Maneuvering target tracking method and device, storage medium and electronic equipment

Abstract

The invention discloses a maneuvering target tracking method and device, a storage medium and electronic equipment. The method comprises the following steps: determining a first noise matrix and a second noise matrix; based on a measurement data set of a plurality of continuous data frames, the first noise matrix and the second noise matrix, a first probability and a second probability corresponding to each data frame are calculated, the first probability is used for indicating the probability that the target motion model of the current data frame is matched with the target motion trail, and the second probability is used for indicating the target motion trail; the second probability is used for indicating the probability that the target motion model of the current data frame is not matched with the target motion track; and under the condition that the first probability and the second probability of the continuous multiple frames meet a preset condition, determining that the target motion track is not matched with the target motion model, and updating parameters of the target motion model. The technical problem that a target tracking algorithm provided by the related technology is low in resource utilization rate is solved.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and more specifically, to a method and device for tracking a maneuvering target, a storage medium, and an electronic device. Background Art

[0002] In the intelligent driving system, the on-board radar can identify the motion of the target vehicle on the road. When the target vehicle suddenly changes speed, turns, or makes a U-turn, the on-board radar needs to identify the mismatch between the current motion model and the actual motion model in order to change the tracking parameters to better track the target vehicle.

[0003] In the process of tracking a target, the related art usually uses a multi-model-based mobile target tracking method, uses multiple motion models to track the target, and uses multiple motion models to calculate the tracking result of the target vehicle, and selects the tracking result of the model with the highest probability as the tracking result of the target vehicle. However, because this method uses multiple motion models, it requires high storage requirements and computational complexity. In other words, the target tracking algorithm provided by the related art has the problem of low resource utilization.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present application provide a mobile target tracking method and device, a storage medium and an electronic device to at least solve the technical problem that the target tracking algorithm provided by the related art has low resource utilization.

[0006] According to one aspect of an embodiment of the present application, a method for tracking a maneuvering target is provided, including: determining a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; calculating a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, the first noise matrix, and the second noise matrix, wherein the first probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame do not match; when the first probability and the second probability of the plurality of consecutive frames meet a preset condition, determining that the target motion trajectory does not match the target motion model, and updating the parameters of the target motion model.

[0007] According to another aspect of an embodiment of the present application, a maneuverable target tracking device is also provided, including: a noise matrix determination unit, used to determine a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; a probability calculation unit, used to calculate the first probability and the second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, the first noise matrix and the second noise matrix, wherein the first probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame do not match; a maneuver judgment unit, used to determine that the target motion trajectory does not match the target motion model when the first probability and the second probability of a plurality of consecutive frames meet preset conditions, and update the parameters of the target motion model.

[0008] Optionally, the above-mentioned probability calculation unit includes: a first probability calculation module, which is used to calculate the first probability and the second probability corresponding to the first data frame based on the measurement data set of the first data frame, the first noise matrix and the second noise matrix, including: calculating the likelihood probability pairs corresponding to each measurement data of the first data frame according to the measurement data set of the first data frame, the first noise matrix and the second noise matrix; using the initial prior probability pair and the likelihood probability pair corresponding to each measurement data of the first data frame to calculate the first probability and the second probability corresponding to the first data frame, wherein the initial prior probability pair is used to indicate the probability of matching and the probability of not matching the target motion model and the target motion trajectory when the measurement data is not received; a second probability calculation module, which is used to calculate the likelihood probability pairs corresponding to each measurement data of the first data frame based on the measurement data set of the kth data frame. The method comprises: calculating a first probability and a second probability corresponding to the k-th data frame according to a measurement data set, a first noise matrix and a second noise matrix, wherein k is an integer greater than 1, and comprising: calculating a prior probability pair corresponding to the k-th data frame according to the first probability and the second probability corresponding to the k-1-th data frame, wherein the prior probability pair is used to indicate the probability of a match and a probability of a mismatch between the target motion model and the target motion trajectory when the measurement data of the k-th data frame is not received; calculating a likelihood probability pair corresponding to each measurement data of the k-th data frame according to the measurement data set of the k-th data frame, the first noise matrix and the second noise matrix; and calculating the first probability and the second probability corresponding to the k-th data frame using the prior probability pair corresponding to the k-th data frame and the likelihood probability pair corresponding to each measurement data of the k-th data frame.

[0009] Optionally, the above-mentioned probability calculation unit includes: a first parameter calculation module, which is used to obtain a first parameter using a first prior probability in a priori probability pair corresponding to an initial radar parameter set and a k-th data frame, and to obtain a second parameter using a second prior probability in a priori probability pair corresponding to an initial radar parameter set and a k-th data frame, wherein the first parameter and the first prior probability correspond to a situation where the target motion model and the target motion trajectory match, and the second parameter and the second prior probability correspond to a situation where the target motion model and the target motion trajectory do not match; a second parameter calculation module, which is used to sum the first likelihood probability in the likelihood probability pair corresponding to each measurement data of the k-th data frame with the first prior probability corresponding to the k-th data frame. The third parameter is obtained by multiplying the first and second likelihood probabilities of the likelihood probability pairs corresponding to the measurement data of the k-th data frame by the second prior probability corresponding to the k-th data frame to obtain a fourth parameter, wherein the third parameter and the first likelihood probability correspond to the case where the target motion model and the target motion trajectory match, and the fourth parameter and the second likelihood probability correspond to the case where the target motion model and the target motion trajectory do not match; a probability determination module is used to determine the first probability as the ratio of the sum of the first parameter and the third parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter; and determine the second probability as the ratio of the sum of the second parameter and the fourth parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter.

[0010] Optionally, the above-mentioned probability calculation unit includes: a mean calculation module, which is used to input the predicted state of the maneuvering target in the kth data frame into the measurement model of the radar to obtain a measurement mean, wherein the predicted state of the maneuvering target is determined according to the target motion model; a covariance matrix calculation module, which is used to obtain a first covariance matrix and a second covariance matrix using a first noise matrix, a second noise matrix and the predicted state of the maneuvering target corresponding to the kth data frame; a Gaussian distribution function calculation module, which is used to obtain a first Gaussian distribution function according to the measurement mean and the first covariance matrix, and to obtain a second Gaussian distribution function according to the measurement mean and the second covariance matrix; a likelihood probability calculation module, which is used to input the measurement data of the kth data frame into the first Gaussian distribution function and the second Gaussian distribution function, respectively, to obtain a likelihood probability pair.

[0011] Optionally, the above-mentioned probability calculation unit includes: a prior probability calculation module, used to multiply the first probability and the second probability corresponding to the k-1th data frame with the transition probability matrix to obtain a prior probability pair corresponding to the kth data frame, wherein the transition probability matrix is ​​used to indicate the probability of a change in the matching relationship between the target motion model and the target motion trajectory in the k-1th data frame and the kth data frame.

[0012] Optionally, the above-mentioned mobile target tracking device also includes: an initialization unit, used to: set the initial values ​​of the radar's clutter density, detection probability, and gate probability, and determine the initial values ​​of the radar's clutter density, detection probability, and gate probability as an initial radar parameter set; set the value of a transfer matrix, wherein the transfer matrix includes the probability that the target motion model and the target motion trajectory of two consecutive data frames are both in a matching state, the probability that the target motion model and the target motion trajectory of two consecutive data frames change from a matching state to a non-matching state, the probability that the target motion model and the target motion trajectory of two consecutive data frames change from a non-matching state to a matching state, and the probability that the target motion model and the target motion trajectory of two consecutive data frames are both in a non-matching state.

[0013] Optionally, the above-mentioned maneuver judgment unit also includes: a counting module, used to add one to the counting parameter when the first probability and the second probability in a data frame meet the preset conditions; clear the counting parameter when the first probability and the second probability in a data frame do not meet the preset conditions; and a judgment module, used to determine that the target motion trajectory does not match the target motion model when the counting parameter is greater than the counting threshold.

[0014] Optionally, the counting module is further used to determine that a preset condition is met when a ratio of the first probability to the second probability is less than a preset ratio.

[0015] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned maneuvering target tracking method when running.

[0016] According to another aspect of the embodiment of the present application, a computer program product is provided, the computer program product including a computer program / instruction, the computer instruction being stored in a computer-readable storage medium. A processor of a computer device reads the computer program / instruction from the computer-readable storage medium, and the processor executes the computer program / instruction, so that the computer device performs the above maneuvering target tracking method.

[0017] According to another aspect of the embodiments of the present application, there is also provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-mentioned maneuvering target tracking method through the above-mentioned computer program.

[0018] In an embodiment of the present application, a first noise matrix and a second noise matrix are determined, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; based on the measurement data set of multiple consecutive data frames, the first noise matrix and the second noise matrix, the first probability and the second probability corresponding to each data frame are calculated, wherein the first probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame do not match; when the first probability and the second probability of the consecutive multiple frames meet the preset conditions, it is determined that the target motion trajectory does not match the target motion model, and the parameters of the target motion model are updated. Different noise matrices are used as process noise covariances to characterize whether the target motion model and the target motion trajectory match, and the target position and Doppler information provided by the multiple measurement data in the gate are combined to determine the first probability of the matching state and the second probability of the non-matching state, and by comparing the first probability and the second probability, it is determined whether the target is maneuvering, and the target is tracked after the maneuvering. The above method solves the technical problem of low resource utilization of the target tracking algorithm provided by the related technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 is a schematic diagram of a hardware environment of an optional mobile target tracking method according to an embodiment of the present application;

[0021] Figure 2 is a flow chart of an optional maneuvering target tracking method according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of an optional maneuvering target tracking method according to an embodiment of the present application;

[0023] Figure 4 is a schematic diagram of another optional maneuvering target tracking method according to an embodiment of the present application;

[0024] Figure 5 is a schematic diagram of another optional maneuvering target tracking method according to an embodiment of the present application;

[0025] Figure 6 is a schematic diagram of another optional maneuvering target tracking method according to an embodiment of the present application;

[0026] Figure 7 is a schematic diagram of another optional maneuvering target tracking method according to an embodiment of the present application;

[0027] Figure 8 is a schematic diagram of an optional mobile target tracking device according to an embodiment of the present application;

[0028] Fig. 9 is a schematic diagram of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] According to one aspect of an embodiment of the present application, a mobile target tracking method is provided. As an optional implementation, the mobile target tracking method can be applied to, but is not limited to, Figure 1 A mobile target tracking system in the hardware environment shown. Optionally, the mobile target tracking method described above can be applied to a vehicle terminal. Figure 1A side view of a vehicle terminal 101 is shown, which is disposed on a travel surface 113 and is capable of traversing the travel surface 113. The vehicle terminal 101 includes an onboard navigation system 103, a computer-readable storage device or medium (memory) 102 including a digitized road map 104, a spatial monitoring system 117, a vehicle controller 109, a GPS (global positioning system) sensor 110, an HMI (human / machine interface) device 111, and also includes an autonomous controller 112 and a telematics controller 114. The vehicle terminal 101 may include, but is not limited to, commercial vehicles, industrial vehicles, agricultural vehicles, passenger vehicles, all-terrain vehicles, personal mobile devices, robots, and similar forms of mobile platforms to achieve the purpose of the present application.

[0032] In one embodiment, the spatial monitoring system 117 includes: one or more spatial sensors and systems arranged to monitor the visible area 105 in front of the vehicle terminal 101; and a spatial monitoring controller 118. The spatial sensors used to monitor the visible area 105 include, for example, a laser radar sensor 106, a radar sensor 107, a camera 108, and the like. The placement of the spatial sensors allows the spatial monitoring controller 118 to monitor traffic flow, including approaching vehicles, intersections, lane markings, and other objects around the vehicle terminal 101. The spatial sensors of the spatial monitoring system 117 may include object positioning sensing devices. The laser radar sensor 106 uses pulsed and reflected laser beams to measure the range or distance to the object. The radar sensor 107 uses radio waves to determine the range, angle, and / or speed of the object. The camera 108 includes an image sensor, a lens, and a camera controller. The laser radar sensor 106 and the radar sensor 107 can collect measurement data sets of targets on the road to generate each data frame.

[0033] The camera 108 is advantageously mounted and positioned on the vehicle terminal 101 in a position that allows capturing images of the visible area 105, wherein at least a portion of the visible area 105 includes a portion of the travel surface 113 in front of the vehicle terminal 101 and including the trajectory of the vehicle terminal 101. The visible area 105 may also include the surrounding environment. Other cameras (not shown) may also be used, for example, including a second camera disposed on the rear or side portion of the vehicle terminal 101 to monitor the rear of the vehicle terminal 101 and one of the right or left sides of the vehicle terminal 101. The camera 108 can collect road information while the vehicle terminal 101 is traveling.

[0034] The autonomous controller 112 is configured to implement autonomous driving or advanced driver assistance system (ADAS) vehicle functionality. Such functionality may include a vehicle onboard control system capable of providing a certain level of driving automation. Driving automation may include a series of dynamic driving and vehicle operations. Driving automation may include simultaneous automatic control of vehicle driving functions (including steering, acceleration and braking), wherein the driver abandons control of the vehicle for a period of time during the trip. Driving automation may include simultaneous automatic control of vehicle driving functions (including steering, acceleration and braking), wherein the driver abandons control of the vehicle terminal 101 for the entire trip. Driving automation includes hardware and a controller, which is configured to monitor the spatial environment in various driving modes to perform various driving tasks during dynamic vehicle operations. By way of non-limiting example, autonomous vehicle functions include adaptive cruise control (ACC) operations, lane guidance and lane keeping operations, lane change operations, steering assistance operations, object avoidance operations, parking assistance operations, vehicle braking operations, vehicle speed and acceleration operations, vehicle lateral motion operations, for example, as part of lane guidance, lane keeping and lane change operations, etc.

[0035] The autonomous controller can be equipped with an operating system and an automatic driving system. Among them, the operating system can be responsible for the management of the hardware resources (including sensors, system buses, networks, etc.) of the vehicle terminal 101 and the scheduling of computing resources. The automatic driving system can implement various algorithms required for automatic driving, including positioning, environmental perception, path planning and control, etc., and can make decisions in situations such as driving on curves, driving in a straight line, driving on complex roads, changing lanes, etc. The autonomous controller can determine the maneuvering target when the vehicle is driving and obtain a set of measurement data.

[0036] The vehicle terminal 101 may include a telematics controller 114, which includes a wireless telematics communication system capable of performing off-vehicle communications (including communicating with a communication network 115 having wireless and wired communication capabilities). Optionally or additionally, the telematics controller 114 directly performs off-vehicle communications by communicating with an off-board server 116 via the communication network 115. The server side may perform steps S1-S3 to determine a first noise matrix and a second noise matrix; calculate a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, a first noise matrix and a second noise matrix; when the first probability and the second probability of the plurality of consecutive frames meet a preset condition, determine that the target motion trajectory does not match the target motion model, and update the parameters of the target motion model.

[0037] In an optional embodiment, the vehicle terminal 101 can collect targets on the road through the space monitoring system 117 using the laser radar sensor 106 and the radar sensor 107 to generate measurement data sets of each data frame, and can track the targets. Mobile targets such as surrounding vehicles, pedestrians and other traffic participants. The collected measurement data sets can be stored in the memory 102.

[0038] In an optional embodiment, the maneuvering target tracking method of the present application can be applied to the vehicle side. During the driving process of the vehicle, the radar system can obtain the measurement data of the surrounding targets and track the surrounding targets to determine the trajectory of the target. The target can be observed first to determine the motion characteristics of the target, and a suitable motion model can be determined based on the motion characteristics of the target to track the target. After the motion model is determined, the target may maneuver (such as sudden turns, acceleration, deceleration, etc.), and the vehicle can adjust the model parameters to adapt to the new motion state of the target. When judging whether the target maneuvers, the probability of the target maneuvering can be calculated.

[0039] In an embodiment of the present application, a first noise matrix and a second noise matrix are determined, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; based on the measurement data set of multiple consecutive data frames, the first noise matrix and the second noise matrix, the first probability and the second probability corresponding to each data frame are calculated, wherein the first probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame do not match; when the first probability and the second probability of the consecutive multiple frames meet the preset conditions, it is determined that the target motion trajectory does not match the target motion model, and the parameters of the target motion model are updated. Different noise matrices are used as process noise covariances to characterize whether the target motion model and the target motion trajectory match, and the target position and Doppler information provided by the multiple measurement data in the gate are combined to determine the first probability of the matching state and the second probability of the non-matching state, and by comparing the first probability and the second probability, it is determined whether the target is maneuvering, and the target is tracked after the maneuvering. The above method solves the technical problem of low resource utilization of the target tracking algorithm provided by the related technology.

[0040] In an optional embodiment, if Figure 2 As shown, the above mobile target tracking method includes the following steps:

[0041] S202, determining a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match;

[0042] It should be noted that the first noise matrix and the second noise matrix can be a process noise covariance matrix Q, which is used to quantify the covariance matrix of the target motion model prediction error. It describes the uncertainty caused by imperfect models or unmodeled dynamic factors during the state transition process. Each element of the Q matrix is ​​the covariance between the element errors of state X.

[0043] It should be noted that the target motion model can be a motion model established by the radar system according to the actual motion of the target during the process of tracking the target. The target motion model is a mathematical model of the position, velocity, acceleration and other states of the target that change over time. The target motion model can be a uniform motion model, which assumes that the target moves along a straight line at a constant speed; it can be a uniform acceleration motion model, which assumes that the target moves at a constant acceleration, which is suitable for situations where the acceleration does not change much; it can be a uniform turning model, where the speed and turn rate of the target remain constant when turning. The target motion model can also be a first-order time-correlated model, etc., and there is no restriction on the specific form of the target motion model.

[0044] It should be noted that the target motion trajectory may be an actual motion model or motion trajectory of the target, and may indicate the actual speed and position of the target at the current moment.

[0045] In an optional embodiment, different events can be set, such as θ k = 1 can represent the event that the target motion model and target motion trajectory of the maneuvering target match at time k, and the corresponding process noise covariance matrix Q can be set 1 is the first noise matrix, indicating the uncertainty level when the model is matched; if θ k = 2 can represent the event that the target motion model and target motion trajectory of the maneuvering target do not match at time k, and the corresponding process noise covariance matrix Q can be set 2 is the second noise matrix to indicate the uncertainty level when the model does not match.

[0046] In an optional embodiment, during the target motion process, the probability of maneuvering is low, and during the target motion process, the target is in a stable state most of the time. Therefore, the first noise matrix can be set to be smaller, and the second noise matrix can be set to be larger, so as to indicate that the uncertainty of the match between the target motion model and the target motion trajectory is small, and the uncertainty of the mismatch between the target motion model and the target motion trajectory is large.

[0047] In an optional embodiment, a reference process covariance matrix Q can be set b , then the first noise matrix and the second noise matrix can be set as follows:

[0048]

[0049] Where a is a positive integer greater than 1, c 1 、c 2 Both can be natural numbers.

[0050] S204, calculating a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, a first noise matrix, and a second noise matrix, wherein the first probability is used to indicate a probability that a target motion model of a current data frame matches a target motion trajectory, and the second probability is used to indicate a probability that a target motion model of a current data frame does not match a target motion trajectory;

[0051] It should be noted that the radar system can collect surrounding measurement data, of which some measurement data originate from the target and some measurement data do not originate from the target. k,i Represents the event "k time measurement z k,i From the goal". k,0 Indicates the event "no measurement originating from the target at time k". Measurement data usually refers to the raw observations obtained directly from the sensor. These data can be continuous or discrete values, which are used to describe certain specific attributes or states of the observed target. In a radar system, measurement data may include: Distance: the straight-line distance between the target and the radar. Speed: the speed of the target relative to the radar, which may be measured by the Doppler effect. Azimuth and elevation: the direction of the target relative to the radar antenna. Reflection intensity: the intensity of the target's reflected signal, which can be used to infer the size or material of the target. Timestamp: the specific time when the measurement data is obtained.

[0052] χ k,i It can represent the i-th measurement data z at time k k,i is directly caused by the target. This measurement data is considered to be the real reflection of the target, not from background clutter or other interference sources. In the radar system, each z k,iIt can be the echo signal generated by the target at the radar receiver, including information such as distance, speed, angle, etc. k,0 It can be said that at time k, no measurement data is caused by the target. That is to say, at this moment, the radar does not receive any valid signal from the target, and all received signals may be clutter, interference or noise. k,0 The occurrence of may mean that the target is not detected at that moment, or the target's signal is masked by environmental factors (such as obstacles, multipath effects, etc.).

[0053] Figure 3 is a schematic diagram of an optional mobile target tracking method according to an embodiment of the present application; Figure 3 As shown, in the wave gate, there can be point cloud data and noise data about the target.

[0054] In an alternative embodiment, by calculating P{θ k |Z k}, it can be identified whether the target has maneuvered. k ={z k ,Z k-1} represents the measured data from the initial moment to the k moment, represents the measurement data at time k, z k,i represents the i∈{1,2,…n at time k k} measurement data, n k is the number of measurements at time k.

[0055] It should be noted that P{θ k |Z k} is given all historical measurement data Z k The probability that the target motion model matches the actual motion trajectory (target motion trajectory) under the condition of . By calculating the posterior probability, it can be determined whether the motion state of the target has changed. Event θ k Can be composed of the following mutually exclusive events:

[0056]

[0057] That is, it includes the case where the measurement data of the target originates from the target and the case where no measurement data originates from the target, so as to comprehensively evaluate the motion state of the target.

[0058] Therefore, the probability can be calculated as follows:

[0059]

[0060] That is, the event θ k The probability P{θ k |Z kThe calculation of} can be divided into two cases. Case 1: Event θ k The probability P{θ k , X k,0 |Z k}; Case 2: Event θ k and event "k moment measurement z k,i The probability of "originating from the target" occurring at the same time.

[0061] It should be noted that P{θ k , X k,0 |Z k} can be calculated by the following formula:

[0062] P{θ k , X k,0 |Z k}∝P{z k |θ k , X k,0 , n k , Z k-1}P{n k |θ k , X k,0 , Z k-1}P{θ k , X k,0 |Z k-1}

[0063] Among them, ∝ is the proportional symbol. Since all measurements in the wave gate are noise, then:

[0064]

[0065] Among them, V k Indicates the area of ​​the wave gate. The wave gate can be understood as the radar continuous illumination area. When using radar to detect a target, the wave gate needs to be used to trap the target after the target is found, that is, the radar continuously illuminates the selected target to obtain a continuous and complete target motion vector.

[0066] Assuming that the number of clutter in the gate follows a Poisson distribution, then:

[0067]

[0068] Where λ is the clutter density. Clutter density can indicate the average number of clutter per unit volume or unit area in the radar detection space. It can be used to describe the distribution characteristics of clutter and is an important parameter for evaluating the clutter intensity in the radar working environment.

[0069] Use P D represents the detection probability, P G represents the gate probability, then:

[0070] P{θ k , χ k,0 |Z k-1}=(1-P D P G )P{θ k |Z k-1}

[0071] Among them, P{θ k |Z k-1} represents the event θ k The detection probability refers to the probability that the system makes a correct judgment of "there is a signal" when there is indeed a signal at the receiving end. The gate probability is used to indicate the probability of containing a target in the wave gate.

[0072] Optionally, the calculation formula of the prior probability may be:

[0073]

[0074] Among them, P{θ k =m|θ k-1 =i,Z k-1} represents the transition probability, m∈{1,2}, P{θ k-1 =i|Z k-1} represents the event θ at time k-1 k-1 The first probability or the second probability.

[0075] It should be noted that for case 2, P{θ k , χ k，i |Z k} can be calculated using the following formula:

[0076] P{θ k , χ k，i |Z k}∝P{z k |θ k , χ k，i , n k , Z k-1}P{n k |θ k , χ k，i , Z k-1}

[0077] ×P{ k，i |θ k , Z k-1}P{θ k |Z k-1}

[0078] in,

[0079]

[0080]

[0081]

[0082] Among them, p(z k，i |θ k , Z k-1 ) is the likelihood probability.

[0083] In an optional implementation, the calculation formulas of the two cases are substituted into the total probability formula to obtain the calculation formula of the first probability:

[0084]

[0085] The calculation formula for the second probability is:

[0086]

[0087] in,

[0088]

[0089]

[0090]

[0091]

[0092] That is, the first probability and the second probability can be calculated based on the prior probability pair and the likelihood probability pair. How to determine the first probability and the second probability in an actual process will be described in detail later, and will not be repeated here.

[0093] S206: When the first probability and the second probability of the plurality of consecutive frames meet the preset condition, it is determined that the target motion trajectory does not match the target motion model, and the parameters of the target motion model are updated.

[0094] It should be noted that in step S204, the system has calculated the probabilities of the target matching (first probability) and mismatching (second probability) for each data frame. In S206, the system will analyze these probability sequences, especially focusing on whether the second probability is significantly higher than the first probability in multiple consecutive frames, which indicates that the current motion model may not correctly describe the actual motion state of the target.

[0095] In an optional embodiment, the condition usually includes one or more thresholds and / or thresholds for determining whether the target motion trajectory does not match the motion model. For example, the system may set a probability ratio threshold, and when the ratio of the second probability to the first probability exceeds this threshold for multiple consecutive frames, or when the second probability itself exceeds an absolute threshold continuously, the system will determine that the above preset conditions are met. Once the first probability and the second probability for multiple consecutive frames meet the above preset conditions, that is, the second probability continues to be high, the system determines that the actual motion trajectory of the target does not match the current motion model, which usually means that the target has performed some kind of maneuver, such as sudden acceleration, deceleration, steering or U-turn.

[0096] It should be noted that in order to respond to the maneuvering state of the target, the system will update the parameters of the target motion model to improve the match between the model and the actual motion state of the target. This may involve adjusting the state transfer matrix, process noise matrix, measurement noise matrix, etc. to reflect the motion characteristics of the target after maneuvering. The updated parameters enable the tracking algorithm to more accurately predict the future position of the target, thereby improving the tracking performance and safety of the radar system. Specifically, the following strategies may be adopted to update the parameters of the target motion model: adjust the process noise matrix (Q) and increase its value to reflect the increased uncertainty of the target motion state. Modify the measurement noise matrix (R) to cope with possible changes in measurement errors during maneuvers, etc.

[0097] In an embodiment of the present application, a first noise matrix and a second noise matrix are determined, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; based on the measurement data set of multiple consecutive data frames, the first noise matrix and the second noise matrix, the first probability and the second probability corresponding to each data frame are calculated, wherein the first probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame do not match; when the first probability and the second probability of the consecutive multiple frames meet the preset conditions, it is determined that the target motion trajectory does not match the target motion model, and the parameters of the target motion model are updated. Different noise matrices are used as process noise covariances to characterize whether the target motion model and the target motion trajectory match, and the target position and Doppler information provided by the multiple measurement data in the gate are combined to determine the first probability of the matching state and the second probability of the non-matching state, and by comparing the first probability and the second probability, it is determined whether the target is maneuvering, and the target is tracked after the maneuvering. The above method solves the technical problem of low resource utilization of the target tracking algorithm provided by the related technology.

[0098] In an optional embodiment, the above-mentioned calculation of the first probability and the second probability corresponding to each data frame based on the measurement data set of multiple consecutive data frames, the first noise matrix and the second noise matrix includes: calculating the first probability and the second probability corresponding to the first data frame based on the measurement data set of the first data frame, the first noise matrix and the second noise matrix, including: calculating the likelihood probability pairs corresponding to each measurement data of the first data frame according to the measurement data set of the first data frame, the first noise matrix and the second noise matrix; calculating the first probability and the second probability corresponding to the first data frame using the initial prior probability pair and the likelihood probability pairs corresponding to each measurement data of the first data frame, wherein the initial prior probability pair is used to indicate the probability of matching and the probability of not matching the target motion model and the target motion trajectory when no measurement data is received.

[0099] It should be noted that when the maneuvering target judgment begins, the first thing to process is the measurement data set of the first data frame. In this data frame, the specific steps of calculating the first probability and the second probability are divided into two parts: 1. Calculate the likelihood probability pair. For each measurement data in the first data frame, the system will calculate the likelihood probability under the "matching state" and "mismatching state" based on the measurement model, the predicted state, the first noise matrix and the second noise matrix. That is, for each measurement data, the likelihood value is calculated under the conditions of matching and mismatching the target motion model and the motion trajectory.

[0100] 2. Calculate the first probability and the second probability using the initial prior probability and the likelihood probability. In the processing of the first data frame, the first probability and the second probability need to be calculated by combining the initial prior probability pair and the likelihood probability pair. The initial prior probability pair reflects the system's prior estimate of whether the target motion model matches or does not match the motion trajectory when no measurement data is received, usually based on the system's preset parameter settings, such as detection probability, gate probability, and expectations of the target's maneuvering behavior.

[0101] It should be noted that the initial prior probability pair is a binary pair, which represents the matching probability and mismatch probability of the target motion model and motion trajectory in the initial state. These probabilities are set at the beginning of the algorithm and are continuously updated in the subsequent data frame processing to reflect the impact of the latest measurement data on the target state estimation.

[0102] It should be noted that once the probability pair of the first data frame is calculated, this pair of probabilities will be used as the prior probability pair of the next data frame, and the measurement data of the subsequent data frames will continue to be processed, and the above process will be repeated. In this way, as each data frame is processed, the system's judgment of the target's maneuvering state will gradually become more accurate. Through the analysis of multiple consecutive frames of data, a more robust judgment can be made on the matching degree between the target's motion trajectory and the motion model.

[0103] In an optional embodiment, the above-mentioned calculation of the first probability and the second probability corresponding to the kth data frame based on the measurement data set of the kth data frame, the first noise matrix and the second noise matrix, wherein k is an integer greater than 1, includes: calculating the prior probability pair corresponding to the kth data frame according to the first probability and the second probability corresponding to the k-1th data frame, wherein the prior probability pair is used to indicate the probability of matching and the probability of not matching the target motion model and the target motion trajectory when the measurement data of the kth data frame is not received; calculating the likelihood probability pairs corresponding to each measurement data of the kth data frame according to the measurement data set of the kth data frame, the first noise matrix and the second noise matrix; calculating the first probability and the second probability corresponding to the kth data frame using the prior probability pair corresponding to the kth data frame and the likelihood probability pairs corresponding to each measurement data of the kth data frame.

[0104] As mentioned above, the first probability and the second probability can be determined by a priori probability pair and a likelihood probability pair. Therefore, in the kth data frame, the priori probability pair and the likelihood probability pair need to be calculated first, and then the first probability and the second probability are determined.

[0105] It should be noted that before processing the kth data frame, the system will infer the prior probability pair of the kth data frame based on the first probability and the second probability calculated from the k-1th data frame. The prior probability pair contains the system's estimate of the probability of the target motion model matching and the probability of mismatching the actual motion trajectory before the measurement data of the kth data frame is received. After determining the prior probability pair, the system will calculate the likelihood probability pair based on the measurement data set of the kth data frame, the first noise matrix and the second noise matrix. The likelihood probability pair respectively represents the probability of the target motion model matching and mismatching the actual target motion trajectory under given measurement data. The first noise matrix and the second noise matrix represent the process noise of the target in the non-maneuvering state and the maneuvering state, respectively, and are used to adjust the uncertainty in the state transfer process.

[0106] In an optional embodiment, after obtaining the prior probability pair and likelihood probability pair of the kth data frame, the system will use the Bayesian formula to calculate the posterior probability pair of the frame. This pair of probabilities is the posterior probability (first probability) that the target motion model matches the actual target motion trajectory under the kth data frame and the posterior probability (second probability) that it does not match.

[0107] Through the above implementation of the present application, each time a new data frame is processed, the posterior probability (first probability and second probability) of the previous frame is used to determine the prior probability of the current frame, and the probability estimate is updated in combination with the measurement data and noise matrix of the current frame. This method makes full use of the continuity of time series data, allowing the system to gradually learn and adapt to changes in the target's motion state, so that in scenarios such as autonomous driving, it can more effectively identify and respond to maneuvering targets and improve the tracking performance and safety of the radar system. It enables the on-board radar system to adjust its judgment of the target's maneuvering state in real time, ensuring that even in complex and dynamic environments, it can maintain accurate tracking of the target, providing key data support for the safe operation of autonomous driving vehicles.

[0108] In an optional embodiment, the above-mentioned calculation of the first probability and the second probability corresponding to the k-th data frame using the prior probability pair corresponding to the k-th data frame and the likelihood probability pair corresponding to each measurement data of the k-th data frame includes: obtaining a first parameter using the first prior probability in the prior probability pair corresponding to the initial radar parameter set and the k-th data frame, and obtaining a second parameter using the second prior probability in the prior probability pair corresponding to the initial radar parameter set and the k-th data frame, wherein the first parameter and the first prior probability correspond to a situation where the target motion model and the target motion trajectory match, and the second parameter and the second prior probability correspond to a situation where the target motion model and the target motion trajectory do not match; The sum of the first likelihood probabilities is multiplied by the first prior probability corresponding to the kth data frame to obtain a third parameter, and the sum of the second likelihood probabilities in the likelihood probability pairs corresponding to the measurement data of the kth data frame is multiplied by the second prior probability corresponding to the kth data frame to obtain a fourth parameter, wherein the third parameter and the first likelihood probability correspond to the situation where the target motion model and the target motion trajectory match, and the fourth parameter and the second likelihood probability correspond to the situation where the target motion model and the target motion trajectory do not match; the ratio of the sum of the first parameter and the third parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter is determined as the first probability; the ratio of the sum of the second parameter and the fourth parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter is determined as the second probability.

[0109] That is, the first probability and the second probability can be calculated by the following formula:

[0110]

[0111] in

[0112]

[0113]

[0114] in, is the first probability of the kth data frame, is the second probability of the kth data frame, is one of the prior probability pairs corresponding to the kth data frame, is one of the likelihood probability pairs corresponding to the i-th measurement data of the k-th data frame. 0,1 is the first parameter, b 0,2 is the second parameter, b 1 is the third parameter, b 2 Is the fourth parameter.

[0115] It should be noted that the system needs to use the initialized radar parameter set and the prior probability calculated from the previous frame (k-1 frame) to generate two preliminary parameters: the first parameter and the second parameter. These parameters are directly related to the matching or mismatching state of the target motion model and motion trajectory.

[0116] It should be noted that the first parameter may indicate that when the target motion model matches the target motion trajectory (i.e., the target has not maneuvered), the system uses the initial radar parameter set and the first prior probability to calculate the first parameter. This parameter reflects the radar system's preliminary estimate of the target motion state under a given matching prior probability. In contrast, the second parameter may indicate that when the target motion model does not match the target motion trajectory (i.e., the target may have maneuvered), the system uses the initial radar parameter set and the second prior probability to calculate the second parameter. The second parameter reflects the radar system's preliminary estimate of the target motion state under the mismatching state.

[0117] In an optional embodiment, the system will then further calculate two parameters: a third parameter and a fourth parameter based on the likelihood probability pairs of all measurement data in the current frame (the kth data frame), which are used for the subsequent calculation of the first probability and the second probability. The third parameter can indicate the sum of the matching degrees of all measurement data when the target motion model matches the target motion trajectory. Specifically, the system adds the first likelihood probability of all measurement data in the current frame (that is, the probability that the measurement data matches the predicted state in the matching state), and then multiplies it by the first prior probability to obtain the third parameter. The fourth parameter is similar to the third parameter, but focuses on the mismatch situation, that is, the second likelihood probability of all measurement data in the current frame (the probability that the measurement data does not match the predicted state in the mismatch state) is added, and then multiplied by the second prior probability to obtain the fourth parameter.

[0118] In an optional embodiment, the system uses the first parameter, the second parameter, the third parameter and the fourth parameter calculated above to calculate the matching probability (first probability) and the mismatching probability (second probability) of the target under the kth data frame. The first probability can indicate the possibility of matching the target motion model with the target motion trajectory after taking into account all the measurement data and prior information. Among them, the first parameter represents the preliminary estimate of the model matching when the measurement data is not received, and the third parameter reflects the total contribution of the measurement data to the model matching. The sum of the two and the ratio of the sum of all parameters is the final probability in the case of matching. The second probability can indicate the possibility of mismatching the target motion model with the target motion trajectory. The second parameter represents the preliminary estimate in the case of mismatching, and the fourth parameter reflects the total contribution of the measurement data to the mismatching state. The ratio of the sum of these two parameters to the sum of all parameters is the final probability in the case of mismatching.

[0119] Through the above-mentioned implementation mode of the present application, the radar system can integrate the previous judgment (ie, the prior probability) and the contribution of the current measurement data (ie, the likelihood probability) to update the estimate of the degree of match between the target motion model and the actual motion trajectory in real time, and then accurately determine whether the target has maneuvered. Not only is the radar system's initial estimate (through the first parameter and the second parameter) taken into account when the target is not maneuvering, but the contribution of the real-time measurement data (through the third parameter and the fourth parameter) is also combined, and finally the maneuvering state of the target is determined by probability ratio calculation. This method improves the response speed and judgment accuracy of the radar system to maneuvering targets, and helps to achieve safer and smarter vehicle driving assistance and automatic driving functions.

[0120] In an optional embodiment, the above-mentioned calculation of the likelihood probability pairs corresponding to each measurement data of the kth data frame based on the measurement data set of the kth data frame, the first noise matrix and the second noise matrix includes: inputting the predicted state of the maneuvering target in the kth data frame into the measurement model of the radar to obtain a measurement mean, wherein the predicted state of the maneuvering target is determined according to the target motion model; obtaining the first covariance matrix and the second covariance matrix using the first noise matrix, the second noise matrix and the predicted state of the maneuvering target corresponding to the kth data frame; obtaining a first Gaussian distribution function based on the measurement mean and the first covariance matrix, and obtaining a second Gaussian distribution function based on the measurement mean and the second covariance matrix; and inputting the measurement data of the kth data frame into the first Gaussian distribution function and the second Gaussian distribution function, respectively, to obtain a likelihood probability pair.

[0121] It should be noted that the system can predict the position, speed and other state information of the target in the kth data frame based on the estimated state of the previous frame and the target motion model (such as constant speed, constant acceleration and other models). This predicted state reflects the expected position and motion characteristics of the target under non-maneuvering conditions. The above predicted state is input into the measurement model of the vehicle-mounted radar, which is usually a mathematical function or model that converts the state space (such as position, speed) to the observation space (such as the point cloud position received by the radar). The output of the measurement model is the measurement mean, which represents the expected value of the target measurement under the predicted state. Then, the first noise matrix (corresponding to the process noise in the non-maneuvering state) and the second noise matrix (corresponding to the process noise in the maneuvering state) are combined with the predicted state of the target to obtain the first covariance matrix and the second covariance matrix. The first covariance matrix reflects the predicted measurement uncertainty in the non-maneuvering state, while the second covariance matrix reflects the predicted measurement uncertainty in the maneuvering state.

[0122] In an optional embodiment, the first Gaussian distribution function and the second Gaussian distribution function can be constructed based on the measurement mean and the above two covariance matrices. These two distribution functions describe the probability distribution of the measurement data given the measurement mean and covariance. The first Gaussian distribution function corresponds to the distribution of measurement data in a non-maneuvering state, while the second Gaussian distribution function corresponds to the distribution of measurement data in a maneuvering state. Then, each measurement data in the kth data frame is substituted into the first Gaussian distribution function and the second Gaussian distribution function, respectively, to obtain two probability values: the likelihood probability of the measurement data in a non-maneuvering state and the likelihood probability of the measurement data in a maneuvering state. This likelihood probability pair reflects the degree of match between the observed data and the predicted state, and is an important input for updating the posterior probability in Bayesian reasoning.

[0123] Specifically, the likelihood probability can be calculated by the following formula:

[0124]

[0125] Among them, h(·) is the measurement model of the radar, X k|k-1 represents the predicted state of the filter at time k, is the covariance matrix of the predicted measurement, calculated as follows:

[0126]

[0127] Among them, H k is the Jacobian matrix at time k, H k T Denotes the matrix H k The transpose of F is the state transfer matrix of the filter, P k-1|k-1 is the state covariance matrix of the filter at time k-1, R k，iTo measure z k，i The covariance matrix of , d∈[0,1] is the influencing factor, which adjusts the influence of process noise in the filter on maneuver recognition.

[0128] Figure 4 is a schematic diagram of another optional mobile target tracking method according to an embodiment of the present application; Figure 4 As shown in Figure 1, the same measurement data has different likelihood values ​​under different covariances. The difference between small variance and large variance is more obvious.

[0129] It should be noted that the Jacobian matrix is ​​the measurement model h(·) with respect to the state X k|k-1 The derivative of . It can be obtained by mathematical differentiation. Specifically, if the measurement model h is a function of the state vector X, then the Jacobian matrix H k Each element of is the partial derivative of h with respect to each element of X. In practice, this usually involves computing the derivatives of the measurement model analytically, or approximating the derivatives numerically using finite difference methods.

[0130] The state transfer matrix F describes the evolution of the system state over time. It is usually determined based on the physical dynamic model of the system. In practical applications, F can be obtained through system identification, physical modeling, or experimental data fitting.

[0131] State covariance matrix P k-1|k-1 describes the uncertainty of the state estimate at time k-1. It can be updated and predicted through the Bayesian filtering process. Specifically, at each time step, the filter updates the covariance matrix of the state estimate based on the new observation data. In the prediction step, P k|k-1 It is based on P k-1|k-1 and process noise Q. In the update step, P k|k It is based on P k|k-1 and observation noise R to update.

[0132] Measurement covariance matrix R k，i Describes the statistical characteristics of the measurement noise. It is usually based on the sensor specifications or determined through experimental data. For example, if the noise characteristics of the sensor are known, then these characteristics can be used directly to set R k，i In practical applications, R k，i It can be obtained through calibration of the measuring equipment, fitting of experimental data or from the technical specifications provided by the manufacturer.

[0133] The impact factor d is a tuning parameter used to adjust the effect of process noise in the filter on maneuver identification. It is usually set based on system characteristics and experimental data. In practical applications, d can be determined through system identification, experimental adjustment, or expert experience. It allows the filter to have better adaptability when facing different system dynamics.

[0134] Through the above implementation of the present application, the system can comprehensively consider the information of all measurement data, further combine the prior probability, and use the Bayesian rule to update the probability estimate of the current state of the target, so as to determine whether the target has maneuvered. Overall, the above implementation uses a mathematical model to quantify the degree of matching between the measurement data and the predicted state into a likelihood probability pair, which provides a solid data foundation for subsequent probability updates and maneuvering state judgments, ensuring that the on-board radar system can accurately and timely identify target maneuvers in complex environments, and provide key information support for the safe operation of autonomous driving vehicles.

[0135] In an optional embodiment, the above-mentioned calculation of the prior probability pair corresponding to the k-th data frame based on the first probability and the second probability corresponding to the k-1-th data frame includes: multiplying the first probability and the second probability corresponding to the k-1-th data frame with the transition probability matrix to obtain the prior probability pair corresponding to the k-th data frame, wherein the transition probability matrix is ​​used to indicate the probability of a change in the matching relationship between the target motion model and the target motion trajectory in the k-1-th data frame and the k-th data frame.

[0136] It should be noted that the transition probability matrix can indicate the probability of a change in the matching state of the target motion model and the target motion trajectory between the two frames of data. It is usually a two-dimensional matrix, in which each row represents the matching state of the previous frame (for example, non-maneuvering state or maneuvering state), and each column represents the matching state of the current frame. The elements of the matrix represent the probability of transitioning from a state of the previous frame to another state of the current frame. For example, it can be a 2-row and 2-column transition probability matrix π=[π ij ],in represents the transition probability from time k-1 to time k.

[0137] In an optional implementation, as mentioned above, the target does not often perform maneuvering movements. To meet the actual situation, the transition probability matrix can satisfy:

[0138]

[0139] In an optional embodiment, before the arrival of the kth data frame, the system calculates the prior probability pair of the kth data frame based on the first probability (the probability that the target motion model and the target motion trajectory match) and the second probability (the probability that the target motion model and the target motion trajectory do not match) corresponding to the k-1th data frame, and the above-mentioned transition probability matrix.

[0140] The first probability and the second probability can be combined into a probability vector, and then the probability vector can be updated using the transition probability matrix to obtain the prior probability vector of the kth data frame. This update process can be achieved by matrix multiplication, and the specific formula is as follows:

[0141]

[0142] This calculation process can combine historical data and the statistical laws of target motion state changes to provide a more accurate prior estimate for determining the maneuvering state of the target in the next frame. By continuously iterating this process, the system can gradually learn and adapt to the pattern of target maneuvering behavior, thereby improving the recognition accuracy of maneuvering targets.

[0143] Through the above implementation of the present application, the method of updating the prior probability based on the historical probability and the transition probability matrix makes full use of the continuity and dynamics of the time series data, ensuring that the system can adjust the estimation of the target motion state in real time. In the real-time target tracking and maneuvering identification of the vehicle-mounted radar system, it can help the system to more accurately predict the possible state of the target at the next moment, and then optimize the parameters of the tracking algorithm, and improve the robustness and accuracy of the radar system in complex environments. An effective mechanism for dynamically updating the estimation of the target maneuvering state is provided for the vehicle-mounted radar system, which is crucial for realizing the target tracking function in advanced driver assistance systems (ADAS) and autonomous driving vehicles.

[0144] In an optional embodiment, before calculating the first probability and the second probability corresponding to each data frame based on the measurement data set of multiple consecutive data frames, the first noise matrix and the second noise matrix, the method includes: setting initial values ​​of the radar's clutter density, detection probability, and gate probability, and determining the initial values ​​of the radar's clutter density, detection probability, and gate probability as an initial radar parameter set; setting the value of a transfer matrix, wherein the transfer matrix includes the probability that the target motion model and the target motion trajectory of two consecutive data frames are both in a matching state, the probability that the target motion model and the target motion trajectory of two consecutive data frames change from a matching state to a non-matching state, the probability that the target motion model and the target motion trajectory of two consecutive data frames change from a non-matching state to a matching state, and the probability that the target motion model and the target motion trajectory of two consecutive data frames are both in a non-matching state.

[0145] It should be noted that before the radar system starts working, a set of initial parameters need to be set. These parameters reflect the basic characteristics of the radar working environment and the system's prior estimates of target detection and judgment. Among them, the clutter density λ is a measure of the strength or density of non-target source signals (i.e., clutter) in the radar received signal. It provides background information for determining whether the measured data is a real target. Usually, the estimation of clutter density needs to consider factors such as the radar operating frequency and environmental conditions (such as weather and terrain). Detection probability P D It refers to the probability that the radar system correctly detects the target when the target does exist. It reflects the reliability of the radar in identifying the target and is closely related to the performance parameters of the radar (such as sensitivity and resolution). G The probability that a radar measurement falls within a certain threshold of the target's predicted position. This probability is used to quantify the degree of agreement between the measurement data and the target's predicted position, and is critical to correctly associating the measurement data with the target.

[0146] In an optional embodiment, the initial prior probability can also be determined and and its corresponding first noise matrix Q 1 and the second noise matrix Q 2 In the actual process, the target does not maneuver at the initial moment, and the initial prior probability can satisfy:

[0147]

[0148] Through the above implementation of the present application, the system can construct a probability model to describe the evolution of the target state over time. It provides a starting point for subsequent dynamic probability updates, allowing the system to gradually adjust and optimize its judgment of the target's maneuvering state based on prior knowledge of the environment and the target, thereby improving the performance and reliability of the vehicle-mounted radar system in complex environments. By reasonably setting the initial parameters and the transition probability matrix, the system can more accurately identify the changes in the target's motion state and provide key safety information for autonomous driving vehicles.

[0149] In an optional embodiment, the above-mentioned determination that the target motion trajectory does not match the target motion model when the first probability and the second probability of multiple consecutive frames meet the preset conditions includes: when the first probability and the second probability in a data frame meet the preset conditions, adding one to the counting parameter; when the first probability and the second probability in a data frame do not meet the preset conditions, clearing the counting parameter; when the counting parameter is greater than the counting threshold, determining that the target motion trajectory does not match the target motion model.

[0150] It should be noted that the system needs to set a series of logical and numerical thresholds for subsequent judgment. These thresholds include preset conditions, counting parameters, and counting thresholds. The counting parameter is used to track the count of the target's possible maneuvering state in multiple consecutive frames of data. This parameter is incremented by one each time a possible maneuver is detected, and is cleared when the detection result does not support the maneuver. It is a pre-set value used to determine whether the target has actually maneuvered. When the accumulated value of the counting parameter exceeds this threshold, the system will confirm that the target's motion trajectory does not match the motion model, indicating that the target has maneuvered.

[0151] In an optional embodiment, for each frame of radar measurement data received, the system will compare the ratio of the calculated first probability and the second probability to see whether they meet the preset conditions. If the ratio of the first probability and the second probability of the current frame is lower than the preset threshold, it indicates that the matching probability in the frame data is lower than the mismatching probability, which may indicate that the target has maneuvered. Based on the analysis results of the above single frame data, the system updates the counting parameter. If the data of the current frame supports signs of maneuvering, that is, the above ratio meets the preset conditions, the counting parameter will be increased by one, indicating that the number of consecutive detections of signs of maneuvering has increased. On the contrary, if the data of the current frame does not support signs of maneuvering, the counting parameter will be cleared to restart the detection series. The accumulation of the counting parameter is the key to the system confirming the maneuvering state of the target. When the counting parameter is greater than the preset counting threshold, it indicates that the system has continuously detected possible signs of maneuvering in multiple consecutive frames of data. After reaching a certain number of times, it is sufficient to confirm that the target motion trajectory does not match the predicted motion model, that is, the target has maneuvered.

[0152] Through the above implementation of the present application, through logical judgment and counting accumulation, the robustness and accuracy of the system in judging the maneuvering state of the target are ensured, and the situation of misjudging the maneuvering of the target due to accidental interference or noise influence of single frame data is avoided. Through the comprehensive analysis of continuous multi-frame data, the vehicle-mounted radar system can more reliably identify the maneuvering behavior of the target, thereby providing more accurate target tracking and decision support for ADAS and autonomous driving vehicles.

[0153] In an optional implementation, the above-mentioned case where the first probability and the second probability in a data frame meet a preset condition includes: when a ratio of the first probability to the second probability is less than a preset ratio, determining that the preset condition is met.

[0154] It should be noted that the preset ratio is a threshold used to quantify and compare the relationship between the first probability (the probability that the target motion trajectory matches the motion model) and the second probability (the probability that the target motion trajectory does not match the motion model). This ratio is preset according to the requirements of system design and the definition of maneuvering state. It usually takes a lot of experiments and data analysis to determine a reasonable value to balance the sensitivity and accuracy of maneuvering judgment.

[0155] In an optional embodiment, when processing each frame of radar data, the system will calculate the posterior probability of the target matching or not matching in the current frame based on the Bayesian reasoning method, that is, the first probability and the second probability. Then, the system compares the ratio of the two probabilities with the preset ratio. If the ratio of the first probability to the second probability is less than the preset ratio, this means that the probability that the target does not match (i.e., may maneuver) in the current data frame is higher than the probability in the matching state. This result is interpreted as a sign that the target in the current frame data may have maneuvered. On the contrary, if the ratio is greater than or equal to the preset ratio, the system believes that the target in the current frame is in a non-maneuvering state, or at least there is no obvious sign of maneuvering.

[0156] Through the above implementation of the present application, a quantitative standard is provided for the vehicle-mounted radar system through the setting and comparison of probability ratios, which is used to instantly identify the maneuvering state of the target. The selection of the preset ratio is crucial. If the preset ratio is set too low, the system may be too sensitive and misjudge some normal non-maneuvering states as maneuvers. On the contrary, if the preset ratio is set too high, the system may become sluggish and it is difficult to capture the maneuvering behavior of the target in time. The sensitivity and accuracy of maneuvering detection are improved by reasonably setting the preset ratio.

[0157] Figure 5 is a schematic diagram of another optional mobile target tracking method according to an embodiment of the present application; Figure 5 As shown, for target tracking, the following steps can be performed in sequence:

[0158] Step S502: Initialize parameters. The parameters that need to be set are clutter density λ, detection probability P D , gate probability P G , the initial event θ 0 The prior probability and and its corresponding process noise covariance matrix Q 1 and Q 2 , 2 rows and 2 columns of the transition probability matrix π = [π ij ],in represents the transition probability from time k-1 to time k.

[0159] Step S504, calculate the prior probability pair at time k. According to the posterior probability at k-1 time and the transition probability π ij To calculate, the formula is as follows:

[0160]

[0161] Step S506, calculate the likelihood probability pair at time k. Under the Gaussian assumption, measure z k,i In the event θ k =m, the likelihood probability calculation formula is:

[0162]

[0163] Among them, h(·) is the measurement model of the radar, X k|k-1 represents the predicted state of the filter at time k, is the covariance matrix of the predicted measurement, calculated as follows:

[0164]

[0165] Among them, H k is the Jacobian matrix at time k, H k T Denotes the matrix H k The transpose of F is the state transfer matrix of the filter, P k-1|k-1 is the state covariance matrix of the filter at time k-1, R k,i To measure z k,i The covariance matrix of , d∈[0,1] is the influencing factor, which adjusts the influence of process noise in the filter on maneuver recognition.

[0166] Step S508, calculating the first probability and the second probability at time k. The first probability and the second probability are calculated by the following formula.

[0167]

[0168] in

[0169]

[0170]

[0171] Step S510: determine whether the first probability and the second probability meet the preset conditions, and calculate the posterior probability ratio

[0172] If the condition is not met, step S512-1 is executed and the counting parameter is cleared. k >G r, the counting parameter is cleared.

[0173] If the conditions are met, step S512-2 is executed and the counting parameter is increased by one. k <G r , the counting parameter is increased by one.

[0174] Step S514, determining whether the counting parameter is greater than a counting threshold. If it is not greater than the counting threshold, executing step S516-1 to determine whether the target motion trajectory matches the target motion model.

[0175] When the count value is greater than the count threshold, step S516 - 2 is executed to determine that the target motion trajectory does not match the target motion model.

[0176] Figure 6 is a schematic diagram of another optional mobile target tracking method according to an embodiment of the present application; Figure 6 As shown, the horizontal axis is time, and the vertical axis is the probability ratio of the first probability to the second probability. Figure 6 This is the result of a simulation. The simulation parameters are sampling time of 0.1s and the target turning maneuver every 30s.

[0177] Figure 7 is a schematic diagram of another optional mobile target tracking method according to an embodiment of the present application; Figure 7 As shown, the horizontal axis is time, and the vertical axis is the probability ratio of the first probability to the second probability. Figure 7 The results are shown in Figure 100, with a sampling time of 0.1s. The target makes a turn at 30s, and the probability ratio is completely lower than 1 after 31s, which lasts for 3s.

[0178] In an optional implementation, multiple noise matrices can be set to determine whether the target has maneuvered, or different measurement noises can be set to determine whether the target has maneuvered. Taking the setting of multiple noise matrices as an example, as mentioned above, only the first noise matrix is ​​set to indicate the matching state, and the second noise matrix indicates the mismatching state, and the state between the matching and mismatching can also be refined. For example, the first noise matrix can be used to indicate a high degree of matching, the second noise matrix can be used to indicate matching, the third noise matrix can be used to indicate mismatching, and the fourth noise matrix can be used to indicate a large deviation. The specific implementation method is similar to the aforementioned process, and the probability corresponding to each noise matrix is ​​calculated, and the state of the target is determined based on the obtained probability.

[0179] In the embodiment of the present application, different noise matrices are used as process noise covariance to characterize whether the target motion model and the target motion trajectory match each other, and the target position and Doppler information provided by multiple measurement data in the gate are combined to determine the first probability of the matching state and the second probability of the mismatching state. By comparing the first probability and the second probability, it is determined whether the target has maneuvered, and the target is tracked after the maneuver. The above method solves the technical problem of low resource utilization of the target tracking algorithm provided by the related art.

[0180] The embodiment of the present application does not use multiple motion models, but shares a motion model with the tracking algorithm. Therefore, there is no need to calculate and save the states of multiple motion models and other related information. Its storage requirements and computational complexity are approximately 1 / N of the multi-model algorithm (N represents the number of motion models of the multi-model algorithm), so it has lower storage requirements and computational complexity. Secondly, all point cloud data in the association gate are comprehensively considered, the likelihood probabilities provided by all point clouds are used, and the posterior probabilities are calculated, the proportion of the likelihood probabilities provided by a single noise point is limited, and the maneuvering state of the target is confirmed in multiple frames, which can effectively reduce the interference of a single noise.

[0181] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0182] According to another aspect of the embodiments of the present application, a mobile target tracking device for implementing the mobile target tracking method is also provided. Figure 8 As shown, the device comprises:

[0183] A noise matrix determination unit 802 is used to determine a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match;

[0184] A probability calculation unit 804 is used to calculate a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, a first noise matrix and a second noise matrix, wherein the first probability is used to indicate a probability that the target motion model of the current data frame matches the target motion trajectory, and the second probability is used to indicate a probability that the target motion model of the current data frame does not match the target motion trajectory;

[0185] The maneuver judgment unit 806 is used to determine that the target motion trajectory does not match the target motion model when the first probability and the second probability of multiple consecutive frames meet the preset conditions, and update the parameters of the target motion model.

[0186] Optionally, the probability calculation unit 804 includes: a first probability calculation module, which is used to calculate the first probability and the second probability corresponding to the first data frame based on the measurement data set of the first data frame, the first noise matrix and the second noise matrix, including: calculating the likelihood probability pairs corresponding to each measurement data of the first data frame according to the measurement data set of the first data frame, the first noise matrix and the second noise matrix; using the initial prior probability pair and the likelihood probability pair corresponding to each measurement data of the first data frame to calculate the first probability and the second probability corresponding to the first data frame, wherein the initial prior probability pair is used to indicate the probability of matching and the probability of not matching the target motion model and the target motion trajectory when the measurement data is not received; a second probability calculation module, which is used to calculate the first probability and the second probability corresponding to the first data frame based on the measurement data set of the kth data frame. The method comprises: calculating a first probability and a second probability corresponding to the k-th data frame according to the first probability and the second probability corresponding to the k-1-th data frame, wherein the prior probability pair is used to indicate the probability of matching and the probability of mismatching the target motion model and the target motion trajectory when the measurement data of the k-th data frame is not received; calculating the likelihood probability pairs corresponding to each measurement data of the k-th data frame according to the measurement data set of the k-th data frame, the first noise matrix and the second noise matrix; and calculating the first probability and the second probability corresponding to the k-th data frame using the prior probability pairs corresponding to the k-th data frame and the likelihood probability pairs corresponding to each measurement data of the k-th data frame.

[0187] Optionally, the above-mentioned probability calculation unit 804 includes: a first parameter calculation module, which is used to obtain a first parameter using a first prior probability in a priori probability pair corresponding to an initial radar parameter set and a k-th data frame, and to obtain a second parameter using a second prior probability in a priori probability pair corresponding to an initial radar parameter set and a k-th data frame, wherein the first parameter and the first prior probability correspond to a situation where the target motion model and the target motion trajectory match, and the second parameter and the second prior probability correspond to a situation where the target motion model and the target motion trajectory do not match; a second parameter calculation module, which is used to sum the first likelihood probability in the likelihood probability pair corresponding to each measurement data of the k-th data frame with the first prior probability corresponding to the k-th data frame. The third parameter is obtained by multiplying the first and second prior probabilities by the third parameter, and the fourth parameter is obtained by multiplying the sum of the second likelihood probabilities in the likelihood probability pairs corresponding to the measurement data of the k-th data frame by the second prior probability corresponding to the k-th data frame, wherein the third parameter and the first likelihood probability correspond to the situation where the target motion model and the target motion trajectory match, and the fourth parameter and the second likelihood probability correspond to the situation where the target motion model and the target motion trajectory do not match; a probability determination module is used to determine the first probability as the ratio of the sum of the first parameter and the third parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter; and determine the second probability as the ratio of the sum of the second parameter and the fourth parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter.

[0188] Optionally, the above-mentioned probability calculation unit 804 includes: a mean calculation module, which is used to input the predicted state of the maneuvering target in the kth data frame into the measurement model of the radar to obtain a measurement mean, wherein the predicted state of the maneuvering target is determined according to the target motion model; a covariance matrix calculation module, which is used to obtain a first covariance matrix and a second covariance matrix using a first noise matrix, a second noise matrix and the predicted state of the maneuvering target corresponding to the kth data frame; a Gaussian distribution function calculation module, which is used to obtain a first Gaussian distribution function according to the measurement mean and the first covariance matrix, and to obtain a second Gaussian distribution function according to the measurement mean and the second covariance matrix; a likelihood probability calculation module, which is used to input the measurement data of the kth data frame into the first Gaussian distribution function and the second Gaussian distribution function respectively to obtain a likelihood probability pair.

[0189] Optionally, the above-mentioned probability calculation unit 804 includes: a prior probability calculation module, used to multiply the first probability and the second probability corresponding to the k-1th data frame with the transition probability matrix to obtain a prior probability pair corresponding to the kth data frame, wherein the transition probability matrix is ​​used to indicate the probability of a change in the matching relationship between the target motion model and the target motion trajectory in the k-1th data frame and the kth data frame.

[0190] Optionally, the above-mentioned mobile target tracking device also includes: an initialization unit, used to: set the initial values ​​of the radar's clutter density, detection probability, and gate probability, and determine the initial values ​​of the radar's clutter density, detection probability, and gate probability as an initial radar parameter set; set the value of a transfer matrix, wherein the transfer matrix includes the probability that the target motion model and the target motion trajectory of two consecutive data frames are both in a matching state, the probability that the target motion model and the target motion trajectory of two consecutive data frames change from a matching state to a non-matching state, the probability that the target motion model and the target motion trajectory of two consecutive data frames change from a non-matching state to a matching state, and the probability that the target motion model and the target motion trajectory of two consecutive data frames are both in a non-matching state.

[0191] Optionally, the above-mentioned maneuver judgment unit 806 also includes: a counting module, used to add one to the counting parameter when the first probability and the second probability in a data frame meet the preset conditions; clear the counting parameter when the first probability and the second probability in a data frame do not meet the preset conditions; and a judgment module, used to determine that the target motion trajectory does not match the target motion model when the counting parameter is greater than the counting threshold.

[0192] Optionally, the counting module is further used to determine that a preset condition is met when a ratio of the first probability to the second probability is less than a preset ratio.

[0193] Optionally, in this embodiment, the embodiments to be implemented by the above-mentioned various unit modules can refer to the above-mentioned various method embodiments, which will not be repeated here.

[0194] According to another aspect of the embodiment of the present application, an electronic device for implementing the above-mentioned mobile target tracking method is also provided. The electronic device may be Fig. 9 The terminal device or server shown in the figure. This embodiment is described by taking the electronic device as a terminal device as an example. Fig. 9 As shown, the electronic device includes a memory 902 and a processor 904. The memory 902 stores a computer program, and the processor 904 is configured to execute the steps in any of the above method embodiments through the computer program.

[0195] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0196] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0197] S1, determining a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match;

[0198] S2, calculating a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, a first noise matrix and a second noise matrix, wherein the first probability is used to indicate a probability that a target motion model of a current data frame matches a target motion trajectory, and the second probability is used to indicate a probability that a target motion model of a current data frame does not match a target motion trajectory;

[0199] S3, when the first probability and the second probability of the continuous multiple frames meet the preset conditions, determining that the target motion trajectory does not match the target motion model, and updating the parameters of the target motion model.

[0200] Alternatively, a person skilled in the art may understand that: Fig. 9 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (MID), a PAD, and other terminal devices. Fig. 9 The electronic device and the electronic equipment described above are not limited in structure. Fig. 9 More or fewer components (such as network interfaces, etc.) as shown in, or with Fig. 9 Different configurations are shown.

[0201] Among them, the memory 902 can be used to store software programs and modules, such as the program instructions / modules corresponding to the mobile target tracking method and device in the embodiments of the present application. The processor 904 executes various functional applications and data processing by running the software programs and modules stored in the memory 902, that is, to implement the above-mentioned mobile target tracking method. The memory 902 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 902 may further include a memory remotely located relative to the processor 904, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Among them, the memory 902 can be specifically, but not limited to, used to store information such as measurement data. As an example, such as Fig. 9As shown, the memory 902 may include, but is not limited to, the noise matrix determination unit 802, the probability calculation unit 804, and the maneuver judgment unit 806 in the maneuvering target tracking device. In addition, it may also include, but is not limited to, other module units in the maneuvering target tracking device, which will not be repeated in this example.

[0202] Optionally, the transmission device 906 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wired network and a wireless network. In one example, the transmission device 906 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices and routers via a network cable so as to communicate with the Internet or a local area network. In one example, the transmission device 906 is a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0203] In addition, the electronic device mentioned above further includes: a display 908 for displaying a simulated road scene; and a connection bus 910 for connecting various module components in the electronic device mentioned above.

[0204] In other embodiments, the terminal device or server may be a node in a distributed system, wherein the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting the multiple nodes through network communication. Among them, the nodes may form a peer-to-peer (P2P, Peer To Peer) network, and any form of computing device, such as a server, terminal or other electronic device, may become a node in the blockchain system by joining the peer-to-peer network.

[0205] According to one aspect of the present application, a computer program product is provided, the computer program product comprising a computer program / instruction, the computer program / instruction comprising a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit, various functions provided by the embodiments of the present application are executed.

[0206] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0207] According to one aspect of the present application, a computer-readable storage medium is provided, and a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above-mentioned maneuvering target tracking method.

[0208] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0209] S1, determining a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match;

[0210] S2, calculating a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, a first noise matrix and a second noise matrix, wherein the first probability is used to indicate a probability that a target motion model of a current data frame matches a target motion trajectory, and the second probability is used to indicate a probability that a target motion model of a current data frame does not match a target motion trajectory;

[0211] S3, when the first probability and the second probability of the continuous multiple frames meet the preset conditions, determining that the target motion trajectory does not match the target motion model, and updating the parameters of the target motion model.

[0212] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0213] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which can be personal computers, servers or network devices, etc.) to execute all or part of the steps of the above methods in each embodiment of the present application.

[0214] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0215] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0216] The units described above as separate components may or may not be physically separated, and the components shown as units 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 units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0217] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0218] The above are only preferred implementations of the present application. It should be pointed out that ordinary technicians in this technical field can make several improvements and modifications without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for tracking a mobile target, characterized in that: include: Determine a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; Calculate the first probability and the second probability corresponding to each data frame based on the measurement data set of a plurality of consecutive data frames, the first noise matrix and the second noise matrix, wherein the first probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate the probability that the target motion model and the target motion trajectory of the current data frame do not match; When the first probability and the second probability of a plurality of consecutive frames meet a preset condition, it is determined that the target motion trajectory does not match the target motion model, and the parameters of the target motion model are updated.

2. The method according to claim 1, characterized in that The calculating the first probability and the second probability corresponding to each data frame based on the measurement data set of a plurality of consecutive data frames, the first noise matrix and the second noise matrix comprises: Calculating a first probability and a second probability corresponding to the first data frame based on the measurement data set of the first data frame, the first noise matrix, and the second noise matrix, including: Calculating likelihood probability pairs corresponding to each measurement data of the first data frame according to the measurement data set of the first data frame, the first noise matrix and the second noise matrix; Calculating a first probability and a second probability corresponding to the first data frame using an initial prior probability pair and a likelihood probability pair corresponding to each measurement data of the first data frame, wherein the initial prior probability pair is used to indicate a probability of matching and a probability of not matching the target motion model and the target motion trajectory when no measurement data is received; Calculating a first probability and a second probability corresponding to the k-th data frame based on the measurement data set of the k-th data frame, the first noise matrix, and the second noise matrix, where k is an integer greater than 1, including: Calculating a priori probability pair corresponding to the k-th data frame according to the first probability and the second probability corresponding to the k-1-th data frame, wherein the priori probability pair is used to indicate the probability of the target motion model and the target motion trajectory matching and the probability of non-matching when the measurement data of the k-th data frame is not received; Calculating likelihood probability pairs corresponding to each measurement data of the k-th data frame according to the measurement data set of the k-th data frame, the first noise matrix, and the second noise matrix; The first probability and the second probability corresponding to the kth data frame are calculated using the prior probability pair corresponding to the kth data frame and the likelihood probability pair corresponding to each measurement data of the kth data frame.

3. The method according to claim 2, characterized in that The calculating the first probability and the second probability corresponding to the kth data frame by using the prior probability pair corresponding to the kth data frame and the likelihood probability pair corresponding to each measurement data of the kth data frame comprises: Obtaining a first parameter using a first prior probability in the prior probability pair corresponding to an initial radar parameter set and a k-th data frame, and obtaining a second parameter using a second prior probability in the prior probability pair corresponding to the initial radar parameter set and the k-th data frame, wherein the first parameter and the first prior probability correspond to a case where the target motion model and the target motion trajectory match, and the second parameter and the second prior probability correspond to a case where the target motion model and the target motion trajectory do not match; A third parameter is obtained by multiplying the sum of the first likelihood probabilities in the likelihood probability pairs corresponding to the measurement data of the k-th data frame by the first prior probability corresponding to the k-th data frame, and a fourth parameter is obtained by multiplying the sum of the second likelihood probabilities in the likelihood probability pairs corresponding to the measurement data of the k-th data frame by the second prior probability corresponding to the k-th data frame, wherein the third parameter and the first likelihood probability correspond to a situation where the target motion model and the target motion trajectory match, and the fourth parameter and the second likelihood probability correspond to a situation where the target motion model and the target motion trajectory do not match; determining a ratio of a sum of the first parameter and the third parameter to a sum of the first parameter, the second parameter, the third parameter and the fourth parameter as the first probability; The ratio of the sum of the second parameter and the fourth parameter to the sum of the first parameter, the second parameter, the third parameter and the fourth parameter is determined as the second probability.

4. The method according to claim 2, characterized in that: The calculating, according to the measurement data set of the k-th data frame, the first noise matrix and the second noise matrix, the likelihood probability pairs corresponding to each measurement data of the k-th data frame comprises: Inputting the predicted state of the maneuvering target in the kth data frame into a measurement model of the radar to obtain a measurement mean, wherein the predicted state of the maneuvering target is determined according to the target motion model; Obtaining a first covariance matrix and a second covariance matrix using the first noise matrix, the second noise matrix, and the predicted state of the maneuvering target corresponding to the k-th data frame; Obtaining a first Gaussian distribution function according to the measurement mean and the first covariance matrix, and obtaining a second Gaussian distribution function according to the measurement mean and the second covariance matrix; The measurement data of the k-th data frame are respectively input into the first Gaussian distribution function and the second Gaussian distribution function to obtain the likelihood probability pair.

5. The method according to claim 2, characterized in that: The calculating the prior probability pair corresponding to the k-th data frame according to the first probability and the second probability corresponding to the k-1th data frame includes: The first probability and the second probability corresponding to the k-1th data frame are multiplied by the transition probability matrix to obtain the prior probability pair corresponding to the k-th data frame, wherein the transition probability matrix is ​​used to indicate the probability of a change in the matching relationship between the target motion model and the target motion trajectory in the k-1th data frame and the kth data frame.

6. The method according to any one of claims 1 to 5, characterized in that Before calculating the first probability and the second probability corresponding to each data frame based on the measurement data set of a plurality of consecutive data frames, the first noise matrix and the second noise matrix, the method includes: Setting initial values ​​of clutter density, detection probability, and gate probability of the radar, and determining the initial values ​​of clutter density, detection probability, and gate probability of the radar as an initial radar parameter set; Set the value of the transfer matrix, wherein the transfer matrix includes the probability that the target motion model and the target motion trajectory are both in a matching state in two consecutive data frames, the probability that the target motion model and the target motion trajectory change from a matching state to a non-matching state in two consecutive data frames, the probability that the target motion model and the target motion trajectory change from a non-matching state to a matching state in two consecutive data frames, and the probability that the target motion model and the target motion trajectory are both in a non-matching state in two consecutive data frames.

7. The method according to claim 1, characterized in that The determining that the target motion trajectory does not match the target motion model when the first probability and the second probability of the continuous multiple frames meet the preset condition includes: When the first probability and the second probability in a data frame meet the preset condition, adding one to the counting parameter; When the first probability and the second probability in a data frame do not meet the preset condition, clearing the counting parameter; When the counting parameter is greater than the counting threshold, it is determined that the target motion trajectory does not match the target motion model.

8. The method according to claim 7, characterized in that When the first probability and the second probability in a data frame meet the preset condition, the method comprises: When the ratio of the first probability to the second probability is less than a preset ratio, it is determined that the preset condition is met.

9. A mobile target tracking device, characterized in that: include: A noise matrix determination unit, configured to determine a first noise matrix and a second noise matrix, wherein the first noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory match, and the second noise matrix is ​​used to indicate the uncertainty of the target motion model when the target motion model and the target motion trajectory do not match; A probability calculation unit, configured to calculate a first probability and a second probability corresponding to each data frame based on a measurement data set of a plurality of consecutive data frames, the first noise matrix, and the second noise matrix, wherein the first probability is used to indicate a probability that the target motion model and the target motion trajectory of the current data frame match, and the second probability is used to indicate a probability that the target motion model and the target motion trajectory of the current data frame do not match; A maneuver judgment unit is used to determine that the target motion trajectory does not match the target motion model when the first probability and the second probability of multiple consecutive frames meet a preset condition, and to update the parameters of the target motion model.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method described in any one of claims 1 to 8 when executed.

11. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

12. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 8 through the computer program.

Images