set through association, an initial confirmation matrix is established through association relationship, and the initial confirmation matrix is grouped, so as to reduce the dimension of the initial confirmation matrix by partitioning the scene corresponding to each radar signal, obtain a plurality of low-dimensional target confirmation matrices, and perform joint probability data association based on the low-dimensional target confirmation matrix, thereby reducing the number of interconnected matrices obtained by splitting the initial target confirmation matrix, changing the exponential growth relationship between the number of joint events corresponding to the interconnected matrix and the number of target point cloud data into an approximately linear relationship, and greatly reducing the computing resources of the joint probability data association algorithm.

[0138] Based on the first embodiment of the present application, the same or similar contents as the above embodiment one can be referred to the above introduction, and the subsequent will not be described in detail. On this basis, please refer to Figure 3 Combined with Figure 2 , step S10, the multi-target tracking method of the millimeter wave radar further includes steps S13-S14:

[0139] Step S13, a plurality of sliding windows are set, and a plurality of radar signals output by the millimeter wave radar are stored in the corresponding sliding windows. When it is detected that the sliding window operation is completed, the radar signals stored in the plurality of sliding windows are extracted to obtain a plurality of point cloud data.

[0140] In order to reduce the random fluctuation of the radar signal, so as to more clearly show the real trend of the radar signal, a plurality of sliding windows are set for storing the radar signal in this embodiment.

[0141] Refer to Figure 2As shown, F1, F2,..., Fn-2, Fn-1, Fn, Fn+1, Fn+2 are radar signals, and the radar signals Fn-2, Fn-1, Fn, Fn+1, Fn+2 are stored in the corresponding sliding windows. If the next moment arrives and a new radar signal Fn-3 is accessed at the next moment, the last frame of data Fn+2 is deleted, the new radar signal Fn-3 is stored in the position where the radar signal Fn-2 is stored, and the radar signals Fn-2, Fn-1, Fn, Fn+1 are sequentially moved one sliding window backward for storage. The radar signals on the rear sliding windows are position compensated, for example, the radar signals Fn-2, Fn-1, Fn, Fn+1 are position compensated. Specifically, according to the time interval between the radar signals Fn-2, Fn-1, Fn, Fn+1 and the new radar signal Fn-3 and the motion state of each point cloud data stored by the radar signals Fn-2, Fn-1, Fn, Fn+1, the position of each point cloud stored by the radar signals Fn-2, Fn-1, Fn, Fn+1 is adjusted. After the adjustment is completed, it is determined that one sliding window operation is completed (that is, step S1 in Figure 2 ).

[0142] Before the radar signals are extracted, in order to ensure the effectiveness of the extracted radar signals, each radar signal on the sliding window is filtered once according to the SNRS, speed, distance, RCS and other attributes. Then, adjacent multiple frames of radar signals are accumulated and extracted to obtain multiple frames of point cloud data (that is, step S2 in Figure 2 ).

[0143] In step S14, the multiple frames of point cloud data are subjected to density-based clustering processing to obtain the to-be-matched point cloud set.

[0144] After each point cloud data on the extracted multiple frames of point cloud data is traversed and the traversed point cloud data is marked as visited, the neighborhood point cloud set of each point cloud data is found. If the number of neighborhood points is less than the minimum number of neighborhood points, the point cloud data is marked as a noise point. If the number of neighborhood points is greater than the minimum number of neighborhood points, a new cluster is created, and the point cloud data and its neighborhood point clouds are added to the cluster. Then, the cluster is expanded. For each point cloud in the neighborhood point cloud set, if the point cloud data has not been visited, the point cloud data is traversed and marked as visited. Then, the neighborhood point cloud set of the point cloud data is found. If the number of neighborhood points is greater than or equal to the minimum number of neighborhood points, the neighborhood point set of the point cloud data is added to the neighborhood point set of the cluster. If the number of neighborhood points is greater than the minimum number of neighborhood points, another new cluster is created. In this way, multiple point cloud clusters are obtained after all the point cloud data is traversed and clustering is completed (that is, step S3 in Figure 2Step S3) in the method 1000, each point cloud cluster is taken as a to-be-matched point cloud set (i.e. Figure 2 Step S4) in the method 1000, the attribute information of the to-be-matched point cloud set is calculated by synthesizing the attribute information of all point cloud data in the point cloud cluster.

[0145] In the embodiment, the radar signals are processed by multi-frame accumulation through the sliding window, the radar signals accessed at previous time are position compensated on the sliding window according to their motion states, the missing of radar signals at individual time caused by the flicker problem of the radar signals is avoided, the to-be-matched point cloud set is extracted based on the density-based clustering processing, and the point cloud data with similar characteristics is obtained, so that the to-be-matched point cloud data with similar characteristics is processed in the subsequent steps.

[0146] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as the above-mentioned first embodiment can be referred to the above description, and the subsequent will not be described in detail. On this basis, please refer to Figure 4 In combination with Figure 2 After step S43, the multi-target tracking method of the millimeter wave radar further includes steps A11-A13.

[0147] Step A11, the effective point cloud data except the second to-be-calculated point cloud data is determined as to-be-judged point cloud data in the false matching state.

[0148] Step A12, whether the to-be-judged point cloud data meets a preset creation condition is judged.

[0149] Step A13, if the to-be-judged point cloud data meets the preset creation condition, the to-be-judged point cloud data is determined as the first next target point cloud data of the next target tracking process (i.e. Figure 2 Step S14) in the method 1000.

[0150] It should be noted that steps S10-S40 are continuously executed according to the multi-frame tracking targets output by the sliding window operation, and the multi-target tracking operation based on the obtained multi-frame tracking target is an operation of tracking the target motion at the current time. Because the association probabilities of the effective point cloud data at different times and the first point cloud data are not the same, the effective point cloud data in the correct matching state is only at the current time, and the to-be-judged point cloud data in the erroneous matching state needs to be judged whether it meets the target tracking condition of the multi-target tracking operation at the next time, that is, the preset creation condition. The point cloud attribute of the to-be-judged point cloud data is judged, such as whether the speed meets the speed requirement of the preset creation condition corresponding to the next time. If it meets, the to-be-judged point cloud data is input into the multi-target tracking operation at the next time, that is, the to-be-judged point cloud data is determined as the first next target point cloud data of the next target tracking process, so as to avoid the situation that the continuity of multi-target tracking is reduced due to the lack of target point cloud data.

[0151] After step S40, the multi-target tracking method of the millimeter wave radar further includes steps S51-S52:

[0152] Step S51, each of the first point cloud data corresponding to the track and the point cloud attribute is traversed, the first point cloud data with the track state of the death track is extracted from each of the tracks, and the first point cloud data with the attribute state of the abnormal attribute is extracted from each of the point cloud attributes. After the first point cloud data with the track state of the death track and / or the first point cloud data with the attribute state of the abnormal attribute are removed, the second next target point cloud data (that is, step S15 in Figure 2 ) is obtained.

[0153] Step S52, the first next target point cloud data and the second next target point cloud data are transmitted into the next target tracking process, and the calculation of each of the tracking targets is performed (that is, step S16 in Figure 2 ).

[0154] In the embodiment, the first point cloud data that cannot be input into the multi-target tracking operation at the next moment is extracted by traversing the track corresponding to each first point cloud data and the point cloud attribute. Specifically, each first point cloud data is traversed, it is judged whether the track corresponding to the first point cloud data overlaps with the track of other first point cloud data, if there is overlapping information, the track state of the track is determined as a dead track, and it is judged whether the track corresponding to the first point cloud data is a track split from other vehicles, if yes, the track state of the track is determined as a dead track, and according to the point cloud attribute, it is judged whether the first point cloud data corresponding to the track has a track abnormal situation, such as overspeed, if yes, the attribute state of the point cloud data is determined as an abnormal attribute, the first point cloud data with the track state of a dead track and / or the attribute state of an abnormal attribute is removed, and the first point cloud data that is not removed is input into the multi-target tracking operation at the next moment as the second next target point cloud data.

[0155] In the embodiment, the next target point cloud data that can be input into the multi-target tracking operation at the next moment is obtained by screening the first point cloud data at the current moment, so as to ensure the continuity of multi-target tracking.

[0156] The application provides a multi-target tracking device of a millimeter wave radar, which comprises at least one processor and a memory in communication connection with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the multi-target tracking method of the millimeter wave radar in the above embodiment one.

[0157] Reference will be made to the following description Figure 5 which shows a structural schematic diagram of the multi-target tracking device of the millimeter wave radar suitable for being used to implement the embodiments of the application. The multi-target tracking device of the millimeter wave radar in the embodiments of the application can include but is not limited to mobile terminals such as notebook computers, PDAs (Personal Digital Assistant), PADs (Portable Application Description), PMPs (Portable Media Player), vehicle-mounted terminals (for example, vehicle-mounted navigation terminals) and the like, and fixed terminals such as digital TVs, desktop computers and the like. Figure 5 The multi-target tracking device of the millimeter wave radar shown is only an example, and should not bring any limitation to the functions and use range of the embodiments of the application.

[0158] As Figure 5As shown, the multi-target tracking device of the millimeter wave radar can include a processing device 1001 (e.g., a central processor, a graphics processor, etc.) that can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 1002 or loaded from a storage device 1003 into a random access memory (RAM) 1004. Various programs and data required for the operation of the multi-target tracking device of the millimeter wave radar are also stored in the RAM 1004. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; the storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the multi-target tracking device of the millimeter wave radar to communicate with other devices wirelessly or by wire to exchange data. Although the multi-target tracking device of the millimeter wave radar with various systems is shown in the figure, it should be understood that all the systems shown are not required to be implemented or possessed. More or fewer systems can be alternatively implemented or possessed.

[0159] In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by a communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are performed.

[0160] The multi-target tracking device of the millimeter wave radar provided by the present disclosure adopts the multi-target tracking method of the millimeter wave radar in the above-mentioned embodiments, and can solve the technical problem that the current multi-target tracking based on the joint probability data association algorithm has too much computing resources. Compared with the prior art, the multi-target tracking device of the millimeter wave radar provided by the present disclosure has the same beneficial effects as the multi-target tracking method of the millimeter wave radar provided by the above-mentioned embodiments, and other technical features in the multi-target tracking device of the millimeter wave radar are the same as the features disclosed in the previous embodiment method, which will not be repeated here.

[0161] It is to be understood that the various parts of the disclosure can be implemented in hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0162] The above description is merely illustrative of the application and not restrictive.

[0163] The application provides a computer readable storage medium having stored thereon computer readable program instructions (i.e., computer programs) for performing the multi-target tracking method of the millimeter wave radar in the above embodiments.

[0164] The computer readable storage medium provided by the application may, for example, be a U disk, but is not limited to an electric, magnetic, optical, electromagnetic, infrared, or semiconductor system, system or device, or any combination thereof. More specific examples of the computer readable storage medium can include, but are not limited to, an electric connection with one or more conductive wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present embodiment, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system or device. The program code contained on the computer readable storage medium can be transmitted by any suitable medium, including but not limited to an electric wire, an optical cable, an RF (Radio Frequency), etc., or any suitable combination thereof.

[0165] The above computer readable storage medium can be included in the multi-target tracking device of the millimeter wave radar; or can exist separately without being assembled into the multi-target tracking device of the millimeter wave radar.

[0166] The computer readable storage medium described above carries one or more programs, when the one or more programs are executed by the multi-target tracking device of the millimeter wave radar, the multi-target tracking device of the millimeter wave radar is caused to: by setting a plurality of sliding windows, performing multi-frame accumulation processing and density-based clustering processing on the accessed radar signal to obtain a to-be-matched point cloud set, and by traversing the to-be-matched point cloud set, a predicted point cloud set is calculated; according to the association relationship between the predicted point cloud set and the to-be-matched point cloud set, an association gate is established; based on the association gate, effective point cloud data in the to-be-matched point cloud set and first point cloud data corresponding to the effective point cloud data in the predicted point cloud set are obtained, and according to the association relationship between the effective point cloud data and the first point cloud data, an initial confirmation matrix is established; a plurality of predicted point cloud indexes associated with each effective point cloud data in the initial confirmation matrix are obtained, effective point cloud data and predicted point cloud data having the same predicted point cloud index are grouped into a group, a plurality of groups are obtained, and a target confirmation matrix belonging to the group is established, and joint probability data association is performed according to the target confirmation matrix of each group, wherein the predicted point cloud set includes a plurality of predicted point cloud data.

[0167] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0168] The flow and block diagrams in the drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flow and block diagrams can represent a module, a segment, or a portion of code, which comprises one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or in the reverse order, depending on the functionality involved. It will also be noted that each block of the block diagrams and / or flowchart illustrations, and combinations thereof, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or combinations of hardware and software.

[0169] The modules involved in the embodiments of the present application can be implemented in the form of software or in the form of hardware. In some cases, the name of the module does not constitute a limitation on the module itself.

[0170] The readable storage medium provided by the present application is a computer readable storage medium, which stores computer readable program instructions (i.e. computer programs) for executing the multi-target tracking method of the millimeter wave radar, and can solve the technical problem of excessive computing resources in the multi-target tracking based on the joint probability data association algorithm. Compared with the prior art, the beneficial effects of the computer readable storage medium provided by the present application are the same as those of the multi-target tracking method of the millimeter wave radar provided by the above-mentioned embodiments, and are not described here.

[0171] The above only describes some embodiments of the present application, and does not limit the patent scope of the present application. Any equivalent structural transformation made by using the content of the present application specification and drawings, or direct / indirect application in other related technical fields is included in the patent protection scope of the present application.

Claims

1. A multi-target tracking method of a millimeter wave radar, characterized by, The multi-target tracking method of the millimeter wave radar comprises the following steps: A plurality of sliding windows are set, and a plurality of radar signals output by the millimeter wave radar are stored in the sliding windows in sequence. The radar signals stored in the sliding windows are extracted to obtain a plurality of point cloud data. The plurality of point cloud data are subjected to density-based clustering processing to obtain a point cloud set to be matched. An association gate is established according to an association relationship between the point cloud set to be matched and the predicted point cloud set.

2. The multi-target tracking method of a millimeter wave radar according to claim 1, characterized by, Based on the association gate, effective point cloud data in the point cloud set to be matched and first point cloud data corresponding to the effective point cloud data in the predicted point cloud set are obtained, and an initial confirmation matrix is established according to an association relationship between the effective point cloud data and the first point cloud data. A plurality of predicted point cloud indexes associated with each effective point cloud data in the initial confirmation matrix are obtained, effective point cloud data and predicted point cloud data having the same predicted point cloud index are grouped into a group, a plurality of groups are obtained, a target confirmation matrix belonging to the group is established, and joint probability data association is performed according to the target confirmation matrix of each group. The step of obtaining the point cloud set to be matched by setting a plurality of sliding windows and performing multi-frame accumulation processing and density-based clustering processing on the accessed radar signals comprises the following steps:

3. The multi-target tracking method of a millimeter wave radar according to claim 2, characterized by, A plurality of sliding windows are set, and a plurality of radar signals output by the millimeter wave radar are stored in the sliding windows in sequence. The radar signals stored in the sliding windows are extracted to obtain a plurality of point cloud data. The plurality of point cloud data are subjected to density-based clustering processing to obtain a point cloud set to be matched.

4. The multi-target tracking method of a millimeter wave radar according to claim 3, characterized by, The millimeter wave radar comprises a plurality of target trackers, and one target tracker is used to track one point cloud data. The target tracker is initialized. After each target point cloud data in the target point cloud set of the previous moment is traversed, a uniform acceleration motion model is used to calculate a motion state prediction value of each target point cloud data at the current moment and a state covariance matrix prediction value at the current moment to obtain the predicted point cloud set.

5. The multi-target tracking method of a millimeter wave radar according to claim 4, characterized by, The step of establishing the association gate according to the association relationship between the predicted point cloud set and the point cloud set to be matched comprises the following steps: Each target point cloud data in the point cloud set to be matched is traversed, and the predicted point cloud data corresponding to the target point cloud data is traversed according to the target point cloud data traversed to obtain target predicted point cloud data corresponding to each target point cloud data. The association gate is established on each target predicted point cloud data according to the point cloud data to be matched. The step of obtaining the effective point cloud data in the point cloud set to be matched based on the association gate comprises the following steps: It is judged whether the target point cloud data falls within the association range of the association gate through the association gate. If the target point cloud data falls into the association range, the to-be-matched point cloud data corresponding to the target point cloud data falling into the association range is extracted as the effective point cloud data.

6. The multi-target tracking method of a millimeter wave radar according to claim 5, wherein, The step of performing joint probability data association according to each target confirmation matrix comprises: Each target confirmation matrix is traversed and split to obtain a plurality of interconnection matrices, wherein each interconnection matrix represents a feasible joint event; Probabilities of the feasible joint events are calculated according to the interconnection matrices to obtain first probability data, and probabilities that the effective point cloud data included in the interconnection matrices belong to the first point cloud data are calculated to obtain second probability data; According to a comparison relationship between the second probability data and a second preset probability threshold, if the second probability data is greater than the second preset probability threshold, a matching state of the effective point cloud data corresponding to the second probability data is marked as a correct matching state, wherein the first point cloud data in the correct matching state is determined as first to-be-calculated point cloud data, and the effective point cloud data in the correct matching state is determined as second to-be-calculated point cloud data; The Kalman gain and the innovation vector of the first to-be-calculated point cloud data under the uniform acceleration motion model are calculated, and the motion state prediction value and the state covariance matrix prediction value are updated according to the Kalman gain and the innovation vector to obtain the motion state and the posterior state covariance of the first to-be-calculated point cloud data.

7. The multi-target tracking method of a millimeter wave radar according to claim 6, wherein After the step of marking the matching state of the effective point cloud data corresponding to the second probability data as the correct matching state if the second probability data is greater than the second preset probability threshold, the method comprises: The effective point cloud data except the second to-be-calculated point cloud data is determined as to-be-judged point cloud data in an error matching state; It is judged whether the to-be-judged point cloud data meets a preset creation condition; If the to-be-judged point cloud data meets the preset creation condition, the to-be-judged point cloud data is determined as first next target point cloud data of a next target tracking process.

8. The multi-target tracking method of a millimeter wave radar according to claim 7, wherein, After the step of performing joint probability data association according to the target confirmation matrix, the method comprises: Each first point cloud data corresponding to a track and a point cloud attribute is traversed, first point cloud data with a track state of a dead track is extracted from each track, and first point cloud data with an attribute state of an abnormal attribute is extracted from each point cloud attribute, and then the first point cloud data with the track state of the dead track and / or the first point cloud data with the attribute state of the abnormal attribute are removed to obtain second next target point cloud data; The first next target point cloud data and the second next target point cloud data are transmitted into a next target tracking process.

9. A multi-target tracking device for a millimeter wave radar, characterized by, The device comprises a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the multi-target tracking method of the millimeter wave radar according to any one of claims 1 to 8.

10. A storage medium, characterized by The storage medium is a computer readable storage medium, and the storage medium stores a computer program. The computer program is executed by a processor to implement the steps of the multi-target tracking method of the millimeter wave radar according to any one of claims 1 to 8.

Images