Vehicle control method, electronic equipment and vehicle

Abstract

The invention relates to the technical field of vehicles, and discloses a vehicle control method, electronic equipment and a vehicle, and the method comprises the steps: obtaining detection data of at least two sensors; obtaining target fusion information based on the detection data and historical fusion information; and controlling the vehicle based on the target fusion information. In the process of calculating the new target fusion information, the historical calculation result of the historical fusion information is applied, so that the current data can be analyzed through the historical data, misjudgment of different sensors on different targets is avoided, the problem of misfusion of the targets of different sensors in the driving process is solved, and the driving safety is improved. And the accuracy of the target fusion information is improved, so that the accuracy of vehicle control is improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle control technology, and in particular to a vehicle control method, electronic equipment and a vehicle. Background Art

[0002] With the development of autonomous driving technology, the algorithms of multi-sensor fusion modules are also constantly improving. As the downstream of the autonomous driving sensor perception module and the upstream of the autonomous driving planning and control module, the sensor fusion module plays a connecting role.

[0003] The current multi-sensor fusion method mainly uses the distance matrix between the observed target of the current sensor and the fusion target Track, and calls the matching algorithm based on the distance matrix for matching. However, in the process of the existing method, historical information is not applied. Therefore, when different targets from different sensors are close to each other, the matching algorithm may determine that the two targets are the same target and perform incorrect fusion, which will have a serious impact on the downstream autonomous driving planning and control modules, resulting in vehicle control errors, such as the possibility of erroneous triggering of the automatic emergency braking system. Summary of the invention

[0004] The embodiments of the present application provide a vehicle control method, an electronic device, and a vehicle, which aim to solve the problem of mis-fusion of targets of different sensors during driving and improve the accuracy of vehicle control.

[0005] The present application provides a vehicle control method, including:

[0006] Acquire detection data of at least two sensors;

[0007] Obtaining target fusion information based on the detection data and historical fusion information;

[0008] The vehicle is controlled based on the target fusion information.

[0009] According to the vehicle control method provided by the present application, the at least two sensors include a first sensor and a second sensor;

[0010] The obtaining of target fusion information based on the detection data and the historical fusion information includes:

[0011] Performing difference calculation based on first detection data of the first sensor and second detection data of the second sensor to obtain a detection data difference;

[0012] The historical fusion information is updated based on the detection data difference to obtain the target fusion information.

[0013] According to the vehicle control method provided by the present application, the detection data includes at least one of a lateral distance of the vehicle in a lateral direction, a longitudinal distance of the vehicle in a longitudinal direction, and a vehicle speed of the vehicle.

[0014] According to the vehicle control method provided by the present application, the difference calculation based on the first detection data of the first sensor and the second detection data of the second sensor to obtain the detection data difference includes:

[0015] The detection data difference is obtained by performing a difference calculation based on at least one of the lateral distance, longitudinal distance and vehicle speed of the first sensor and the corresponding at least one of the lateral distance, longitudinal distance and vehicle speed of the second sensor.

[0016] According to the vehicle control method provided by the present application, the detection data difference includes at least one of a lateral distance difference, a longitudinal distance difference and a vehicle speed difference.

[0017] According to the vehicle control method provided by the present application, the method further includes:

[0018] If at least one of the lateral distance difference, the longitudinal distance difference and the vehicle speed difference is greater than a corresponding preset threshold, determining that the detection data difference does not meet a preset calculation threshold;

[0019] If the lateral distance difference, the longitudinal distance difference and the vehicle speed difference are all less than or equal to corresponding preset thresholds, it is determined that the detection data difference meets the calculation threshold.

[0020] According to the vehicle control method provided by the present application, the updating of the historical fusion information based on the detection data difference to obtain the target fusion information includes:

[0021] If the detection data difference does not meet the preset calculation threshold, the historical fusion information is updated based on the first identifier of the first sensor, the second identifier of the second sensor and a preset threshold result to obtain the target fusion information.

[0022] According to the vehicle control method provided by the present application, the updating of the historical fusion information based on the first identification of the first sensor, the second identification of the second sensor and the preset threshold result to obtain the target fusion information includes:

[0023] The threshold result is used to replace the historical calculation result in which the column is marked with the first identifier and the row is marked with the second identifier in the historical fusion information to obtain the target fusion information.

[0024] According to the vehicle control method provided by the present application, the updating of the historical fusion information based on the detection data difference to obtain the target fusion information includes:

[0025] If the detection data difference satisfies a preset calculation threshold, a Mahalanobis distance calculation is performed based on the first detection data and the second detection data to obtain a current calculation result;

[0026] According to the first identifier of the first sensor and the second identifier of the second sensor, matching is performed in the historical fusion information to obtain a matching result;

[0027] The historical fusion information is updated based on the matching result and the current calculation result to obtain target fusion information.

[0028] According to the vehicle control method provided by the present application, matching is performed in the historical fusion information according to the first identifier of the first sensor and the second identifier of the second sensor to obtain a matching result, including:

[0029] The first identifier is used as a column identifier, and the second identifier is used as a row identifier, matching is performed in the historical fusion information, and the obtained historical calculation result is used as the matching result.

[0030] According to the vehicle control method provided by the present application, the updating of the historical fusion information based on the matching result and the current calculation result to obtain the target fusion information includes:

[0031] If the matching result is a historical calculation result corresponding to the first identifier and the second identifier in the historical fusion information, a calculation is performed based on the historical calculation result and the current calculation result and their corresponding weight values ​​to obtain a first final result;

[0032] The historical calculation result in the historical fusion information is updated based on the first final result to obtain the target fusion information.

[0033] According to the vehicle control method provided by this application, the specific calculation formula of the first final result is:

[0034] D1=Dist×(1-gain)+Dist_old×gain;

[0035] Among them, D1 represents the first final result, Dist represents the current calculation result, Dist_old represents the historical calculation result, 1-gain represents the weight value of the current calculation result, and gain represents the weight value of the historical calculation result.

[0036] According to the vehicle control method provided by the present application, the updating of the historical fusion information based on the matching result and the current calculation result to obtain the target fusion information includes:

[0037] If the matching result is that the historical calculation result corresponding to the first identifier and the second identifier does not exist in the historical fusion information, then calculating based on the current calculation result and the preset threshold result and its corresponding weight value to obtain a second final result;

[0038] The historical calculation result in the historical fusion information is updated based on the second final result to obtain the target fusion information.

[0039] According to the vehicle control method provided by this application, the specific calculation formula of the second final result is:

[0040] D2 = Dist × n + Gate × m;

[0041] Among them, D2 represents the second final result, Dist represents the current calculation result, Gate represents the threshold result, n represents the weight value of the current calculation result, and m represents the weight value of the threshold result.

[0042] According to the vehicle control method provided by the present application, the step of acquiring detection data of at least two sensors includes:

[0043] Acquire first detection data of the first sensor;

[0044] Based on the first identification of the first sensor, matching is performed in the detection data list, and detection data in the detection data list with a second identification identical to the first identification is used as second detection data of the second sensor.

[0045] According to the vehicle control method provided by the present application, the sensor includes a camera and a radar.

[0046] According to the vehicle control method provided by the present application, the detection data list includes a camera detection data list and / or a radar detection data list.

[0047] According to the vehicle control method provided by the present application, the first identification based on the first sensor is matched in the detection data list, and the detection data in the detection data list having the same second identification as the first identification is used as the second detection data of the second sensor, including:

[0048] Based on the first identifier, matching is performed in the camera detection data list and / or the radar detection data list, and detection data in the camera detection data list and / or the radar detection data list whose second identifier is the same as the first identifier is used as the second detection data of the second sensor.

[0049] The present application also provides a vehicle control device, including:

[0050] An acquisition module, used to acquire detection data of at least two sensors;

[0051] A determination module, used for obtaining target fusion information based on the detection data and historical fusion information;

[0052] A vehicle control module is used to control the vehicle based on the target fusion information.

[0053] An embodiment of the present application also provides an electronic device, including a processor and a memory, wherein the memory stores multiple computer programs; the processor loads the computer program from the memory to execute any vehicle control method provided in the embodiment of the present application.

[0054] An embodiment of the present application also provides a computer-readable storage medium, which stores multiple computer programs that can be executed by a vehicle. When the computer programs are run on the vehicle, the vehicle executes any one of the vehicle control methods provided in the embodiments of the present application.

[0055] An embodiment of the present application also provides a computer program product, which includes a computer program, and the computer program is executed by a processor to perform any one of the vehicle control methods provided in the embodiments of the present application.

[0056] An embodiment of the present application also provides a vehicle, which includes the vehicle control device, or / and, the electronic device, or / and, the computer-readable storage medium, or / and, the computer program product.

[0057] The vehicle control method provided in the embodiment of the present application obtains detection data of at least two sensors; obtains target fusion information based on the detection data and historical fusion information; and controls the vehicle based on the target fusion information. Therefore, in the process of calculating new target fusion information, the historical calculation results of the historical fusion information are applied, so that current data can be analyzed through historical data, avoiding misjudgment of different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, improving the accuracy of target fusion information, and thus improving the accuracy of vehicle control. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0059] Figure 1 is a flow chart of a vehicle control method provided in one embodiment of the present application;

[0060] Figure 2 It is one of the flowcharts of the second detection data matching provided in one embodiment of the present application;

[0061] Figure 3 This is the second schematic diagram of the flow chart of the second detection data matching provided in one embodiment of the present application;

[0062] Figure 4 is a schematic diagram of the structure of a distance matrix provided in one embodiment of the present application;

[0063] Figure 5 is a flow chart of a matching fusion solution provided in one embodiment of the present application;

[0064] Figure 6 is a schematic diagram of a horizontal scene provided in an embodiment of the present application;

[0065] Figure 7 is a schematic diagram of a vertical scene provided in one embodiment of the present application;

[0066] Figure 8 is a schematic diagram for comparing matching results of a scenario provided in an embodiment of the present application;

[0067] Fig. 9 is a schematic structural diagram of a vehicle control device provided in one embodiment of the present application;

[0068] Fig.10 It is a schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0069] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application. At the same time, in the description of the embodiments of the present application, the terms "first", "second", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0070] It should be noted that the description order of the following embodiments is not intended to limit the preferred order of the embodiments. Although the logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in an order different from that shown in the drawings.

[0071] With the development of autonomous driving technology, the algorithm of multi-sensor fusion module is also constantly improving. As the downstream of autonomous driving sensor perception module and the upstream of autonomous driving planning and control module, the sensor fusion module plays a connecting role. The current multi-sensor fusion method mainly uses the distance matrix between the observation target of the current sensor and the fusion target Track, and calls the matching algorithm based on the distance matrix for matching. However, in the process of the existing method, historical information is not applied. Therefore, when different targets from different sensors are close to each other, the matching algorithm may determine that the two targets are the same target and perform incorrect fusion. For example, when pedestrians traveling horizontally relative to the vehicle and fences on the side of the road are close in position, mis-fusion will occur. However, due to the limitation of sensor performance, there will be a situation where target detection is lost. For example, the camera only detects pedestrians and the radar only detects pillars. The trust in different sensors is different. Therefore, the target after mis-fusion will trust the radar more in the x direction and the camera more in the y direction, which will cause the mis-fused target to be closer to the vehicle relative to the position detected by the camera, and the actual situation is very far away, which will have a serious impact on the downstream autonomous driving planning and control modules, resulting in vehicle control errors and may trigger the automatic emergency braking system by mistake.

[0072] The inventor of the present application proposed a method for analyzing current data through historical data, that is, in the process of calculating new target fusion information, the historical calculation results of historical fusion information are applied, so that current data can be analyzed through historical data, avoiding misjudgment between different sensors for different targets, solving the problem of misfusion of targets of different sensors during driving, improving the accuracy of target fusion information, and thus improving the accuracy of vehicle control.

[0073] Glossary

[0074] Track is a storage unit for the fusion result of the fusion module. It has three main forms: Track composed of a single camera, Track composed of a single radar, and Track composed of camera and radar fusion. In addition to recording the target IDs that make up the Track, Track also records the target information after fusion.

[0075] The Track list is a list of fusion results of the entire fusion system, and its constituent unit is Track.

[0076] Autonomous Emergency Braking (AEB).

[0077] Gate is the threshold value of each element of the distance matrix and is the maximum value in the distance matrix.

[0078] Kalman Filter is an algorithm that uses the linear system state equation to optimally estimate the system state through the system input and output observation data. Since the observation data includes the influence of noise and interference in the system, the optimal estimation can also be regarded as a filtering process.

[0079] The execution subject of the embodiment of the present application is described by taking a vehicle control device as an example. Figure 1 , Figure 1 1 is a flow chart of a vehicle control method provided by an embodiment of the present application. The vehicle control method is specifically described in steps 10 to 30:

[0080] Step 10: Acquire detection data from at least two sensors.

[0081] In this embodiment, the at least two sensors include a first sensor and a second sensor. The second sensor is a sensor constituting a Track storage unit. Since the Track storage unit includes a Track consisting of a single camera, a Track consisting of a single radar, and a Track consisting of a camera and a radar fusion, it can be understood that the second sensor can be a camera, a radar, or a combination of a camera and a radar. The first sensor is a single perception sensor, so the first sensor can be a camera or a radar.

[0082] Therefore, the vehicle control device obtains the detection data of the observed target from the first sensor and the second sensor constituting the Track storage unit, as described in step 101 to step 102. At the same time, the vehicle state data and the vehicle body state data such as the speed and the rate of change of the vehicle body heading angle are obtained from the vehicle-mounted sensor.

[0083] It should be noted that since the acquired detection data is the relative speed and relative distance to the vehicle, it is necessary to pre-process the detection data through the vehicle status data to modify the relative speed and relative distance of the detection data to the vehicle into the absolute speed and absolute distance to the vehicle.

[0084] Step 20, obtaining target fusion information based on the detection data and historical fusion information.

[0085] The vehicle control device matches the historical data in the historical fusion information, calculates new data based on the detection data and the historical data, and then updates the historical data in the historical fusion information based on the new data to obtain target fusion information, as described in steps 201 to 202.

[0086] Step 30: Control the vehicle based on the target fusion information.

[0087] The vehicle control device controls the vehicle through the target fusion information. In one embodiment, the target fusion information may be a target distance matrix, and therefore specifically includes:

[0088] The target distance matrix is ​​matched in rows and columns by the nearest neighbor association algorithm to obtain the matching result. Specifically, the minimum value of the number of rows and columns of the target distance matrix is ​​calculated, that is, n = min (number of rows, number of columns), and the minimum value in the entire target distance matrix is ​​found by looping from 1 to n. After the minimum value is found, its row i and column j are locked. Row i, column j has been matched and will no longer participate in subsequent loops. Continue to search for the minimum value in the target distance matrix excluding row i and column j until the entire target distance matrix is ​​traversed to obtain the matching result.

[0089] According to the matching results, the Kalman filter method is used to complete the update of the Track in the Track list. The Tracks that have not been updated and the Tracks that need to be created are managed to complete the track management.

[0090] Calculate the priority and quality of each track and adjust it to the format required by the downstream. For the downstream, the target has a priority. For example, if the vehicle is following the vehicle in front, the priority of the vehicle in front must be higher than other targets farther away. Therefore, post-processing will assign priorities to all tracks and assign quality information to the target to allow the downstream to make better choices.

[0091] In the process of calculating new target fusion information, the embodiment of the present application applies the historical calculation results of historical fusion information, so that current data can be analyzed through historical data, avoiding misjudgment between different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, and improving the accuracy of target fusion information, thereby improving the accuracy of vehicle control.

[0092] In one embodiment, the description of step 101 to step 102 includes:

[0093] Step 101, obtaining first detection data of the first sensor;

[0094] Step 102 : matching in a detection data list based on the first identifier of the first sensor, and taking detection data in the detection data list whose second identifier is the same as the first identifier as second detection data of the second sensor.

[0095] The vehicle control device obtains the first detection data of the current frame from the first sensor, and obtains the second detection data of the second sensor constituting the Track storage unit. The process of obtaining the second detection data is as follows:

[0096] The vehicle control device obtains the first identifier of the first sensor, matches the detection data list according to the first identifier of the first sensor, and uses the detection data in the detection data list with the same second identifier as the first identifier as the second detection data of the second sensor.

[0097] In this embodiment, the second sensor may be a camera, a radar, or a combination of a camera and a radar. Therefore, the detection data list includes a camera detection data list, a radar detection data list, or a combination list of camera detection data and radar detection data.

[0098] Therefore, the vehicle control device matches the camera detection data list and / or the radar detection data list according to the first identifier, and takes the detection data in the camera detection data list and / or the radar detection data list whose second identifier is the same as the first identifier as the second detection data of the second sensor.

[0099] Further, taking the case where the input of this frame is the detection data of the camera, that is, the input of this frame is the target information detected by the camera, refer to Figure 2 , Figure 2This is one of the flowcharts of the second detection data matching provided in one embodiment of the present application. [camera id, -1] is a Track composed of a single camera, [-1, radar id] is a Track composed of a single radar, and [cameraid, radar id] is a Track composed of a camera and a radar, and any target detection data in the detection data of the camera is taken. Therefore, for the Track of [camera id, -1], a match is performed in the Track of [camera id, -1] according to the first identifier of the target detection data, and the detection data corresponding to the second identifier in the Track of [camera id, -1] that is the same as the first identifier is obtained. For the Track of [-1, radar id], a match is performed in the Track of [-1, radar id] according to the first identifier of the target detection data, and the detection data corresponding to the second identifier in the Track of [-1, radar id] that is the same as the first identifier is obtained. For the track with [camera id, radar id], first search for radarid according to the first identifier of the target detection data, then search for camera id, and obtain the detection data corresponding to the second identifier that is the same as the first identifier. Finally, the detection data obtained in the three tracks is used as the second detection data of the second sensor.

[0100] Similarly, the input for this frame is the detection data of the radar, that is, the input for this frame is the target information detected by the radar, refer to Figure 3 , Figure 3 This is the second flowchart of the second detection data matching provided in one embodiment of the present application.

[0101] In one embodiment, the description of step 201 to step 202 includes:

[0102] Step 201, performing difference calculation based on first detection data of the first sensor and second detection data of the second sensor to obtain a detection data difference;

[0103] Step 202: Update the historical fusion information based on the detection data difference to obtain the target fusion information.

[0104] In this embodiment, the detection data may include the lateral distance of the vehicle in the lateral direction, the longitudinal distance of the vehicle in the longitudinal direction and at least one of the vehicle speed of the vehicle. Therefore, the detection data difference may include at least one of the lateral distance difference, the longitudinal distance difference and the vehicle speed difference.

[0105] Therefore, it can be understood that the vehicle control device will obtain at least one of the lateral distance of the vehicle in the lateral direction, the longitudinal distance of the vehicle in the longitudinal direction and the vehicle speed of the vehicle from the first sensor, and perform difference calculation with at least one of the corresponding lateral distance, longitudinal distance and vehicle speed obtained from the second sensor to obtain at least one of the lateral distance difference, longitudinal distance difference and vehicle speed difference of the observed target in the first sensor and the second sensor.

[0106] In one embodiment, the detection data may include lateral distance, longitudinal distance and vehicle speed. Therefore, the difference between the lateral distance in the first sensor and the lateral distance in the second sensor is calculated to obtain the lateral distance difference, the difference between the longitudinal distance in the first sensor and the longitudinal distance in the second sensor is calculated to obtain the longitudinal distance difference, and the difference between the vehicle speed in the first sensor and the vehicle speed in the second sensor is calculated to obtain the vehicle speed difference.

[0107] The vehicle control device updates the historical fusion information according to the detection data difference to obtain the target fusion information, that is, updates the historical fusion information according to the lateral distance difference, the longitudinal distance difference and the vehicle speed difference to obtain the target fusion information, which specifically includes the description of steps 2021 to 2024.

[0108] In the process of calculating new target fusion information, the embodiment of the present application applies the historical calculation results of historical fusion information, so that current data can be analyzed through historical data, avoiding misjudgment between different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, and improving the accuracy of target fusion information, thereby improving the accuracy of vehicle control.

[0109] In one embodiment, the description of step 2021 includes:

[0110] Step 2021: if the detection data difference does not meet the preset calculation threshold, the historical fusion information is updated based on the first identifier of the first sensor, the second identifier of the second sensor and the preset threshold result to obtain the target fusion information.

[0111] In this embodiment, the historical fusion information can be a historical distance matrix, referring to Figure 4 , Figure 4 It is a schematic diagram of the structure of the distance matrix provided in one embodiment of the present application.

[0112] It should be noted that, before this, it is necessary to determine whether the detection data difference satisfies a preset calculation threshold, wherein the calculation threshold is set according to actual conditions.

[0113] If it is determined that at least one of the lateral distance difference, the longitudinal distance difference, and the vehicle speed difference is greater than the corresponding preset threshold, it is determined that the detection data difference does not meet the preset calculation threshold. If it is determined that the lateral distance difference, the longitudinal distance difference, and the vehicle speed difference are all less than or equal to the corresponding preset threshold, it is determined that the detection data difference meets the calculation threshold.

[0114] Therefore, if it is determined that the detection data difference does not meet the preset calculation threshold, that is, the lateral distance difference is greater than the lateral distance threshold, or / and, the longitudinal distance difference is greater than the longitudinal distance threshold, or / and, the vehicle speed difference is greater than the speed threshold, the vehicle control device obtains the preset threshold result, wherein the lateral distance threshold, the longitudinal distance threshold, the speed threshold and the preset threshold result are set according to the actual situation, for example, the lateral distance threshold is 12m, the longitudinal distance threshold is 2m, the speed threshold is 12m / s, and the preset threshold result is Gate.

[0115] The vehicle control device updates the historical fusion information according to the first identifier of the first sensor, the second identifier of the second sensor and the preset threshold result to obtain the target fusion information. Specifically, the threshold result is used to replace the historical calculation results in the historical distance matrix, in which the column identifier is the first identifier of the first sensor and the row identifier is the second identifier of the second sensor, to complete the update of the historical distance matrix and obtain the target fusion information. The target fusion information is the updated historical distance matrix.

[0116] In the process of calculating new target fusion information, the embodiment of the present application applies the historical calculation results of historical fusion information, so that current data can be analyzed through historical data, avoiding misjudgment between different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, and improving the accuracy of target fusion information, thereby improving the accuracy of vehicle control.

[0117] In one embodiment, the description of step 2022 to step 2024 includes:

[0118] Step 2022, if the difference in the detection data meets the preset calculation threshold, a Mahalanobis distance calculation is performed based on the first detection data and the second detection data to obtain a current calculation result;

[0119] Step 2023, matching the first identifier of the first sensor and the second identifier of the second sensor in the historical fusion information to obtain a matching result;

[0120] Step 2024: Update the historical fusion information based on the matching result and the current calculation result to obtain the target fusion information.

[0121] If the detection data difference meets the preset calculation threshold, that is, the lateral distance difference is less than or equal to the lateral distance threshold, and the longitudinal distance difference is less than or equal to the longitudinal distance threshold, and the vehicle speed difference is less than or equal to the speed threshold, the vehicle control device obtains the current calculation result based on the first detection data and the second detection data.

[0122] The current calculation result in this embodiment can be understood as the Mahalanobis distance. The Mahalanobis distance is a measure of distance and can be regarded as a correction of the Euclidean distance. It corrects the problem that the scales of various dimensions in the Euclidean distance are inconsistent and related. Therefore, it can be understood that the Mahalanobis distance is calculated according to the first detection data and the second detection data to obtain the current calculation result. The Mahalanobis distance formula for a single data point is as follows:

[0123]

[0124] The Mahalanobis distance commonly used in the fusion module mainly considers three measurement values: the lateral distance x, the longitudinal distance y, and the x-direction speed vx. The x-direction in this embodiment refers to the direction perpendicular to the rear axle of the vehicle. Therefore, the Mahalanobis distance formula is:

[0125]

[0126] Among them, Σ is the covariance matrix of the multidimensional random variable, μ is the sample mean, if the covariance matrix is ​​a unit vector, that is, each dimension is independent and identically distributed, the Mahalanobis distance degenerates into the Euclidean distance.

[0127] The vehicle control device performs matching in the historical fusion information according to the first identifier of the first sensor and the second identifier of the second sensor to obtain a matching result. Specifically, the first identifier of the first sensor is used as the column identifier of the distance matrix, and the second identifier of the second sensor is used as the row identifier of the distance matrix, and matching is performed in the historical distance matrix, and the obtained historical calculation result is used as the matching result.

[0128] The vehicle control device updates the historical fusion information according to the matching results and the current calculation results to obtain the target fusion information, that is, updates the historical distance matrix according to the matching results and the current calculation results to obtain the updated target distance matrix, as described in steps 20241 to 20244.

[0129] In the process of calculating new target fusion information, the embodiment of the present application applies the historical calculation results of historical fusion information, so that current data can be analyzed through historical data, avoiding misjudgment between different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, and improving the accuracy of target fusion information, thereby improving the accuracy of vehicle control.

[0130] In one embodiment, the description of step 20241 to step 20242 includes:

[0131] Step 20241, if the matching result is that there is a historical calculation result corresponding to the first identifier and the second identifier in the historical fusion information, then calculation is performed based on the historical calculation result and the current calculation result and their corresponding weight values ​​to obtain a first final result;

[0132] Step 20242, based on the first final result, the historical calculation results in the historical fusion information are updated to obtain target fusion information.

[0133] If the matching result is a historical calculation result corresponding to the first identifier and the second identifier in the historical fusion information, the vehicle control device calculates according to the historical calculation result and the current calculation result and their corresponding weight values ​​to obtain a first final result. Specifically, the historical calculation result and the current calculation result are respectively multiplied by their corresponding weight values, and the products of the two are summed to obtain the first final result.

[0134] The specific calculation formula of the first final result is: D1 = Dist × (1-gain) + Dist_old × gain. Wherein, D1 represents the first final result, Dist represents the current calculation result, Dist_old represents the historical calculation result, 1-gain represents the weight value of the current calculation result, and gain represents the weight value of the historical calculation result. For example, in this embodiment, gain can be set to 0.5.

[0135] The vehicle control device updates the historical calculation results in the historical fusion information according to the first final result, that is, replaces the historical calculation results with the first final result to obtain the target fusion information.

[0136] In the process of calculating new target fusion information, the embodiment of the present application applies the historical calculation results of historical fusion information, so that current data can be analyzed through historical data, avoiding misjudgment between different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, and improving the accuracy of target fusion information, thereby improving the accuracy of vehicle control.

[0137] In one embodiment, the description of step 20243 to step 20244 includes:

[0138] Step 20243, if the matching result is that there is no historical calculation result corresponding to the first identifier and the second identifier in the historical fusion information, then a calculation is performed based on the current calculation result and the preset threshold result and its corresponding weight value to obtain a second final result;

[0139] Step 20244, based on the second final result, the historical calculation results in the historical fusion information are updated to obtain target fusion information.

[0140] If the matching result is that the historical calculation result corresponding to the first identifier and the second identifier does not exist in the historical fusion information, the vehicle control device obtains a preset threshold result.

[0141] Furthermore, the vehicle control device multiplies the current calculation result by the preset threshold result and its corresponding weight value, and then sums the products of the two to obtain a second final result.

[0142] The specific calculation formula for the second final result is: D2=Dist×n+Gate×m, where D2 represents the second final result, Dist represents the current calculation result, Gate represents the threshold result, n represents the weight value of the current calculation result, and m represents the weight value of the threshold result. For example, in this embodiment, n can be set to 1 and m can be set to 0.9.

[0143] The vehicle control device updates the historical calculation results in the historical fusion information according to the second final result, that is, replaces the historical calculation results with the second final result to obtain the target fusion information.

[0144] In the process of calculating new target fusion information, the embodiment of the present application applies the historical calculation results of historical fusion information, so that current data can be analyzed through historical data, avoiding misjudgment between different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, and improving the accuracy of target fusion information, thereby improving the accuracy of vehicle control.

[0145] Reference Figure 5 , Figure 5 It is a flow chart of the matching fusion solution provided in one embodiment of the present application.

[0146] (1) Input: The detection data of this frame obtained from a certain perception sensor and the vehicle status data obtained from the vehicle-mounted sensor.

[0147] (2) Preprocessing: Filter the targets detected by the sensor and filter out invalid and other detection data that does not need to be stored.

[0148] (3) Iterative maintenance of previous frame detection data: The detection data format is to assign an ID to each target, and include its position information, speed information, and acceleration information. The sensor will track each target internally to ensure that the same target in the previous and next frames has the same ID information in most cases. Therefore, the previous frame information queue of the sensor is cleared and all observation information of this frame is stored in the information queue.

[0149] (4) Get the rows and columns of the distance matrix calculation: Figure 2 and Figure 3 In the Track list, each Track inside it is the storage medium after the target is fused. Track has three main fusion results: [camera id, -1], [-1, radar id], [camera id, radar id]. That is, Track composed of a single camera, Track composed of a single radar, and Track composed of a camera and radar. The Track list is used as the row of the distance matrix calculation, and the target list of the input single sensor is used as the column to calculate the distance matrix. The initial value of the distance matrix is ​​the threshold value gate. Figure 4 The specific format of the distance matrix. The target information involved in the calculation in each column is the single-frame target information input by the sensor. The target information involved in the calculation in each row is the current frame information of a sensor that constitutes the current Track.

[0150] (5) Distance matrix calculation: After the row and column information involved in the calculation is determined, the distance matrix will be calculated according to the following function. Except for the part in (4), it does not participate in the calculation and is directly assigned a threshold value Gate. In addition, if the difference in each dimension of the two target information involved in the calculation exceeds 12m in the x direction, exceeds 2m in the y direction, and the speed vx exceeds 12m / s, it does not participate in the calculation and is directly assigned a threshold value Gate. The rest of the distance matrix is ​​solved according to the Mahalanobis distance calculation formula.

[0151] Dist is the Mahalanobis distance between two points, taking into account their position x, position y and longitudinal velocity vx. Therefore, according to the ID of Track and the ID of the input sensor target, the last calculated result Dist_old is searched in the historical distance matrix. The final result D is calculated according to the following formula, where gain is the historical influence factor, which can be dynamically adjusted or set to a fixed empirical value.

[0152]

[0153] Reference Figure 6 and Figure 7 , Figure 6 is a schematic diagram of a horizontal scene provided in an embodiment of the present application, Figure 7 It is a schematic diagram of a vertical scene provided in an embodiment of the present application.

[0154] After verification, the embodiment of the present application can solve various mis-fusion scenarios that occur in the single-frame matching solution. The mis-fusion problem in the real car mainly includes the following two scenarios: horizontal scenario, where a pedestrian crossing horizontally is mis-fused with stationary obstacles such as fences and vehicles on the roadside; vertical scenario, where a non-motorized vehicle traveling vertically is mis-fused with stationary obstacles such as fences and vehicles on the roadside. Figure 6 and Figure 7 shown.

[0155] by Figure 6 and Figure 7 Taking the typical scenario in Figure 2 as an example, the different sensor IDs that have misfused are (1) ID9 and ID7382, and (2) ID23 and ID936. Among them, the ID detected by the radar is less than 100, and the ID detected by the camera is more than 100. Figure 6 and Figure 7 Shown in the figure are the camera target enclosed by the red dotted frame and a radar target below it, which is located at the position of the green triangle after misfusion.

[0156] Reference Figure 8 , Figure 8 It is a schematic diagram comparing the matching results of the scenes provided in an embodiment of the present application. By comparing the results of the single-frame matching scheme in a typical scene and the embodiment of the present application, it can be seen that the embodiment of the present application solves the above-mentioned horizontal scene, longitudinal scene and the scene of the vehicle starting at rest after waiting for a red light. The two detection targets that are mismatched in the single-frame matching scheme are separated into two tracks. Therefore, the solution proposed in the embodiment of the present application can solve the problem of misfusion that currently occurs in real vehicles. After verification, for the two major categories of mismatches that occurred in real vehicles in the scene library, a total of 40 specific mismatching scenarios, the pass rate of the embodiment of the present application is 95%. Therefore, the embodiment of the present application can solve the problem of misfusion of sensor targets in a variety of typical scenarios, and can complete the two tasks of mismatching the two misfused targets and decomposing them into two targets after matching. After verification, the embodiment of the present application has application value, can output more superior results to the downstream planning and control module, and has a certain improvement in the performance of the entire automatic driving.

[0157] The vehicle control device provided by the embodiment of the present application is described below. The vehicle control device described below and the vehicle control method described above can be referred to each other. Fig. 9 As shown, Fig. 9 is a schematic diagram of the structure of a vehicle control device provided in an embodiment of the present application, and the vehicle control device includes.

[0158] An acquisition module 901 is used to acquire detection data of at least two sensors;

[0159] A determination module 902, configured to obtain target fusion information based on the detection data and the historical fusion information;

[0160] The vehicle control module 903 is used to control the vehicle based on the target fusion information.

[0161] The vehicle control device provided in the embodiment of the present application obtains detection data of at least two sensors; obtains target fusion information based on the detection data and historical fusion information; and controls the vehicle based on the target fusion information. Therefore, in the process of calculating new target fusion information, the historical calculation results of the historical fusion information are applied, so that current data can be analyzed through historical data, avoiding misjudgment of different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, improving the accuracy of target fusion information, and thus improving the accuracy of vehicle control.

[0162] Reference Fig.10 , Fig.10 1 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: a processor 1010, a communication interface 1020, a memory 1030 and a communication bus 1040, wherein the processor 1010, the communication interface 1020, and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 may call a computer program in the memory 1030 to execute the steps of the vehicle control method, including:

[0163] Acquire detection data from at least two sensors; obtain target fusion information based on the detection data and historical fusion information; control the vehicle based on the target fusion information, so that in the process of calculating new target fusion information, the historical calculation results of the historical fusion information are applied, so that current data can be analyzed through historical data, avoiding misjudgment of different targets by different sensors, solving the problem of misfusion of targets of different sensors during driving, improving the accuracy of target fusion information, and thus improving the accuracy of vehicle control.

[0164] In addition, the logical computer program in the above-mentioned memory 1030 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several computer programs to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0165] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0166] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including several computer programs to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0167] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A vehicle control method, characterized in that: include: Acquire detection data of at least two sensors; Based on the detection data and the historical fusion information, obtaining target fusion information; The vehicle is controlled based on the target fusion information.

2. The vehicle control method according to claim 1, characterized in that: The at least two sensors include a first sensor and a second sensor; The obtaining of target fusion information based on the detection data and the historical fusion information includes: Performing difference calculation based on first detection data of the first sensor and second detection data of the second sensor to obtain a detection data difference; The historical fusion information is updated based on the detection data difference to obtain the target fusion information.

3. The vehicle control method according to claim 2, characterized in that: The detection data includes at least one of a lateral distance of the vehicle in a lateral direction, a longitudinal distance of the vehicle in a longitudinal direction, and a vehicle speed of the vehicle.

4. The vehicle control method according to claim 3, characterized in that: The performing difference calculation based on the first detection data of the first sensor and the second detection data of the second sensor to obtain the detection data difference includes: Based on at least one of the lateral distance, longitudinal distance and vehicle speed of the first sensor, a difference calculation is performed with at least one of the lateral distance, longitudinal distance and vehicle speed corresponding to the second sensor to obtain the detection data difference.

5. The vehicle control method according to claim 4, characterized in that: The detection data difference includes at least one of a lateral distance difference, a longitudinal distance difference and a vehicle speed difference.

6. The vehicle control method according to claim 5, characterized in that: The method further comprises: If at least one of the lateral distance difference, the longitudinal distance difference and the vehicle speed difference is greater than a corresponding preset threshold, determining that the detection data difference does not meet a preset calculation threshold; If the lateral distance difference, the longitudinal distance difference and the vehicle speed difference are all less than or equal to corresponding preset thresholds, it is determined that the detection data difference meets the calculation threshold.

7. The vehicle control method according to claim 2, characterized in that: The updating of the historical fusion information based on the detection data difference to obtain the target fusion information includes: If the detection data difference does not meet the preset calculation threshold, the historical fusion information is updated based on the first identifier of the first sensor, the second identifier of the second sensor and a preset threshold result to obtain the target fusion information.

8. The vehicle control method according to claim 7, characterized in that: The updating of the historical fusion information based on the first identifier of the first sensor, the second identifier of the second sensor and a preset threshold result to obtain the target fusion information includes: The threshold result is used to replace the historical calculation result in which the column is marked with the first identifier and the row is marked with the second identifier in the historical fusion information to obtain the target fusion information.

9. The vehicle control method according to claim 2, characterized in that: The updating of the historical fusion information based on the detection data difference to obtain the target fusion information includes: If the detection data difference satisfies a preset calculation threshold, a Mahalanobis distance calculation is performed based on the first detection data and the second detection data to obtain a current calculation result; According to the first identifier of the first sensor and the second identifier of the second sensor, matching is performed in the historical fusion information to obtain a matching result; The historical fusion information is updated based on the matching result and the current calculation result to obtain target fusion information.

10. The vehicle control method according to claim 9, characterized in that: The performing matching in the historical fusion information according to the first identifier of the first sensor and the second identifier of the second sensor to obtain a matching result includes: The first identifier is used as a column identifier, and the second identifier is used as a row identifier, matching is performed in the historical fusion information, and the obtained historical calculation result is used as the matching result.

11. The vehicle control method according to claim 9, characterized in that: The updating of the historical fusion information based on the matching result and the current calculation result to obtain target fusion information includes: If the matching result is a historical calculation result corresponding to the first identifier and the second identifier in the historical fusion information, a calculation is performed based on the historical calculation result and the current calculation result and their corresponding weight values ​​to obtain a first final result; The historical calculation result in the historical fusion information is updated based on the first final result to obtain the target fusion information.

12. The vehicle control method according to claim 11, characterized in that: The specific calculation formula of the first final result is: D1=Dist×(1-gain)+Dist_old×gain; Among them, D1 represents the first final result, Dist represents the current calculation result, Dist_old represents the historical calculation result, 1-gain represents the weight value of the current calculation result, and gain represents the weight value of the historical calculation result.

13. The vehicle control method according to claim 9, characterized in that: The updating of the historical fusion information based on the matching result and the current calculation result to obtain target fusion information includes: If the matching result is that the historical calculation result corresponding to the first identifier and the second identifier does not exist in the historical fusion information, then calculating based on the current calculation result and the preset threshold result and its corresponding weight value to obtain a second final result; The historical calculation result in the historical fusion information is updated based on the second final result to obtain the target fusion information.

14. The vehicle control method according to claim 13, characterized in that: The specific calculation formula of the second final result is: D2 = Dist × n + Gate × m; Among them, D2 represents the second final result, Dist represents the current calculation result, Gate represents the threshold result, n represents the weight value of the current calculation result, and m represents the weight value of the threshold result.

15. The vehicle control method according to any one of claims 2 to 14, characterized in that: The acquiring detection data of at least two sensors includes: Acquire first detection data of the first sensor; Based on the first identification of the first sensor, matching is performed in the detection data list, and detection data in the detection data list with a second identification identical to the first identification is used as second detection data of the second sensor.

16. The vehicle control method according to claim 15, characterized in that: The sensors include cameras and radars.

17. The vehicle control method according to claim 16, characterized in that: The detection data list includes a camera detection data list and / or a radar detection data list.

18. The vehicle control method according to claim 17, characterized in that: The matching in the detection data list based on the first identifier of the first sensor and taking the detection data in the detection data list having the same second identifier as the first identifier as the second detection data of the second sensor includes: Based on the first identifier, matching is performed in the camera detection data list and / or the radar detection data list, and detection data in the camera detection data list and / or the radar detection data list whose second identifier is the same as the first identifier is used as the second detection data of the second sensor.

19. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores a plurality of computer programs; the processor loads the computer program from the memory to execute the vehicle control method according to any one of claims 1 to 18.

20. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program executable by a vehicle, and when the computer program is run on the vehicle, the vehicle is caused to execute the vehicle control method according to any one of claims 1 to 18.

21. A computer program product, characterized in that The computer program product comprises a computer program, and the computer program is executed by a processor to execute the vehicle control method according to any one of claims 1 to 18.

22. A vehicle, characterized in that: The vehicle includes the electronic device of claim 19.

Images