Target tracking method and device for vehicle and vehicle driving system

Abstract

The invention relates to a target tracking method and device for a vehicle and a vehicle driving system, and the method comprises the steps: obtaining an existing track of a tracking target, and obtaining the current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors; based on the mahalanobis distance and a target confidence coefficient corresponding to the current measurement data, carrying out association matching on the current measurement data and an existing track to obtain a matching result; and when determining that the measurement data matched with the existing track exists in the current measurement data according to the matching result, performing Kalman updating on the existing track according to the current measurement data. According to the method, data collected by a sensor in real time can be associated and matched with an existing track, and the matched track is associated with front and back frames, so that stable tracking of a foresight target can be improved, and the stability of foresight perception of a vehicle is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent vehicle driving, and in particular to a target tracking method, device and vehicle driving system for a vehicle. Background Art

[0002] If smart cars want to use data obtained by sensors to achieve smart driving, the data obtained by the sensors must be processed by perception algorithms and compiled into data results to achieve information exchange between vehicles, roads, and people, so that the vehicle can automatically analyze whether the vehicle is in a safe or dangerous state, allowing the vehicle to achieve smart driving according to human wishes, and ultimately replace humans to make decisions and achieve unmanned driving goals.

[0003] At present, the forward-looking target is often tracked based on the data collected by a single sensor, and then the current driving status of the vehicle is analyzed. In this way, the collected data is often affected by the limitations of the single sensor itself, resulting in low stability of the perception results. For example, the data collected by the forward-looking camera is more easily affected by rainy weather.

[0004] With regard to the problem of low stability of forward-looking perception results of target tracking in related technologies, no effective solution has been proposed so far. Summary of the invention

[0005] In this embodiment, a target tracking method, device and vehicle driving system for a vehicle are provided to solve the problem of low stability of forward-looking perception results of target tracking in related technologies.

[0006] In a first aspect, a target tracking method for a vehicle is provided in this embodiment, comprising:

[0007] Acquire an existing track of the tracked target and acquire current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors;

[0008] Based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result;

[0009] When it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data, a Kalman update is performed on the existing track according to the current measurement data.

[0010] In some embodiments, obtaining current measurement data collected by at least one sensor includes:

[0011] Acquire the first measurement data currently collected by the forward-looking camera, and / or the second measurement data currently collected by the millimeter-wave radar; wherein, when the currently collected measurement data includes the first measurement data, convert the first measurement data from the camera coordinates to the preset absolute coordinates according to the preset coordinate conversion relationship.

[0012] In some embodiments, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result, including:

[0013] Determining a fusion order of the current measurement data according to timestamp information corresponding to the current measurement data;

[0014] According to the fusion order, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, each current measurement data is sequentially associated and matched with the existing track to obtain a matching result.

[0015] In some embodiments, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result, including:

[0016] According to the historical track, the Kalman prediction is performed on the current track to obtain the existing track;

[0017] Calculate the one-to-one matching degree between each track in the existing track and each measurement data in the current measurement data according to the Mahalanobis distance and the target confidence corresponding to the current measurement data;

[0018] A correlation matrix is ​​constructed with the number of existing tracks as rows and the number of measurements of current measurement data as columns, and the one-to-one matching degree between the existing tracks and the current measurement data is assigned to each element in the correlation matrix to obtain a matching result.

[0019] In some embodiments, when it is determined according to the matching result that there is measurement data in the current measurement data that matches the existing track, performing a Kalman update on the existing track according to the current measurement data includes:

[0020] When the target measurement data that best matches the existing track is determined from the matching results according to a preset matching algorithm, a Kalman update is performed on the existing track according to the target measurement data.

[0021] In some embodiments, after determining the target measurement data that best matches the existing track from the matching results, and before performing a Kalman update on the existing track according to the current measurement data, the method further includes:

[0022] Determine whether the position difference between the kinematic predicted position corresponding to the target measurement data and the kinematic predicted position corresponding to the existing track is less than a preset difference threshold; if so, perform Kalman update on the existing track according to the target measurement data; otherwise, discard the target measurement data.

[0023] In some embodiments, the method further comprises:

[0024] According to the matching result, a new track is created for the measurement data in the current measurement data that is not matched with the existing track.

[0025] In some embodiments, the method further comprises:

[0026] If there are multiple tracks created based on the measurement data provided by the same sensor, then the time information and position information of the multiple tracks are compared to obtain a track distance comparison result;

[0027] At least two tracks among the multiple tracks whose track distance comparison results are less than a preset distance threshold are merged into one track.

[0028] In a second aspect, in this embodiment, a target tracking device for a vehicle is provided, comprising: an acquisition module, a matching module and an update module; wherein:

[0029] The acquisition module is used to acquire an existing track of the tracked target and to acquire current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors;

[0030] The matching module is used to associate and match the current measurement data with the existing track based on the Mahalanobis distance and the target confidence corresponding to the current measurement data to obtain a matching result;

[0031] The updating module is used to perform Kalman update on the existing track according to the current measurement data when it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data.

[0032] In the third aspect, a vehicle driving system is provided in this embodiment, comprising: a forward-looking camera, a millimeter-wave radar and a server; the forward-looking camera and the millimeter-wave radar are respectively connected to the server; the forward-looking camera and the millimeter-wave radar are respectively used to collect measurement data; the server is used to execute the target tracking method for a vehicle described in the first aspect above.

[0033] Compared with the related art, the present embodiment provides a target tracking method, device and vehicle driving system for a vehicle. The target tracking method for a vehicle obtains an existing track of the tracked target and obtains current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors; based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result; when it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data, the existing track is Kalman updated according to the current measurement data. It can associate and match the data collected by the sensor in real time with the existing track, and associate the matched track with the front and back frames, thereby improving the stable tracking of the forward-looking target and improving the stability of the vehicle's forward-looking perception.

[0034] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] 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:

[0036] Figure 1 is a hardware structure block diagram of a terminal of the target tracking method for a vehicle according to the present embodiment;

[0037] Figure 2 is a flow chart of a target tracking method for a vehicle according to the present embodiment;

[0038] Figure 3 is a flow chart of a target tracking method of some of the embodiments;

[0039] Figure 4 is a structural block diagram of a target tracking device for a vehicle according to the present embodiment;

[0040] Figure 5 Schematic diagram of the structure of the vehicle driving system of this embodiment. DETAILED DESCRIPTION

[0041] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0042] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the", "these" and the like in this application do not represent quantitative restrictions, and they can be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether directly or indirectly. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. Usually, the character " / " indicates that the objects associated with each other are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0043] The method embodiment provided in this embodiment can be executed in a terminal, a computer or a similar computing device. For example, running on a terminal, Figure 1 is a hardware structure block diagram of a terminal of the target tracking method for a vehicle in this embodiment. Figure 1 As shown, the terminal may include one or more ( Figure 1 Only one is shown in the figure) processor 102 and memory 104 for storing data, wherein processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA. The above terminal may also include a transmission device 106 and an input and output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0044] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the target tracking method for vehicles in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 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 examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0045] The transmission device 106 is used to receive or send data via a network. The above network includes a wireless network provided by the communication provider of the terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.

[0046] In this embodiment, a target tracking method for a vehicle is provided. Figure 2 is a flow chart of the target tracking method for a vehicle in this embodiment. Figure 2 As shown, the process includes the following steps:

[0047] Step S210, obtaining an existing track of the tracked target and obtaining current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors.

[0048] The above-mentioned tracking target is a target within the preset detection range around the vehicle where the sensor is located (that is, the self-vehicle), such as other vehicles, pedestrians or other dynamic obstacles. The existing track of the tracking target is constructed by the historical measurement data previously detected by multiple sensors. The state quantity contained in the existing track can be: the relative distance Px in the x direction and the relative distance Py in the y direction between the self-vehicle and the tracking target; the absolute speed Velx in the x direction and the absolute speed Vely in the y direction of the tracking target; the absolute acceleration Accx in the x direction and the absolute acceleration Accy in the y direction of the tracking target; the speed Vx in the x direction and the speed Vy in the y direction of the tracking target relative to the self-vehicle; the angular velocity Ax in the x direction and the angular velocity Ay in the y direction of the tracking target relative to the self-vehicle. The global existing track can be constructed based on the historical measurement data collected by multiple sensors, such as forward-looking cameras, millimeter-wave radars, etc.

[0049] Among them, the existing track can be obtained based on Kalman prediction, which is specifically based on the kinematic formula and sensor delay △t:

[0050]

[0051] where x k It can represent the existing track currently predicted, x k is based on the track x at the previous moment k-1 The various state quantities of are predicted. It can be simplified as:

[0052] x k =Ax k-1 +w k-1 (2)

[0053] where w k-1 is the noise matrix, which can be set to 0 matrix here.

[0054] Step S220 , based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result.

[0055] The existing track can be associated with the currently collected measurement results. For each track of the existing track and each measurement data in the current measurement data, the Mahalanobis distance between the two is calculated as the distance cost, and multiplied by the target confidence transmitted by the sensor perception result, and then the corresponding matching result is obtained. Among them, the target confidence sent by the sensor represents the sensor's measure of whether the detected target actually exists and the accuracy of its attributes (such as position, speed, category, etc.). It is a value between 0 and 1, and the higher the value, the higher the accuracy. In this step, the current measurement data is associated with the existing track in combination with the Mahalanobis distance between the state quantity of the measurement data and the existing track, as well as the target confidence, which can achieve accurate track association of the previous and next frame data, thereby improving the accuracy and stability of target tracking.

[0056] Step S230: When it is determined according to the matching result that there is measurement data in the current measurement data that matches the existing track, Kalman update is performed on the existing track according to the current measurement data.

[0057] After the matching result is obtained in step S220, the existing track is updated accordingly. First, the state matrix P of the existing track is predicted to obtain the predicted covariance matrix

[0058]

[0059] Afterwards, the Kalman gain K is calculated based on the predicted covariance matrix k :

[0060]

[0061] Where H is the unit matrix, representing the quantity to be observed; R is the inherent noise of the sensor, which can be provided by the sensor manufacturer.

[0062] Afterwards, the posterior estimate is calculated based on the Kalman gain:

[0063]

[0064] in, is the final updated state, z km It is the measurement data that best matches the existing track in the current measurement data. By updating, the state quantities such as Px, Py, Velx, Vely, Accx, Vely, etc. in the existing track can be updated.

[0065] Finally, update the error covariance matrix:

[0066]

[0067] in, is the final updated covariance matrix.

[0068] Based on equations (1) to (6), the measurement data from different sensors can be continuously fused according to the above steps to form a stable forward tracking.

[0069] In the related art, the forward-looking target is tracked based on the data collected by a single sensor. The collected data is often affected by the limitations of the single sensor itself, resulting in low stability of the perception results. For example, the data collected by the forward-looking camera is more susceptible to rainy weather. The data collected by the millimeter-wave radar is not easy to identify the specific shape of the obstacle. In addition, the method of using laser radar for data collection is not suitable for all models due to its high cost. Therefore, the stability of forward-looking target tracking based on a single sensor in the related art is low.

[0070] The forward-looking perception scheme formed by the above steps S210 to S230 in this embodiment supports data fusion using different sensors. By collecting the current measurement data collected by at least one sensor, based on the combination of target confidence and Mahalanobis distance, the existing track previously constructed by fusion of multiple sensor data is correlated and matched, so as to achieve accurate and stable forward-looking perception. In addition, the state quantities used in this embodiment are perception attributes such as longitudinal distance, speed, acceleration, etc., which can be applied to intelligent driving of low-cost vehicles to eliminate the dependence on lidar.

[0071] The above steps S210 to S230 obtain the existing track of the tracked target and the current measurement data collected by at least one sensor; wherein the existing track is constructed based on the historical measurement data collected by at least two sensors; based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result; when it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data, the existing track is Kalman updated according to the current measurement data. It can associate and match the data collected by the sensor in real time with the existing track, and associate the previous and next frames of the matched track, thereby improving the stable tracking of the forward-looking target and the stability of the vehicle's forward-looking perception.

[0072] Furthermore, in one embodiment, obtaining current measurement data collected by at least one sensor may specifically include:

[0073] Acquire the first measurement data currently collected by the forward-looking camera, and / or the second measurement data currently collected by the millimeter-wave radar; wherein, when the currently collected measurement data includes the first measurement data, convert the first measurement data from the camera coordinates to the preset absolute coordinates according to the preset coordinate conversion relationship.

[0074] In this embodiment, measurement data can be collected by two different sensors. Specifically, the measurement data is collected by a forward-looking camera and a millimeter-wave radar. If any one of the forward-looking camera and the millimeter-wave radar collects measurement data, or both sensors collect measurement data, then the measurement data is associated and matched with the existing track.

[0075] Among them, since the measurement data output by the millimeter wave radar includes the absolute speed and absolute acceleration of the detected target, in order to unify the data, it is necessary to convert the coordinate system of the measurement data collected by the forward-looking camera, and convert the first measurement data from the camera coordinate system to the absolute coordinate system:

[0076] Velx=-wPy+Vx+speed (7)

[0077] Vely=-wPx+Vy (8)

[0078] Accx=-w 2 Px-2wVy+Ax+acc (9)

[0079] Accy=-w 2 Py-2wVx+Ay+speed×w (10)

[0080] Among them, speed is the speed of the vehicle, which is a scalar; w is the angular velocity of the vehicle; acc represents the longitudinal relative acceleration.

[0081] In this embodiment, the respective characteristics of the forward-looking camera and the millimeter-wave radar are taken into account in the forward-looking perception. Among them, the forward-looking camera can identify the attributes and shapes of obstacles, but has certain limitations in measuring the speed, acceleration, and longitudinal distance of obstacles, while the millimeter-wave radar has certain capabilities for measuring the longitudinal distance, speed, and acceleration of obstacles, but has certain limitations in measuring the target confidence, shape, etc. In this embodiment, by fusing the forward-looking camera with the millimeter-wave radar, the measurement data sensed by different sensors are converted to the same coordinate system, and then subsequent correlation matching is performed, so that the perception advantages of the millimeter-wave radar for longitudinal distance, speed, and acceleration can be combined with the perception advantages of the forward-looking camera for attributes and shapes, to achieve multi-modal target-level fusion, make up for the disadvantages of each sensor, and thus ensure stable forward-looking perception. In addition, the use of forward-looking cameras and millimeter-wave radars is also relatively low in cost.

[0082] Additionally, in one embodiment, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result, which may specifically include:

[0083] According to the timestamp information corresponding to the current measurement data, the fusion order of the current measurement data is determined; according to the fusion order, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, each current measurement data is sequentially associated and matched with the existing track to obtain the matching result.

[0084] Specifically, it is possible to determine whether each sensor has been updated based on the timestamp information corresponding to the current measurement data. This can be done by establishing a two-dimensional list List(flag1, flag2). Among them, flag1 indicates whether the front-view camera has been updated, and flag2 indicates whether the millimeter-wave radar has been updated; the timestamp threshold can be set to 0.05 seconds. Based on the timestamp information corresponding to the measurement data and the timestamp threshold, when the timestamp of the newly updated measurement data is 0.05 seconds apart from the timestamp corresponding to the fused measurement data, it is confirmed that the sensor has been updated. Among them:

[0085] When the sensor update value is (-1, -1), it means that neither the front-view camera nor the millimeter-wave radar has been updated, so there is no need for fusion. Among them, "-1" means that the sensor has not been updated, "0" means that the front-view camera has been updated, and "1" means that the millimeter-wave radar has been updated.

[0086] When the value of the sensor update status is (0, -1), it indicates that the front-view camera is updated, but the millimeter-wave radar is not updated. Therefore, only the measurement data of the front-view camera is fused, that is, List[1]=0 is updated in the update list.

[0087] When the value of the above sensor update status is (1, -1), it indicates that the front-view camera is not updated, the millimeter-wave radar is updated, and only the measurement data of the millimeter-wave radar is integrated, that is, List[0]=0 is updated in the update list.

[0088] The value of the sensor update status is (1, 0), indicating that both the forward-looking camera and the millimeter-wave radar are updated. The measurement data of the millimeter-wave radar is fused first, and then the measurement data of the forward-looking camera is fused. List[i] is updated to 1, and when the measurement data of each sensor is fused, they are fused in chronological order.

[0089] When neither the forward-looking camera nor the millimeter-wave radar is updated, since the time to trigger a fusion is 0.025 seconds, the delay of all tracks needs to increase by the corresponding time interval, for example, 0.025 seconds, so as to achieve time compensation for the track.

[0090] In this embodiment, the fusion order of the measurement data from different sensors is determined according to the timestamps, so that the orderly fusion of the measurement data from different sensors can be achieved, thereby improving the accuracy of subsequent perception fusion results.

[0091] In particular, in one embodiment, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result, which may specifically include:

[0092] According to the historical tracks, the Kalman prediction is performed on the current track to obtain the existing track; according to the Mahalanobis distance and the target confidence corresponding to the current measurement data, the one-to-one matching degree between each track in the existing track and each measurement data in the current measurement data is calculated; with the number of existing tracks as rows and the number of measurements of the current measurement data as columns, an association matrix is ​​constructed, and the one-to-one matching degree between the existing track and the current measurement data is assigned to each element in the association matrix to obtain the matching result.

[0093] First, an n×m cost matrix M is established as an association matrix, where n represents the number of existing tracks and m represents the number of measurements of the current measurement data. The Mahalanobis distance between the existing track and the current measurement data is calculated as the distance cost, and multiplied by the target confidence from the sensor perception result to obtain the one-to-one matching degree between each track and each measurement data:

[0094]

[0095] in represents the mean value of each state quantity of the existing track, z k Indicates the measurement from the sensor. represents the mean value of the sensor measurement, Σ-1 represents x k and z k The covariance matrix between them, score represents the target confidence.

[0096] Based on this, the matching degree between each track in the existing track and each measurement data in the current measurement data can be calculated according to formula (11), and then assigned to each element in the matrix M respectively, and the assigned matrix is ​​used as the matching result.

[0097] In this embodiment, the correlation matching between the current measurement data and the existing track is achieved based on the Mahalanobis distance and the target confidence, which can improve the accuracy of the matching result.

[0098] Additionally, in one embodiment, when it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data, Kalman updating is performed on the existing track according to the current measurement data, which may specifically include:

[0099] When the target measurement data that best matches the existing track is determined from the matching results according to the preset matching algorithm, the existing track is Kalman updated according to the target measurement data.

[0100] For example, the nearest neighbor matching algorithm can be used to determine the measurement data with the highest degree of matching with the existing track from the association matrix calculated above as the target measurement data, and the existing track is Kalman updated according to the target measurement data. The specific Kalman update process can be based on the above formulas (3) to (6), which will not be described in detail in this embodiment. In addition, other matching algorithms can also be used, such as the K-nearest neighbor algorithm or other distance-based matching algorithms.

[0101] By determining the measurement data that best matches the existing track based on the matching algorithm and performing Kalman update, the accuracy of track matching can be further improved.

[0102] In addition, in one embodiment, after determining the target measurement data that best matches the existing track from the matching results, and before performing Kalman update on the existing track according to the current measurement data, the target tracking method may further include:

[0103] Determine whether the position difference between the kinematic predicted position corresponding to the target measurement data and the kinematic predicted position corresponding to the existing track is less than a preset difference threshold; if so, perform Kalman update on the existing track according to the target measurement data; otherwise, discard the target measurement data.

[0104] In this embodiment, noise filtering of the measurement data is implemented. For the new measurement data that is matched, before performing the Kalman update, it is necessary to determine whether the new measurement data conforms to the kinematic law, that is, whether the position difference between Px and Py representing the position information in the matched current measurement data and Px and Py in the corresponding existing track is less than a preset difference threshold. If not, it is considered that the measurement data does not conform to kinematics and is discarded and no longer used to update the existing track.

[0105] In this embodiment, by adding noise filtering before Kalman updating, accurate updating of the track can be further achieved, avoiding interference of noise on the track update.

[0106] Optionally, in one embodiment, unstable measurement data can also be filtered. For example, after the sensor detects a target, the target needs to be detected continuously for at least 6 frames before outputting the measurement data corresponding to the target. Otherwise, the target is identified as noise. Based on this, the stability of the sensor output result can be improved.

[0107] Additionally, in one embodiment, the target tracking method may further include:

[0108] According to the matching results, a new track is created for the measurement data in the current measurement data that does not match the existing track. After the above-mentioned association matching process is completed, if there is measurement data in the current measurement data that does not match the existing track, a new track is created based on such measurement data. By creating a new track for the unmatched measurement data, it is possible to track the newly appeared tracking target, thereby providing more comprehensive data for the subsequent vehicle driving decision-making steps.

[0109] Additionally, in one embodiment, the target tracking method may further include:

[0110] If there are multiple tracks created based on the measurement data provided by the same sensor, the time information and position information of the multiple tracks are compared to obtain a track distance comparison result; at least two tracks among the multiple tracks whose track distance comparison results are less than a preset distance threshold are merged into one track.

[0111] Specifically, for single-sensor tracks detected only by millimeter-wave radar or forward-looking camera, the Euclidean distances of multiple tracks of the single sensor can be compared according to their appearance time and position information during the process, and similar tracks can be merged based on the comparison results. Similar tracks can be determined based on the time information and position information of different tracks, and then similar tracks can be merged, thereby simplifying track data and improving the efficiency of track maintenance and management.

[0112] Figure 3 is a flow chart of a target tracking method in some of the embodiments, such as Figure 3 As shown, the target tracking method includes the following steps:

[0113] Step S301, obtaining the acquisition result of the forward-looking camera; wherein, in the case that the forward-looking camera currently acquires measurement data, the measurement data currently acquired by the forward-looking camera is used as the first measurement data.

[0114] Step S302, obtaining the acquisition result of the millimeter wave radar; wherein, when the millimeter wave radar currently acquires measurement data, the measurement data acquired by the millimeter wave radar is used as the second measurement data. The execution order of step S301 and step S302 is not specifically limited, and the two steps can be performed simultaneously or successively.

[0115] Step S303, determining the fusion order of the measurement data of different sensors; wherein, according to the update status of the measurement data of the forward-looking camera and the millimeter-wave radar, the fusion order of the measurement data is determined in combination with the timestamp information corresponding to the measurement data, so as to associate and match with the existing track according to the fusion order.

[0116] Step S304, associate and match the current measurement data with the existing track, and determine whether there is measurement data matching the existing track; if so, execute step S305; otherwise, execute step S306; wherein, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, associate and match the current measurement data with the existing track to obtain a matching result.

[0117] Step S305: Perform Kalman update on the existing track according to the current measurement data.

[0118] Step S306: creating a new track for the unmatched measurement data.

[0119] In the above steps S301 to S306, the data collected by the sensor in real time is correlated and matched with the existing track, and the matched track is correlated with the front and rear frames, so as to improve the stable tracking of the forward-looking target and improve the stability of the vehicle's forward-looking perception.

[0120] In this embodiment, a target tracking device for a vehicle is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. The terms "module", "unit", "subunit", etc. used below can implement a combination of software and / or hardware of predetermined functions. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0121] Figure 4 4 is a structural block diagram of the target tracking device 40 for a vehicle according to the present embodiment. Figure 4 As shown, the target tracking device 40 for a vehicle includes: an acquisition module 42, a matching module 44 and an update module 46; wherein:

[0122] An acquisition module 42 is used to acquire an existing track of the tracked target and current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors; a matching module 44 is used to associate and match the current measurement data with the existing track based on the Mahalanobis distance and the target confidence corresponding to the current measurement data to obtain a matching result; an update module 46 is used to perform a Kalman update on the existing track according to the current measurement data when it is determined that there is measurement data matching the existing track in the current measurement data according to the matching result.

[0123] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0124] In this embodiment, a vehicle driving system is also provided. Figure 5 FIG. 5 is a schematic diagram of the structure of the vehicle driving system 50 of this embodiment. Figure 5 As shown, the vehicle driving system 50 includes: a forward-looking camera 52, a millimeter-wave radar 54 and a server 56; the forward-looking camera 52 and the millimeter-wave radar 54 are respectively connected to the server 56; the forward-looking camera 52 and the millimeter-wave radar 54 are respectively used to collect measurement data; the server 56 is used to execute the target tracking method for the vehicle provided in any of the above embodiments.

[0125] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and will not be repeated in this embodiment.

[0126] It should be understood that the specific embodiments described herein are only used to explain the application, rather than to limit it. Based on the embodiments provided in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the protection scope of this application.

[0127] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0128] Obviously, the drawings are only some examples or embodiments of the present application. For ordinary technicians in the field, the present application can also be applied to other similar situations based on these drawings without creative work. In addition, it is understandable that although the work done in this development process may be complicated and lengthy, for ordinary technicians in the field, certain changes in design, manufacturing or production based on the technical content disclosed in this application are only conventional technical means and should not be regarded as insufficient content disclosed in this application.

[0129] The term "embodiment" in this application refers to a specific feature, structure or characteristic described in conjunction with the embodiment that can be included in at least one embodiment of the present application. The appearance of this phrase in various locations in the specification does not necessarily mean the same embodiment, nor does it mean that it is mutually exclusive with other embodiments and is independent or optional. It is clearly or implicitly understood by those of ordinary skill in the art that the embodiments described in this application can be combined with other embodiments without conflict.

[0130] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of patent protection. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the attached claims.

Claims

1. A target tracking method for a vehicle, characterized in that: include: Acquire an existing track of the tracked target and acquire current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors; Based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result; When it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data, a Kalman update is performed on the existing track according to the current measurement data.

2. The target tracking method according to claim 1, characterized in that: Obtaining current measurement data collected by at least one sensor, including: Acquire the first measurement data currently collected by the forward-looking camera, and / or the second measurement data currently collected by the millimeter-wave radar; wherein, when the currently collected measurement data includes the first measurement data, convert the first measurement data from the camera coordinates to the preset absolute coordinates according to the preset coordinate conversion relationship.

3. The target tracking method according to claim 1, characterized in that: Based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result, including: Determining a fusion order of the current measurement data according to timestamp information corresponding to the current measurement data; According to the fusion order, based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, each current measurement data is sequentially associated and matched with the existing track to obtain a matching result.

4. The target tracking method according to claim 1 or 3, characterized in that: Based on the Mahalanobis distance and the target confidence corresponding to the current measurement data, the current measurement data is associated and matched with the existing track to obtain a matching result, including: According to the historical track, the Kalman prediction is performed on the current track to obtain the existing track; Calculate the one-to-one matching degree between each track in the existing track and each measurement data in the current measurement data according to the Mahalanobis distance and the target confidence corresponding to the current measurement data; A correlation matrix is ​​constructed with the number of existing tracks as rows and the number of measurements of current measurement data as columns, and the one-to-one matching degree between the existing tracks and the current measurement data is assigned to each element in the correlation matrix to obtain a matching result.

5. The target tracking method according to claim 1, characterized in that: When it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data, Kalman updating is performed on the existing track according to the current measurement data, including: When the target measurement data that best matches the existing track is determined from the matching results according to a preset matching algorithm, a Kalman update is performed on the existing track according to the target measurement data.

6. The target tracking method according to claim 5, characterized in that: After determining the target measurement data that best matches the existing track from the matching results, and before performing a Kalman update on the existing track according to the current measurement data, the method further includes: Determine whether the position difference between the kinematic predicted position corresponding to the target measurement data and the kinematic predicted position corresponding to the existing track is less than a preset difference threshold; if so, perform Kalman update on the existing track according to the target measurement data; otherwise, discard the target measurement data.

7. The target tracking method according to claim 1, characterized in that: The method further comprises: According to the matching result, a new track is created for the measurement data in the current measurement data that is not matched with the existing track.

8. The target tracking method according to claim 1, characterized in that: The method further comprises: If there are multiple tracks created based on the measurement data provided by the same sensor, then the time information and position information of the multiple tracks are compared to obtain a track distance comparison result; At least two tracks among the multiple tracks whose track distance comparison results are less than a preset distance threshold are merged into one track.

9. A target tracking device for a vehicle, characterized in that: include: Get modules, match modules, and update modules; where: The acquisition module is used to acquire an existing track of the tracked target and to acquire current measurement data collected by at least one sensor; wherein the existing track is constructed based on historical measurement data collected by at least two sensors; The matching module is used to associate and match the current measurement data with the existing track based on the Mahalanobis distance and the target confidence corresponding to the current measurement data to obtain a matching result; The updating module is used to perform Kalman update on the existing track according to the current measurement data when it is determined according to the matching result that there is measurement data matching the existing track in the current measurement data.

10. A vehicle driving system, characterized in that: include: A forward-looking camera, a millimeter-wave radar and a server; the forward-looking camera and the millimeter-wave radar are connected to the server respectively; The forward-looking camera and the millimeter-wave radar are respectively used to collect measurement data; The server is used to execute the target tracking method for a vehicle according to any one of claims 1 to 8.

Images