Multi-sensor fusion target tracking system confidence acquisition method, storage medium and electronic device

Abstract

The application discloses a multi-sensor fusion target tracking system confidence acquisition method, a storage medium and an electronic device. When a single sensor of a sensor perception system detects a target at T moment, the current sensor target confidence of a current sensor at the T moment and the fusion target confidence at T-1 moment are used to update the correction coefficient of the current sensor, and when the same target is detected by the sensor next time, the fusion target confidence is calculated after the sensor target confidence of the sensor is corrected by using the correction coefficient of the current sensor, so that the accuracy of the sensor target confidence of each sensor is improved, and the precision of the output target of the fusion system is improved.

Description

Technical Field

[0001] This application relates to the field of target tracking technology, and in particular to a method for obtaining confidence in a multi-sensor fusion target tracking system, a storage medium, and an electronic device. Background Technology

[0002] Obstacle detection in vehicles involves a perception system comprised of multiple sensors mounted on the vehicle body. The target output by this system is the fused result of the multi-sensor detections. However, the stability of this multi-sensor fusion target tracking system is easily affected by the performance of individual sensors. Individual sensors are susceptible to factors such as lighting, weather, wear, occlusion, external interference, and target shape, resulting in temporary or permanent degradation in their perception performance. Sensors with degraded performance will have lower accuracy in identifying targets.

[0003] In current multi-sensor fusion target tracking systems, the confidence level of a single sensor is generally set to the factory confidence level value. During the target tracking process, the confidence level of a single sensor remains unchanged, making it impossible to distinguish the recognition results of sensors with degraded perception performance, thus affecting the accuracy of the target results output by the multi-sensor fusion target tracking system. Summary of the Invention

[0004] The purpose of this application is to overcome the shortcomings of existing technologies that cannot distinguish the recognition results of some sensors with degraded perception performance, which affects the accuracy of the target results output by the multi-sensor fusion target tracking system. It provides a method, storage medium, and electronic device for obtaining the confidence of a multi-sensor fusion target tracking system that can dynamically evaluate the recognition performance of a single sensor, adjust the target confidence of a single sensor, and improve the accuracy of target tracking.

[0005] The technical solution of this application provides a method for obtaining confidence in a multi-sensor fusion target tracking system, including:

[0006] In response to the sensor sensing system detecting a target at time T, the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T are obtained;

[0007] Based on the target tracking algorithm, the fused target confidence at time T is obtained by taking the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor.

[0008] The correction coefficients of the current sensor are updated based on the sensor target confidence at time T and the fused target confidence at time T-1.

[0009] Furthermore, when time T is the initial time, then:

[0010] In response to the sensor sensing system detecting a target at time T, the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T are obtained, wherein the correction coefficient of the current sensor is a preset initial value;

[0011] Based on the target tracking algorithm, the fused target confidence at time T is obtained by taking the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor.

[0012] Furthermore, the step of updating the correction coefficient of the current sensor based on the sensor target confidence at time T and the fused target confidence at time T-1 specifically includes:

[0013] If the product of the current sensor target confidence and the current sensor correction coefficient at time T is less than the fused target confidence at time T-1, then the correction coefficient of the current sensor is increased.

[0014] If the product of the current sensor target confidence and the current sensor correction coefficient at time T is greater than the fusion target confidence at time T-1, then the correction coefficient of the current sensor is reduced.

[0015] Furthermore, the step of updating the correction coefficient of the current sensor based on the sensor target confidence at time T and the fused target confidence at time T-1 specifically includes:

[0016] An error function is constructed by multiplying the current sensor target confidence at time T by the correction coefficient of the current sensor and the fused target confidence at time T-1.

[0017] Construct a target function for the correction coefficients relative to the current sensor based on the error function;

[0018] The correction coefficients of the current sensor are updated according to the objective function.

[0019] Furthermore, updating the correction coefficients of the current sensor according to the objective function specifically includes:

[0020] Obtain the correction coefficient θ(k) of the current sensor, where k is an integer greater than or equal to 0. When k = 0, θ(0) is a preset initial value.

[0021] The gradient Δθ of the correction coefficient change is determined based on the objective function. θ Jfunc(θ), θ Jfunc(θ) is the partial derivative of the objective function Jfunc(θ) with respect to θ. When k = 0, Δθ = 0.

[0022] The correction coefficient of the current sensor is updated according to the gradient of the change of the correction coefficient, θ(k+1)=θ(k)-μ*Δθ, where μ is the proportional coefficient.

[0023] Furthermore, obtaining the fused target confidence at time T based on the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor specifically includes:

[0024] The false alarm rate and matching probability of the sensor are determined by multiplying the current sensor target confidence level and the current sensor correction coefficient at time T.

[0025] The change in confidence level of the fused target is determined based on the false alarm rate of the sensor and the matching probability.

[0026] The fusion target confidence level at time T is determined based on the change in the fusion target confidence level.

[0027] Furthermore, the target tracking algorithm is a Kalman filter tracking algorithm, a particle filter tracking algorithm, or a multiple hypothesis tracking algorithm.

[0028] The technical solution of this application also provides a storage medium that stores computer instructions. When the computer executes the computer instructions, it is used to perform the confidence acquisition method of the multi-sensor fusion target tracking system as described above.

[0029] The technical solution of this application also provides an electronic device, including at least one processor; and,

[0030] A memory communicatively connected to the at least one processor; wherein,

[0031] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform the confidence acquisition method for the multi-sensor fusion target tracking system as described above.

[0032] This application enables a sensor in a sensor perception system to update the correction coefficient of a single sensor when it detects a target at time T, based on the sensor target confidence of the current sensor at time T and the fused target confidence at time T-1. When the same sensor detects the same target again, the correction coefficient of the current sensor is used to correct the sensor target confidence of the sensor before calculating the fused target confidence, thereby improving the accuracy of the sensor target confidence of each sensor and thus improving the accuracy of the target output by the fusion system. Attached Figure Description

[0033] The disclosure of this application will become more readily understood with reference to the accompanying drawings. It should be understood that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this application. In the drawings:

[0034] Figure 1 This is a flowchart of a confidence acquisition method for a multi-sensor fusion target tracking system according to an embodiment of this application;

[0035] Figure 2 This is a flowchart of a method for obtaining confidence in a multi-sensor fusion target tracking system according to a preferred embodiment of this application;

[0036] Figure 3 This is a schematic diagram of the hardware structure of an electronic device in one embodiment of this application. Detailed Implementation

[0037] The specific embodiments of this application will be further described below with reference to the accompanying drawings.

[0038] It is readily understood that, based on the technical solution of this application, various structural and implementation methods can be interchanged by those skilled in the art without altering the essential spirit of this application. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this application and should not be considered as the entirety of this application or as limitations or restrictions on the technical solution of the application.

[0039] The directional terms such as up, down, left, right, front, back, front, back, top, and bottom mentioned or possibly used in this specification are defined relative to the structures shown in the accompanying drawings. These are relative concepts and may therefore vary depending on their location and usage. Therefore, these or other directional terms should not be interpreted as restrictive. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0040] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meanings of the above in this application according to the specific circumstances.

[0041] The confidence acquisition method of the multi-sensor fusion target tracking system in this application embodiment, such as Figure 1 As shown, it includes:

[0042] S101: In response to the sensor sensing system detecting a target at time T, obtain the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T;

[0043] S102: Based on the target tracking algorithm, the fused target confidence at time T is obtained according to the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor.

[0044] S103: Update the correction coefficients of the current sensor based on the sensor target confidence of the current sensor at time T and the fused target confidence at time T-1.

[0045] The sensor perception system consists of a variety of sensors installed on the vehicle body, including lidar, millimeter-wave radar, and cameras. When performing target detection, the same target may be detected by different sensors at different times. The multi-sensor fusion target tracking system can fuse the detection results of the same target detected by different sensors and finally output the target result.

[0046] As an example, for the same target, the lidar detects the first target information at the first moment, and the camera detects the second target information at the second moment. The multi-sensor fusion target tracking system then fuses the first target information and the second target information to obtain the fused target information. If the millimeter-wave radar detects the third target information at the third moment, the multi-sensor fusion target tracking system continues to fuse the third target information with the previously fused target information to update the fused target information. It continuously fuses the detected target information and updates the fused target information.

[0047] The sensor target confidence score is the confidence score of the target information detected by the current sensor, while the fused target confidence score is the confidence score of the fused target information. The fused target confidence score is affected by the sensor confidence score at each moment. Each time target fusion is performed, the fused target confidence score is updated according to the current sensor target confidence score until the score of the fused target confidence score is greater than or equal to the preset score. Then, the multi-sensor fusion target tracking system outputs the target result.

[0048] In existing technologies, the sensor target confidence level of each sensor is set to a fixed value. For example, when the lidar detects the first target information, the sensor target confidence level of the lidar is obtained to update the fused target confidence level. Even if the perception performance of the lidar deteriorates, the fused target confidence level is still updated with a higher sensor target confidence level, which ultimately affects the accuracy of the output target result.

[0049] In this embodiment, the sensor target confidence of each sensor is set to a dynamically adjusted correction coefficient for correction. When one of the sensors detects target information at time T, the correction coefficient of the current sensor is determined based on the fused target confidence at time T-1 and the sensor target confidence of the current sensor at time T. The correction coefficient of the current sensor is used to correct the sensor target confidence of the sensor when the sensor detects the same target information again.

[0050] Since the fused target information at time T-1 incorporates target information from multiple sensors prior to time T-1, the fused target confidence at time T-1 is closer to the true confidence. By determining the correction coefficient of the current sensor based on the fused target confidence at time T-1 and correcting the current sensor target confidence, the accuracy of the sensor target confidence can be improved.

[0051] Specifically, in step S101, when any sensor of the sensor perception system detects a target at time T, it acquires the correction coefficient of the current sensor (the correction coefficient is updated when the current sensor last detected a target) and the sensor target confidence of the current sensor at time T.

[0052] Step S102 corrects the current sensor target confidence by multiplying the current sensor's correction coefficient by the current sensor target confidence at time T. Then, a target tracking algorithm is used to obtain the fused target confidence at time T based on the corrected current sensor target confidence and the target information detected at time T, thereby updating the fused target confidence.

[0053] The target tracking algorithm can be a Kalman filter tracking algorithm, a particle filter tracking algorithm, or a multiple hypothesis tracking algorithm.

[0054] Step S103 is used to update the correction coefficients of the current sensor. It compares the fused target confidence at time T-1 with the sensor target confidence of the current sensor at time T, and corrects the correction coefficients of the current sensor accordingly. Specifically:

[0055] If the product of the current sensor target confidence and the current sensor correction coefficient at time T is less than the fused target confidence at time T-1, then the correction coefficient of the current sensor is increased.

[0056] If the product of the current sensor target confidence and the current sensor correction coefficient at time T is greater than the fusion target confidence at time T-1, then the correction coefficient of the current sensor is reduced.

[0057] Using the fusion target confidence level at time T-1 as a benchmark, the product of the current sensor's target confidence level and the current sensor's correction coefficient at time T is compared with the fusion target confidence level at time T-1. By adjusting the magnitude of the current sensor's correction coefficient, the difference between the product of the current sensor's target confidence level and the current sensor's correction coefficient at time T and the fusion target confidence level at time T-1 is reduced, thus adjusting the product of the current sensor's target confidence level and the current sensor's correction coefficient at time T towards the fusion target confidence level at time T-1.

[0058] Step S103 compares the product of the current sensor target confidence at time T and the current sensor correction coefficient with the fused target confidence at time T-1 to dynamically correct the current sensor correction coefficient. The corrected current sensor correction coefficient is used to correct the sensor target confidence the next time the current sensor detects the same target, thereby improving the reliability of the current sensor target confidence.

[0059] Furthermore, when time T is the initial time, then:

[0060] In response to the sensor sensing system detecting a target at time T, the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T are obtained, wherein the correction coefficient of the current sensor is a preset initial value;

[0061] Based on the target tracking algorithm, the fused target confidence at time T is obtained by taking the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor.

[0062] When time T is the initial time of target detection, the correction coefficient of the current sensor is the preset initial value. The fused target confidence at the initial time is determined directly based on the target detected at the initial time, the preset initial value of the correction coefficient of the current sensor, and the sensor target confidence of the current sensor at the initial time.

[0063] Furthermore, since there is no fusion target confidence level from the previous time step at the initial time step, the correction coefficient of the current sensor is not corrected at this time, and the correction coefficient of the current sensor remains at the preset initial value of the correction coefficient of the current sensor.

[0064] The correction coefficients of the current sensor are updated the next time the same target is detected by the current sensor.

[0065] It should be noted that each sensor in the sensor perception system is equipped with a corresponding correction coefficient. When each sensor detects the same target for the first time, the correction coefficient used is the sensor's preset initial value. Different sensors are set with different preset initial values.

[0066] As an example:

[0067] If the lidar detects the first target information of the current target at time t0 (initial time), then obtain the lidar sensor target confidence level b0 and the lidar correction coefficient α0 (α0 is a preset initial value) at time t0.

[0068] Based on the target tracking algorithm, the fused target confidence level a0 at time t0 is determined according to the first target information, the correction coefficient α0 of the lidar, and the sensor target confidence level b0 of the lidar.

[0069] If the camera detects the second target information of the current target at time t1, then the sensor target confidence level c0 of the camera and the camera correction coefficient β0 (β0 is a preset initial value) of the camera at time t1 are obtained.

[0070] Based on the target tracking algorithm, the fused target confidence a1 at time t1 is determined according to the second target information, the camera correction coefficient β0, and the camera sensor target confidence c0;

[0071] The camera's correction coefficient β1 is updated based on the sensor target confidence c0 of the camera at time t1 and the fused target confidence a0 at time t0.

[0072] If the camera detects the third target information of the current target at time t2, then the sensor target confidence level c1 and the camera correction coefficient β1 of the camera at time t2 are obtained.

[0073] Based on the target tracking algorithm, the fused target confidence a2 at time t2 is determined according to the third target information, the camera correction coefficient β1, and the camera sensor target confidence c1.

[0074] The camera's correction coefficient β2 is updated based on the sensor target confidence c1 of the camera at time t2 and the fused target confidence a1 at time t1.

[0075] If the lidar detects the fourth target information of the current target at time t3, then obtain the sensor target confidence level b1 and the lidar correction coefficient α0 of the lidar at time t3.

[0076] Based on the target tracking algorithm, the fused target confidence a3 at time t3 is determined according to the fourth target information, the correction coefficient α0 of the lidar, and the target confidence b1 of the lidar sensor.

[0077] The correction coefficient for the laser is updated to α1 based on the sensor target confidence level b1 of the lidar at time t3 and the fused target confidence level a2 at time t2.

[0078] In this embodiment, a dynamic correction coefficient is set for each sensor. Each time a sensor detects a target, the correction coefficient of the current sensor is corrected according to the fused target confidence level of the previous moment. The next time the sensor detects the same target, the correction coefficient is used to correct the target confidence level of the sensor at that time. The fused target confidence level is determined by the sensor target confidence level after correction coefficient. This can eliminate the interference of the detection results of the degraded sensor on the final output of the detected target and improve the accuracy of target recognition and target tracking.

[0079] In one embodiment, updating the correction coefficient of the current sensor based on the sensor target confidence at time T and the fused target confidence at time T-1 specifically includes:

[0080] An error function is constructed by multiplying the current sensor target confidence at time T by the correction coefficient of the current sensor and the fused target confidence at time T-1.

[0081] Construct a target function for the correction coefficients relative to the current sensor based on the error function;

[0082] The correction coefficients of the current sensor are updated according to the objective function.

[0083] In this embodiment, the error function of the current sensor's correction coefficient with respect to the fusion target confidence at time T is first constructed based on the product of the sensor target confidence at time T and the correction coefficient of the current sensor, and the fusion target confidence at time T-1. Then, based on the error function, the objective function of the current sensor's correction coefficient with respect to the fusion target confidence at time T-1 is constructed. Based on the objective function, the correction coefficient that minimizes the error between the current sensor's sensor target confidence and the fusion target confidence at time T-1 can be calculated.

[0084] Specifically, determining the correction coefficient based on the objective function includes:

[0085] Get the current correction coefficient θ(k), where k is an integer greater than or equal to 0. When k = 0, θ(0) is initialized.

[0086] The gradient Δθ of the correction coefficient change is determined based on the objective function. θ Jfunc(θ), θ Jfunc(θ) is the partial derivative of the objective function Jfunc(θ) with respect to θ. When k = 0, Δθ = 0.

[0087] The correction coefficient θ(k+1) = θ(k) - μ*Δθ is corrected according to the gradient of the change of the correction coefficient, where μ is the proportionality coefficient.

[0088] Using the gradient descent approach, the partial derivative of the objective function with respect to the correction coefficient θ is taken to determine the gradient Δθ of the correction coefficient change. The gradient Δθ of the correction coefficient change is multiplied by the scaling factor μ to obtain the change in the correction coefficient. The scaling factor μ is set according to the specific requirements of the sensor sensing system.

[0089] The embodiments of this application use the idea of ​​gradient descent to correct the sensor target confidence. It can obtain a more appropriate correction coefficient based on the error between the target confidence at time T-1 and the current sensor target confidence at time T, thereby improving the accuracy of the current sensor target confidence.

[0090] In one embodiment, obtaining the fused target confidence at time T based on the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor specifically includes:

[0091] The false alarm rate and matching probability of the sensor are determined by multiplying the current sensor target confidence level and the current sensor correction coefficient at time T.

[0092] The change in confidence level of the fused target is determined based on the false alarm rate of the sensor and the matching probability.

[0093] The fusion target confidence level at time T is determined based on the change in the fusion target confidence level.

[0094] In a multi-sensor fusion target tracking system, the presence of a target needs to be determined based on the fusion target confidence score. Each time the sensor perception system detects a target, it adds the fusion target confidence score change value to the previous fusion target confidence score until the fusion target confidence score reaches the set value.

[0095] Specifically, the false alarm rate P(DK|H0) and the matching probability P(DK|H1) can be determined by multiplying the current sensor target confidence level and the current sensor correction coefficient at time T.

[0096] The sensor likelihood ratio is calculated by dividing the matching probability by the sensor false alarm rate.

[0097] LR = P(DK|H1) / P(DK|H0);

[0098] The change in the fused target confidence is obtained by taking the logarithm of the sensor likelihood ratio (LR).

[0099] ΔLLR=ln(P(DK|H1) k / P(DK|H0) k );

[0100] This leads to the fusion target confidence level at time T.

[0101] LLR(k)=LLR(k-1)+ΔLLR.

[0102] In this embodiment, the fused target confidence at time T is determined by multiplying the current sensor target confidence at time T by the correction coefficient of the current sensor. The higher the sensor target confidence, the higher the change in the fused target confidence, and thus the faster the target is determined based on the fused target confidence. Therefore, the detection results of sensors with better performance have a greater weight in the fused target confidence, while the detection results of sensors with poorer performance have a smaller weight, thereby achieving the stability and reliability of the entire perception fusion system.

[0103] Figure 2 This invention presents a flowchart illustrating a method for obtaining confidence in a multi-sensor fusion target tracking system according to a preferred embodiment of the present application, which specifically includes:

[0104] Step S201: In response to the sensor sensing system detecting a target at the initial moment, the correction coefficient of the current sensor and the sensor target confidence of the current sensor at the initial moment are obtained, wherein the correction coefficient of the current sensor is a preset initial value;

[0105] Step S202: Determine the sensor false alarm rate and matching probability based on the product of the sensor target confidence level of the current sensor at the initial moment and the correction coefficient of the current sensor;

[0106] Step S203: Determine the change in confidence level of the fused target based on the false alarm rate of the sensor and the matching probability;

[0107] Step S204: Determine the initial fusion target confidence level based on the change value of the fusion target confidence level;

[0108] Step S205: In response to the sensor sensing system detecting a target at time T, obtain the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T;

[0109] Step S206: Determine the sensor false alarm rate and matching probability based on the product of the sensor target confidence level of the current sensor at time T and the correction coefficient of the current sensor;

[0110] Step S207: Determine the change in confidence of the fused target based on the false alarm rate of the sensor and the matching probability;

[0111] Step S208: Determine the fusion target confidence at time T based on the change value of the fusion target confidence;

[0112] Step S209: Construct an error function based on the product of the current sensor target confidence at time T and the correction coefficient of the current sensor, and the fused target confidence at time T-1;

[0113] Step S210: Construct an objective function for the correction coefficients relative to the current sensor based on the error function;

[0114] Step S211: Obtain the correction coefficient of the current sensor;

[0115] Step S212: Determine the gradient of the correction coefficient change based on the objective function;

[0116] Step S213: Update the correction coefficient of the current sensor according to the gradient of the change of the correction coefficient, and return to step S205.

[0117] The technical solution of this application also provides a storage medium that stores computer instructions. When the computer executes the computer instructions, it is used to execute the confidence acquisition method of the multi-sensor fusion target tracking system in any of the foregoing embodiments.

[0118] Figure 3 An electronic device according to this application is shown, comprising:

[0119] At least one processor 301; and,

[0120] The memory 302 is communicatively connected to the at least one processor 301; wherein,

[0121] The memory 302 stores instructions that can be executed by the at least one processor 301, which, when executed by the at least one processor 301, enables the at least one processor 301 to perform all steps of the confidence acquisition method for the multi-sensor fusion target tracking system in any of the foregoing method embodiments.

[0122] The electronic device is preferably an in-vehicle electronic control unit (ECU), and more specifically a microcontroller unit (MCU) within the in-vehicle electronic control unit.

[0123] Figure 3 Taking processor 301 as an example:

[0124] The electronic device may also include an input device 303 and an output device 304.

[0125] The processor 301, memory 302, input device 303 and output device 304 can be connected by a bus or other means. The figure shows an example of connection by bus.

[0126] Memory 302, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the confidence acquisition method of the multi-sensor fusion target tracking system in the embodiments of this application, for example, Figure 1 The method flow is shown in Figure 2. The processor 301 executes various functional applications and data processing by running non-volatile software programs, instructions, and modules stored in the memory 302, thereby realizing the confidence acquisition method of the multi-sensor fusion target tracking system in the above embodiments.

[0127] Memory 302 may include a program storage area and a data storage area. The program storage area may store an operating system and applications required for at least one function; the data storage area may store data created based on the use of the multi-sensor fusion target tracking system confidence acquisition method, etc. Furthermore, memory 302 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 302 may optionally include memory remotely located relative to processor 301, and this remote memory may be connected via a network to the apparatus performing the multi-sensor fusion target tracking system confidence acquisition method. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0128] The input device 303 can receive user clicks and generate signal inputs related to user settings and function control of the multi-sensor fusion target tracking system confidence acquisition method. The output device 304 may include a display device such as a screen.

[0129] When one or more modules are stored in the memory 302, and are run by one or more processors 301, the confidence acquisition method of the multi-sensor fusion target tracking system in any of the above method embodiments is executed.

[0130] The above description is merely the principle and preferred embodiment of this application. It should be noted that for those skilled in the art, implementation methods obtained by appropriately combining the technical solutions disclosed in different embodiments are also included within the technical scope of this invention. Based on the principle of this application, several other modifications can also be made, which should also be considered within the protection scope of this application.

Claims

1. A method for obtaining confidence in a multi-sensor fusion target tracking system, characterized in that, include: In response to the sensor sensing system detecting a target at time T, the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T are obtained; Based on the target tracking algorithm, the fused target confidence at time T is obtained by taking the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor. The correction coefficient of the current sensor is updated based on the sensor target confidence of the current sensor at time T and the fused target confidence at time T-1. The correction coefficient of the current sensor is used to correct the sensor target confidence when the sensor detects the same target information again. The step of updating the correction coefficients of the current sensor based on the sensor target confidence at time T and the fused target confidence at time T-1 specifically includes: If the product of the current sensor target confidence and the current sensor correction coefficient at time T is less than the fused target confidence at time T-1, then the correction coefficient of the current sensor is increased. If the product of the current sensor target confidence and the current sensor correction coefficient at time T is greater than the fusion target confidence at time T-1, then the correction coefficient of the current sensor is reduced.

2. The confidence acquisition method for a multi-sensor fusion target tracking system according to claim 1, characterized in that, When time T is the initial time, then: In response to the sensor sensing system detecting a target at time T, the correction coefficient of the current sensor and the sensor target confidence of the current sensor at time T are obtained, wherein the correction coefficient of the current sensor is a preset initial value; Based on the target tracking algorithm, the fused target confidence at time T is obtained by taking the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor.

3. The confidence acquisition method for a multi-sensor fusion target tracking system according to claim 1, characterized in that, The step of updating the correction coefficients of the current sensor based on the sensor target confidence at time T and the fused target confidence at time T-1 specifically includes: An error function is constructed by multiplying the current sensor target confidence at time T by the correction coefficient of the current sensor and the fused target confidence at time T-1. Construct a target function for the correction coefficients relative to the current sensor based on the error function; The correction coefficients of the current sensor are updated according to the objective function.

4. The confidence acquisition method for a multi-sensor fusion target tracking system according to claim 3, characterized in that, The step of updating the correction coefficient of the current sensor according to the objective function specifically includes: Obtain the correction coefficient θ(k) of the current sensor, where k is an integer greater than or equal to 0. When k = 0, θ(0) is a preset initial value. The gradient Δθ of the correction coefficient change is determined based on the objective function. θ Jfunc(θ), θ Jfunc(θ) is the partial derivative of the objective function Jfunc(θ) with respect to θ. When k = 0, Δθ = 0. The correction coefficient of the current sensor is updated according to the gradient of the change of the correction coefficient, θ(k+1)=θ(k)-μ*Δθ, where μ is the proportional coefficient.

5. The confidence acquisition method for a multi-sensor fusion target tracking system according to claim 1, characterized in that, The step of obtaining the fused target confidence at time T based on the target information detected at time T, the sensor target confidence of the current sensor at time T, and the correction coefficient of the current sensor specifically includes: The false alarm rate and matching probability of the sensor are determined by multiplying the current sensor target confidence level and the current sensor correction coefficient at time T. The change in confidence level of the fused target is determined based on the false alarm rate of the sensor and the matching probability. The fusion target confidence level at time T is determined based on the change in the fusion target confidence level.

6. The confidence acquisition method for a multi-sensor fusion target tracking system according to claim 1, characterized in that, The target tracking algorithm is a Kalman filter tracking algorithm, a particle filter tracking algorithm, or a multiple hypothesis tracking algorithm.

7. A storage medium, characterized in that, The storage medium stores computer instructions, which, when executed by the computer, are used to perform the confidence acquisition method for the multi-sensor fusion target tracking system as described in any one of claims 1-6.

8. An electronic device, characterized in that, Includes at least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the confidence acquisition method for a multi-sensor fusion target tracking system as described in any one of claims 1-6.

Images