False detection of brakes, methods, devices, electronic equipment and autonomous vehicles

Abstract

The present disclosure provides a false detection braking detection method, device, electronic device and autonomous driving vehicle, which relate to the field of artificial intelligence technology, and in particular to the field of autonomous driving and intelligent transportation technology. The specific implementation scheme is: from the obstacles detected during the vehicle's driving, screen out suspicious obstacles that have a risk of collision with the vehicle, and screen out false detection obstacles that have the possibility of false detection from each suspicious obstacle. Then, the detection time of each false detection obstacle is obtained respectively to obtain a detection time set. The braking moment of the vehicle is obtained, and by comparing the braking moment with the detection moment set, it is determined whether the braking reason at the braking moment is a false detection of an obstacle. Thereby, it is realized to detect whether the braking of the vehicle is caused by the false detection of an obstacle.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to the field of autonomous driving and intelligent transportation technology. Background Art

[0002] During the autonomous driving process, when the autonomous driving vehicle is close to surrounding obstacles, the autonomous driving vehicle needs to slow down or brake to ensure the safety of the autonomous driving vehicle and the safety of road traffic. Summary of the invention

[0003] The present invention provides a false brake detection method, device, electronic equipment and an autonomous driving vehicle.

[0004] According to a first aspect of an embodiment of the present disclosure, a method for detecting a misdetected brake is provided, comprising:

[0005] Screening out suspicious obstacles with a risk of collision with the vehicle from an obstacle set; the obstacle set is obstacles detected during the vehicle's driving process;

[0006] Screen out misdetected obstacles with the possibility of misdetection from various suspicious obstacles;

[0007] The detection time of each misdetected obstacle is obtained respectively to obtain a detection time set;

[0008] The braking moment of the vehicle is obtained, and by comparing the braking moment with the detection moment set, it is determined whether the braking cause at the braking moment is a false detection of an obstacle.

[0009] According to a second aspect of the embodiments of the present disclosure, there is provided a misdetected brake detection device, comprising:

[0010] A screening module is used to screen out suspicious obstacles with a risk of collision with the vehicle from a set of obstacles; the obstacle set is obstacles detected during the driving of the vehicle;

[0011] The screening module is further used to screen out misdetected obstacles with a possibility of misdetection from various suspicious obstacles;

[0012] An acquisition module, used to respectively acquire the detection time of each misdetected obstacle screened by the screening module to obtain a detection time set;

[0013] The determination module is used to obtain the braking moment of the vehicle and determine whether the braking cause at the braking moment is a false detection of an obstacle by comparing the braking moment with the detection moment set obtained by the acquisition module.

[0014] According to a third aspect of the embodiments of the present disclosure, there is provided an electronic device, including:

[0015] at least one processor; and

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform any method according to the first aspect.

[0018] According to a fourth aspect of an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute a method according to any one of the first aspects.

[0019] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the method according to any one of the first aspects is implemented.

[0020] A sixth aspect of an embodiment of the present disclosure provides an autonomous driving vehicle, comprising the electronic device as described in the third aspect.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0023] Figure 1 is a flow chart of a misdetected brake detection method provided by an embodiment of the present disclosure;

[0024] Figure 2 is a flow chart of a method for determining a suspicious obstacle provided by an embodiment of the present disclosure;

[0025] Figure 3 is a first exemplary schematic diagram of the positional relationship between a vehicle and an obstacle provided in an embodiment of the present disclosure;

[0026] Figure 4 is a flow chart of a method for screening misdetected obstacles provided by an embodiment of the present disclosure;

[0027] Figure 5 is a second exemplary schematic diagram of the positional relationship between a vehicle and an obstacle provided in an embodiment of the present disclosure;

[0028] Figure 6 is a third exemplary schematic diagram of the positional relationship between a vehicle and an obstacle provided in an embodiment of the present disclosure;

[0029] Figure 7 is a fourth exemplary schematic diagram of the positional relationship between a vehicle and an obstacle provided in an embodiment of the present disclosure;

[0030] Figure 8 is an exemplary schematic diagram of a misdetected brake detection process provided by an embodiment of the present disclosure;

[0031] Fig. 9 is a structural schematic diagram of a misdetected brake detection device provided by an embodiment of the present disclosure;

[0032] Fig.10 4 is a block diagram of an electronic device used to implement the false detection brake detection method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0033] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0034] During the autonomous driving process, autonomous driving vehicles sometimes mistakenly believe that objects such as cracks on the ground or fallen leaves on the ground will hinder the vehicle's progress, and thus take emergency braking measures, which seriously affects the user's riding experience and also affects the driving safety of road vehicles.

[0035] In order to detect whether the braking of a vehicle is caused by a false detection of an obstacle, an embodiment of the present disclosure provides a false detection braking detection method, which can be applied to an electronic device, such as an onboard computer in an autonomous vehicle, or a server connected to the autonomous vehicle. Figure 1 As shown, the method comprises the following steps:

[0036] S101. Filter out suspicious obstacles that may collide with the vehicle from a set of obstacles.

[0037] The obstacle set is obstacles detected during vehicle driving.

[0038] In the disclosed embodiment, the electronic device can predict suspicious obstacles that may collide with the vehicle in the future, i.e., have a risk of collision with the vehicle, in real time based on the detected obstacle information during the vehicle's driving process. Alternatively, after obtaining all obstacles detected during the complete driving process of the vehicle, the electronic device can predict suspicious obstacles that may collide with the vehicle at that time based on the obstacle information offline.

[0039] The obstacle information includes: the location of the obstacle, the detection time, the identity document (ID), the type, speed, size and heading angle, etc.

[0040] S102: Filter out misdetected obstacles with a possibility of misdetection from among the suspicious obstacles.

[0041] Falsely detected obstacles are obstacles that do not actually affect the vehicle's driving. For example, falsely detected obstacles can be static obstacles such as ground cracks, leaves and pebbles, and dynamic obstacles such as fallen leaves and catkins in the air.

[0042] S103, respectively obtain the detection time of each falsely detected obstacle to obtain a detection time set, wherein the detection time of the obstacle is the initial time when the obstacle is detected.

[0043] S104: Obtain the braking moment of the vehicle, and determine whether the braking reason at the braking moment is a false detection of an obstacle by comparing the braking moment with the detection moment set.

[0044] Among them, braking can also be called emergency braking.

[0045] It is understandable that if the braking moment is close to a detection moment in the detection moment set, then the braking is likely to be caused by a false detection of an obstacle.

[0046] In the disclosed embodiment, misdetected obstacles with the possibility of misdetection are screened out from the suspicious obstacles with the risk of collision with the vehicle, and the detection time set of each misdetected obstacle is obtained. Then, by comparing the braking time of the vehicle with the detection time set, it is determined whether the braking reason at the braking time is the misdetection of an obstacle. In this way, it is possible to detect whether the braking of the vehicle is caused by the misdetection of an obstacle.

[0047] This enables subsequent analysis of misdetected obstacles, thereby reducing the impact of misdetected obstacles on the driving of the autonomous driving vehicle.

[0048] The following is a detailed description of the false detection braking detection method provided by the embodiment of the present disclosure:

[0049] The method of screening suspicious obstacles that may have a collision risk with the vehicle in S101 includes the following steps:

[0050] Step 1: From the obstacle set, select obstacles that are located in the vehicle's forward direction, are less than a preset distance from the vehicle, and whose shortest distance from the vehicle's predicted trajectory planning line is less than a specified safety threshold.

[0051] In the disclosed embodiment, for obstacles detected within a specified range around the vehicle, the obstacles can be projected onto the vehicle body coordinate system first, and according to the vehicle's driving direction, the obstacles located in the vehicle's forward direction can be filtered out, that is, obstacles in the opposite direction of the driving direction can be filtered out. For example, if the vehicle is driving forward, obstacles in the rear direction of the vehicle are filtered out, and if the vehicle is reversing, obstacles in the front direction of the vehicle are filtered out. Since obstacles in the opposite direction of the vehicle's forward direction will not collide with the vehicle, the disclosed embodiment filters obstacles in the vehicle's forward direction.

[0052] For each obstacle after filtering, the obstacles whose straight-line distance from the vehicle at the time of detection is less than a preset distance are screened. Since there is generally no risk of collision when the straight-line distance from the vehicle is too far, the embodiment of the present disclosure screens obstacles whose distance from the vehicle is less than a preset distance.

[0053] The shortest distance between the selected obstacles and the vehicle's predicted trajectory planning line is compared, where the predicted trajectory planning line is the future driving trajectory predicted based on the vehicle's current driving speed. Since there is generally no collision risk when the distance from the vehicle's future driving trajectory is too far, the embodiment of the present disclosure selects obstacles whose shortest distance from the predicted trajectory planning line is less than a specified safety threshold.

[0054] When comparing, the position of the obstacle at the time of detection can be set as P obs , the real-time predicted trajectory planning line of the vehicle is composed of multiple predicted trajectory points, that is, the predicted trajectory planning line is Each element of S is a predicted trajectory point, for example, is the first predicted trajectory point, and N is the number of predicted trajectory points. You can select the predicted trajectory segment closest to the obstacle detection moment, and the predicted trajectory segment is the line connecting the adjacent predicted trajectory points. i=1,2,…,N+1. And calculate the vertical distance between the obstacle detection moment and the predicted trajectory segment. The vertical distance is less than the specified safety threshold L s Obstacles. Among them, L s It can be determined by formula (1):

[0055]

[0056] Among them, l 车宽 is the width of the vehicle, l 障碍物宽度 is the obstacle width, l 阈值 is the preset distance threshold.

[0057] Step 2: Among the screened obstacles, predict the suspicious obstacles that still pose a risk of collision even if the vehicle starts braking from the moment of detection based on the vehicle's speed.

[0058] According to the vehicle speed at the time of detection and the moving speed of the obstacle, obstacles that can still be caught up by the vehicle after the vehicle starts braking from the time the obstacle is detected can be determined as suspicious obstacles.

[0059] Through the above method, the disclosed embodiment can screen out suspicious obstacles that have a risk of collision with the vehicle from obstacles around the vehicle. Since the risk of collision may cause the vehicle to brake, detecting suspicious obstacles can facilitate subsequent judgment of whether the cause of braking is related to the suspicious obstacles.

[0060] See also Figure 2 In the above step 2, the method of predicting a suspicious obstacle that still poses a risk of collision when the vehicle brakes from the moment of detection includes the following steps:

[0061] S201. Obtain obstacles whose moving speed in the forward direction of the vehicle is less than the vehicle speed from the screened obstacles.

[0062] It is understandable that if the moving speed of the obstacle in the vehicle's forward direction is greater than or equal to the vehicle's speed, since the selected obstacle is located in the vehicle's forward direction, the vehicle cannot catch up with the obstacle, so there is no risk of collision. Therefore, the disclosed embodiment obtains obstacles whose moving speed in the direction is less than the vehicle's speed in the direction from the selected obstacles for the vehicle's forward direction, that is, obstacles that may be caught up by the vehicle, for subsequent judgment.

[0063] S202: For each obstacle obtained, predict the moment when the vehicle starts braking from the moment of detection and the position of the vehicle in the forward direction is the same as the position of the obstacle, as the predicted collision moment.

[0064] like Figure 3 As shown in the figure, assuming that the vehicle's forward direction is the y direction and the direction perpendicular to the forward direction is the x direction, in the vehicle's body coordinate system, at the moment of obstacle detection, the position of the obstacle is (x obs ,y obs ), the moving speed of the obstacle is (v x ,v y ), the vehicle's moving speed in the y direction is v c , the predicted collision time is solved by formula (2):

[0065]

[0066] Among them, a is the preset maximum acceleration of the vehicle during braking, and t is the predicted collision time between the obstacle and the vehicle.

[0067] If there is no positive real number solution to formula (2), it is determined that there is no collision risk between the obstacle and the vehicle, that is, the obstacle is a non-suspicious obstacle; otherwise, the minimum positive real number solution of formula (2) is used as the predicted collision time between the obstacle and the vehicle.

[0068] S203: Determine the predicted position of the obstacle at the predicted collision moment according to the moving speed of the obstacle.

[0069] In the embodiment of the present disclosure, it can be assumed that the obstacle has been moving at the speed at the time of detection, and the predicted position is (x obs +t×v x ,y obs +t×v y ).

[0070] S204: Determine whether the shortest distance between the predicted position and the vehicle predicted trajectory planning line is less than a specified safety threshold. If so, execute S205; if not, execute S206.

[0071] The determination method of S204 is the same as that of the above step 1, and reference may be made to the above description, which will not be repeated here.

[0072] S205: Determine that the obstacle is a suspicious obstacle.

[0073] S206: Determine that the obstacle is a non-suspicious obstacle.

[0074] Through the above method, the disclosed embodiment can determine obstacles that still have a collision risk when the vehicle starts braking from the moment the obstacle is detected, based on the vehicle speed and the moving speed of the obstacle. Since these obstacles may cause the vehicle to brake, screening these obstacles can facilitate subsequent judgment of whether the braking reason is related to these obstacles.

[0075] In the disclosed embodiment, since it is necessary to analyze whether the braking is caused by the false detection of obstacles, after screening out suspicious obstacles, it is necessary to further screen out the falsely detected obstacles.

[0076] See also Figure 4 The method of screening misdetected obstacles in the above S102 includes the following steps:

[0077] S401: for each suspicious obstacle, determine whether the suspicious obstacle is a dynamic obstacle. If the suspicious obstacle is a dynamic obstacle, execute S402; if the suspicious obstacle is a static obstacle, execute S404.

[0078] When the positions of the obstacles are different at different times, the obstacles can be determined to be dynamic obstacles; on the contrary, when the positions of the obstacles are the same at different times, the obstacles can be determined to be static obstacles. Alternatively, the dynamic and static states of the obstacles can be determined by other methods, which are not specifically limited in the embodiments of the present disclosure.

[0079] S402: Determine whether the detection state of the suspicious obstacle is stable detection according to the time when the suspicious obstacle is detected.

[0080] It can be understood that if a suspicious obstacle can be detected stably and continuously, the detection state of the suspicious obstacle can be determined as stable detection; conversely, if the suspicious obstacle cannot be detected stably and continuously, that is, the duration of detecting the suspicious obstacle is short, for example, only one frame of a suspicious obstacle is detected, then the detection state of the suspicious obstacle can be determined as unstable detection.

[0081] S403: Treat the non-stably detected dynamic obstacle as a falsely detected obstacle.

[0082] It is understandable that the non-stably detected dynamic obstacles are most likely mis-detected obstacles, such as obstacles that do not actually exist, and therefore should not be regarded as suspicious obstacles that may cause the vehicle to brake. Therefore, the non-stably detected dynamic obstacles are regarded as mis-detected obstacles.

[0083] S404: Determine whether the suspicious obstacle is run over by the vehicle based on the historical driving trajectory of the vehicle, wherein the historical driving trajectory is the actual driving trajectory of the vehicle.

[0084] When the historical driving track of the vehicle and the position of the suspicious obstacle have an intersection, it can be determined that the suspicious obstacle is run over by the vehicle. Conversely, when the historical driving track of the vehicle and the position of the suspicious obstacle do not have an intersection, it can be determined that the suspicious obstacle is not run over by the vehicle.

[0085] S405: The static obstacle crushed by the vehicle is regarded as a misdetected obstacle.

[0086] Since a static obstacle can be crushed by a vehicle, it means that the static obstacle does not actually affect the driving of the vehicle and should not be regarded as a suspicious obstacle that may cause the vehicle to brake. Therefore, a static obstacle crushed by a vehicle is regarded as a falsely detected obstacle.

[0087] Through the above method, the embodiment of the present disclosure can treat the non-stably detected dynamic obstacles and the static obstacles crushed by the vehicle as falsely detected obstacles. Since the non-stably detected dynamic obstacles and the static obstacles crushed by the vehicle will not affect the normal driving of the vehicle with a high probability, they should not be identified as suspicious obstacles. Therefore, these obstacles are treated as falsely detected obstacles, which improves the accuracy of determining falsely detected obstacles.

[0088] In the embodiment of the present disclosure, the above-mentioned S402 method for determining the detection status of a suspicious obstacle can be implemented as follows: starting from the moment the suspicious obstacle is detected, determining whether the cumulative detection time of the suspicious obstacle within a preset time period exceeds the preset time period, and whether the suspicious obstacle is detected at the specified time.

[0089] If the judgment results are all yes, it is determined that the detection state of the suspicious obstacle is stable detection.

[0090] If any of the judgment results is no, it is determined that the detection state of the suspicious obstacle is unstable detection.

[0091] In the disclosed embodiment, the detection time of the suspicious obstacle can be set as t, the preset time period can be t to t+T, the preset duration can be T / 2, and the specified time can be t+T. T is a preset period of time, for example, T is 5 seconds.

[0092] Since the cumulative detection time does not exceed the preset time, and the obstacles that are not detected at the specified time are obstacles that appear for a short time, these obstacles are most likely caused by false detections and may not actually exist or do not affect the vehicle's driving. Therefore, these dynamic obstacles can be regarded as falsely detected obstacles, thereby improving the accuracy of detecting falsely detected obstacles.

[0093] In the embodiment of the present disclosure, the method of determining whether the static obstacle is crushed by the vehicle in the above S404 includes the following steps:

[0094] Step 1: Determine whether the shortest distance between the center point of the suspected obstacle and the historical driving trajectory of the vehicle is less than or equal to the specified distance. If yes, execute step 3; if no, execute step 2.

[0095] like Figure 3 As shown in the figure, assuming that in the vehicle body coordinate system, the vehicle's forward direction is y, and the position of the obstacle center is (x obs ,y obs ).like and|x obs |≤L s , then execute step 3, otherwise execute step 2. s To specify the safety threshold, the calculation method can refer to formula (1), l 阈值 is the preset length threshold, l 车长 is the vehicle length.

[0096] Step 2: According to the historical driving trajectory of the vehicle, determine the boundary position of the vehicle when the vehicle is closest to the suspicious obstacle, and determine whether the boundary position of the suspicious obstacle and the boundary position of the vehicle intersect. If so, execute step 3; if not, execute step 4.

[0097] In the disclosed embodiment, when the length, width and heading angle of the obstacle are detected, it can be determined that the obstacle is a convex hull. The position of the obstacle's bounding box is determined based on the length, width and heading angle of the obstacle when the vehicle is closest to the suspected obstacle. The position of the vehicle's bounding box is determined based on the length, width and heading angle of the vehicle when the vehicle is closest to the suspected obstacle. Based on the positions of the two bounding boxes, it is determined whether the two bounding boxes intersect, wherein the bounding box can be a rectangle or other shapes.

[0098] Figure 5 and Figure 6 The rectangle where the vehicle is located is the vehicle calibration box, and the hexagon where the obstacle is located is the obstacle calibration box. When judging, the straddle algorithm can be used to determine whether any side of a calibration box intersects with any side of another calibration box. If so, it is determined that the two calibration boxes intersect, that is, the boundary position of the suspected obstacle and the boundary position of the vehicle have an intersection. For example, the position relationship between the vehicle calibration box and the obstacle calibration box is as follows: Figure 5 If not, see Figure 6 , it may be that the obstacle calibration box is inside the vehicle calibration box, then we can determine whether the number of intersections between any end point of the obstacle calibration box and the ray drawn in any direction and the vehicle calibration box is an odd number; if so, the obstacle calibration box is inside the vehicle calibration box, that is, the boundary position of the obstacle and the boundary position of the vehicle have an intersection. On the contrary, the obstacle calibration box is not inside the vehicle calibration box, that is, the boundary position of the obstacle and the boundary position of the vehicle do not have an intersection.

[0099] If the length, width and heading angle of the obstacle are not detected, the obstacle is determined to be a concave hull, for example, the obstacle shape is Figure 7 As shown, Figure 7 The rectangle where the vehicle is located is the vehicle calibration box. At this time, when any boundary point of the obstacle is inside the vehicle calibration box, it is determined that the boundary position of the obstacle and the boundary position of the vehicle have an intersection; otherwise, it is determined that the boundary position of the obstacle and the boundary position of the vehicle do not have an intersection.

[0100] Step 3: Determine whether the suspicious obstacle is run over by the vehicle.

[0101] Step 4: Determine whether the suspected obstacle is not run over by the vehicle.

[0102] Through the above method, the disclosed embodiment can determine whether the vehicle has run over the suspicious obstacle based on the distance between the static suspicious obstacle and the historical driving trajectory of the vehicle. Since the static obstacle is run over by the vehicle, it means that the obstacle does not affect the driving of the vehicle and should not be identified as a suspicious obstacle that affects the driving of the vehicle. Therefore, these obstacles are regarded as falsely detected obstacles, which can improve the accuracy of determining falsely detected obstacles.

[0103] After determining the misdetected obstacle, it can be determined whether the actual braking of the vehicle is related to the misdetected obstacle. That is, the above S104 determines whether the braking reason at the braking moment is the misdetection of an obstacle, which can be implemented by: judging whether there is a target detection moment that meets the preset conditions in the detection moment set. The preset conditions include: the detection moment is less than or equal to the braking moment, and the difference between the braking moment and the detection moment is less than the preset time difference.

[0104] If it exists, it is determined that the braking reason at the braking moment is a false detection of an obstacle;

[0105] If not, it is determined that the braking reason at the braking moment is not a false detection of an obstacle. For example, the braking reason may be that there is indeed an obstacle in the vehicle's forward direction that affects the vehicle's travel, or the vehicle needs to wait for the red light to end when driving to the intersection.

[0106] In the embodiment of the present disclosure, the moment when the acceleration of the vehicle is less than the preset acceleration threshold can be used as the braking moment t brake , for example, the preset acceleration threshold is -2 or -5. Assume that the set of falsely detected obstacles is S0, and the set of detection times of each falsely detected obstacle is T c ={t1,t2,…,t m}, where m is the number of misdetected obstacles. i ∈T c , t i ≤t brake , and t brake -t i ≤t 阈值 , then determine the braking time t brake The reason for braking is the false detection of an obstacle. If there is no obstacle, determine the braking time t brake The reason for braking is not the false detection of obstacles. 阈值 is the preset time difference, for example, t 阈值 is 3 seconds.

[0107] Since the brakes may be applied to avoid collision with the misdetected obstacle only after the misdetected obstacle is detected, the target detection time should be less than or equal to the braking time. Moreover, when the braking time is close to the obstacle detection time, the braking reason may be related to the misdetected obstacle, so the difference between the braking time and the target detection time should be less than or equal to the preset time difference. Therefore, in the case of the target detection time, the embodiment of the present disclosure determines that the braking reason at the braking time is the misdetection of the obstacle, thereby improving the accuracy of determining the braking reason.

[0108] After determining that the reason for braking is the false detection of an obstacle, in order to reduce the occurrence of such braking events, the perception model used to detect obstacles can be optimized by using the false detection of obstacles associated with the braking moment. The specific method includes the following steps:

[0109] Step (1), obtaining obstacle information of the misdetected obstacle corresponding to the target detection moment.

[0110] The obstacle information includes the obstacle detection time and obstacle ID. Optionally, the obstacle information may also include: obstacle type, speed, size, heading angle, and calibration box position.

[0111] Optionally, the perceived false detection of emergency braking information may also be obtained, such as the vehicle ID and the acceleration of the vehicle at the moment of braking, so that subsequent analysts can perform manual analysis based on the perceived false detection of emergency braking information and obstacle information.

[0112] Step (2), obtaining a false detection label of the obstacle information, wherein the false detection label is used to indicate that the obstacle corresponding to the obstacle information is a false detection obstacle.

[0113] Step (3), using obstacle information and false positive labels, train a perception model. The perception model is used to detect obstacles around the vehicle. For example, the obstacles around the vehicle detected in the above step 1 may be non-false positive obstacles detected by the perception model.

[0114] During training, obstacle information can be input into the perception model to obtain the output of the perception model to determine whether the obstacle is a falsely detected obstacle. The output of the perception model is compared with the falsely detected label to obtain a loss value, which is used to adjust the model parameters of the perception model and iterate continuously until the perception model converges.

[0115] Through the above method, the embodiment of the present disclosure can train the perception model based on the falsely detected obstacles, so that the perception model can identify the falsely detected obstacles more and more accurately, thereby reducing the braking caused by the falsely detected obstacles and improving the driving safety and stability of the vehicle.

[0116] See also Figure 8The following describes the overall process of the false detection brake detection method provided by the embodiment of the present disclosure in combination with actual application scenarios:

[0117] Obtain vehicle-side information. The vehicle-side information includes: real-time location information, real-time perception reporting, real-time trajectory prediction and real-time vehicle condition. The real-time location information includes the current location of the obstacle, the real-time perception reporting includes the detection time of the obstacle, the real-time trajectory prediction includes the predicted trajectory planning line of the vehicle, and the real-time vehicle condition includes the speed and heading angle of the vehicle. The detection of obstacles can be obtained by the perception model.

[0118] When making a false detection emergency brake judgment, a false detection judgment is first performed, that is, irrelevant obstacles are first filtered out, that is, obstacles that are not in the direction of vehicle travel are filtered out, and then the risk of collision between the vehicle and the obstacle is calculated to obtain suspicious obstacles that have a risk of collision with the vehicle.

[0119] Then, among the suspicious obstacles, the dynamic obstacle detection stability is judged to obtain the unstable detected dynamic obstacles. The static obstacle crushing judgment is also performed to obtain the static obstacles crushed by the vehicle. After that, the detection time set of the unstable detected dynamic obstacles and the static obstacles crushed by the vehicle is obtained, and the obstacle detection time and emergency braking time are judged to determine whether the braking reason at the emergency braking moment is a false detection of an obstacle.

[0120] In the subsequent data usage process, when the braking reason is the false detection of obstacles, the perception false detection emergency braking scene analysis is performed, that is, the obstacle information of the false detection obstacle that caused the emergency braking is obtained. Then the obstacle information is labeled to obtain the false detection label of the obstacle information. Then the obstacle information and false detection label are used to train the perception model. Then the perception model obtained after training is deployed so that the vehicle can use the perception model obtained after training to detect obstacles around the vehicle.

[0121] The disclosed embodiment combines real-time collision risk prediction to obtain suspicious obstacles, and a posteriori stability detection judgment and crushing judgment to obtain falsely detected obstacles, thereby realizing false detection analysis of various types of obstacles, so as to accurately identify whether various falsely detected obstacles lead to emergency braking.

[0122] Based on the same inventive concept, corresponding to the above method embodiment, the embodiment of the present disclosure also provides a false detection brake detection device, such as Fig. 9 As shown, the device includes: a screening module 901, an acquisition module 902 and a determination module 903;

[0123] The screening module 901 is used to screen out suspicious obstacles with a risk of collision with the vehicle from the obstacle set; the obstacle set is obstacles detected during the vehicle's driving process;

[0124] The screening module 901 is also used to screen out misdetected obstacles with a possibility of misdetection from the suspicious obstacles;

[0125] The acquisition module 902 is used to respectively acquire the detection time of each misdetected obstacle screened by the screening module 901 to obtain a detection time set;

[0126] The determination module 903 is used to obtain the braking moment of the vehicle, and determine whether the braking reason at the braking moment is a false detection of an obstacle by comparing the braking moment with the detection moment set obtained by the acquisition module 902.

[0127] In some embodiments of the present disclosure, the screening module 901 is specifically used to:

[0128] For each suspicious obstacle, determine whether the suspicious obstacle is a dynamic obstacle;

[0129] If the suspected obstacle is a dynamic obstacle, determining whether the detection state of the suspected obstacle is stable detection according to the time when the suspected obstacle is detected;

[0130] Treat the non-stably detected dynamic obstacles as falsely detected obstacles;

[0131] If the suspected obstacle is a static obstacle, determine whether the suspected obstacle has been run over by the vehicle based on the vehicle's historical driving trajectory;

[0132] Static obstacles that are crushed by the vehicle are considered as falsely detected obstacles.

[0133] In some embodiments of the present disclosure, the screening module 901 is specifically used to:

[0134] Starting from the detection time of the suspicious obstacle, determine whether the cumulative detection time of the suspicious obstacle exceeds the preset time within the preset time period, and whether the suspicious obstacle is detected at the specified time;

[0135] If the judgment results are all yes, then the detection state of the suspicious obstacle is determined to be stable detection;

[0136] If any of the judgment results is no, it is determined that the detection state of the suspicious obstacle is unstable detection.

[0137] In some embodiments of the present disclosure, the screening module 901 is specifically used to:

[0138] Determine whether the shortest distance between the center point of the suspected obstacle and the historical driving trajectory of the vehicle is less than or equal to a specified distance;

[0139] If yes, it is determined that the suspected obstacle is run over by the vehicle;

[0140] If not, then determine the boundary position of the vehicle when the vehicle is closest to the suspected obstacle based on the historical driving trajectory of the vehicle, and determine whether the boundary position of the suspected obstacle intersects with the boundary position of the vehicle;

[0141] If it exists, it is determined that the suspicious obstacle is run over by the vehicle;

[0142] If not, it is determined that the suspicious obstacle has not been run over by the vehicle.

[0143] In some embodiments of the present disclosure, the screening module 901 is specifically used to:

[0144] From the obstacle set, select obstacles that are located in the vehicle's forward direction, are less than a preset distance from the vehicle, and whose shortest distance from the vehicle's predicted trajectory planning line is less than a specified safety threshold;

[0145] Among the screened obstacles, there are some suspected obstacles that may still pose a risk of collision, based on the vehicle's speed, and are predicted to have braking effects from the moment they are detected.

[0146] In some embodiments of the present disclosure, the screening module 901 is specifically used to:

[0147] Obtaining obstacles whose moving speed in the forward direction of the vehicle is less than the vehicle speed from the screened obstacles;

[0148] For each obstacle obtained, the moment when the vehicle starts braking from the moment of detection and the position of the vehicle in the forward direction is the same as the position of the obstacle is predicted as the collision moment;

[0149] Determining a predicted position of the obstacle at a predicted collision moment according to a moving speed of the obstacle;

[0150] Determine whether the shortest distance between the predicted position and the vehicle's predicted trajectory planning line is less than a specified safety threshold;

[0151] If so, the obstacle is determined to be a suspicious obstacle;

[0152] If not, it is determined that the obstacle is a non-suspicious obstacle.

[0153] In some embodiments of the present disclosure, the determining module 903 is specifically configured to:

[0154] Determine whether there is a target detection time that meets the preset conditions in the detection time set; wherein the preset conditions include: the detection time is less than or equal to the braking time, and the difference between the braking time and the detection time is less than the preset time difference;

[0155] If it exists, it is determined that the braking reason at the braking moment is a false detection of an obstacle;

[0156] If not, it is determined that the braking reason at the braking moment is not a false detection of an obstacle.

[0157] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of vehicle information and obstacle information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0158] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0159] Fig.10 A schematic block diagram of an example electronic device 1000 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0160] like Fig.10 As shown, the electronic device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the electronic device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0161] Multiple components in the electronic device 1000 are connected to the I / O interface 1005, including: an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a disk, an optical disk, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the electronic device 1000 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0162] The computing unit 1001 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 1001 performs the various methods and processes described above, such as the false detection brake detection method. For example, in some embodiments, the false detection brake detection method may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 1008. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 1000 via the ROM 1002 and / or the communication unit 1009. When the computer program is loaded into the RAM 1003 and executed by the computing unit 1001, one or more steps of the false detection brake detection method described above may be performed. Alternatively, in other embodiments, the computing unit 1001 may be configured to execute the false brake detection method in any other appropriate manner (eg, by means of firmware).

[0163] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0164] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0165] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0166] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0167] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0168] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0169] The disclosed embodiment also provides an autonomous driving vehicle, comprising the above-mentioned electronic device.

[0170] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0171] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A false brake detection method, comprising: From the obstacle set, select suspicious obstacles that may have a collision risk with the vehicle; The obstacle set is obstacles detected during the vehicle's driving process; Screen out misdetected obstacles with the possibility of misdetection from various suspicious obstacles; The detection time of each misdetected obstacle is obtained respectively to obtain a detection time set; Obtaining a braking moment of the vehicle, and determining whether a braking cause at the braking moment is a false detection of an obstacle by comparing the braking moment with the detection moment set; The step of screening out suspicious obstacles with a risk of collision with the vehicle from the obstacle set includes: Obtaining obstacles whose moving speed in the forward direction of the vehicle is less than the speed of the vehicle from the screened obstacles, wherein the screened obstacles are obstacles other than obstacles that will not collide with the vehicle; For each obstacle acquired, predicting the moment when the vehicle starts braking from the moment of detection and the position of the vehicle in the forward direction is the same as the position of the obstacle as the predicted collision moment; Determining a predicted position of the obstacle at the predicted collision moment according to the moving speed of the obstacle; Determining whether the shortest distance between the predicted position and the vehicle predicted trajectory planning line is less than a specified safety threshold; If so, the obstacle is determined to be a suspicious obstacle; If not, it is determined that the obstacle is a non-suspicious obstacle.

2. The method according to claim 1, wherein: The step of screening out misdetected obstacles with a possibility of misdetection from each suspicious obstacle includes: For each suspicious obstacle, determine whether the suspicious obstacle is a dynamic obstacle; If the suspected obstacle is a dynamic obstacle, determining whether the detection state of the suspected obstacle is stable detection according to the time when the suspected obstacle is detected; The non-stably detected dynamic obstacle is regarded as the falsely detected obstacle; If the suspected obstacle is a static obstacle, determining whether the suspected obstacle has been run over by the vehicle based on the historical driving trajectory of the vehicle; The static obstacle that is crushed by the vehicle is regarded as the misdetected obstacle.

3. The method according to claim 2, wherein: The determining, based on the time when the suspicious obstacle is detected, whether the detection state of the suspicious obstacle is stable detection includes: Starting from the detection time of the suspicious obstacle, determine whether the cumulative detection time of the suspicious obstacle exceeds the preset time within the preset time period, and whether the suspicious obstacle is detected at the specified time; If the judgment results are all yes, then the detection state of the suspicious obstacle is determined to be stable detection; If any of the judgment results is no, it is determined that the detection state of the suspicious obstacle is unstable detection.

4. The method according to claim 2, wherein: The determining, based on the historical driving trajectory of the vehicle, whether the suspicious obstacle is run over by the vehicle comprises: Determine whether the shortest distance between the center point of the suspicious obstacle and the historical driving trajectory of the vehicle is less than or equal to a specified distance; If yes, it is determined that the suspected obstacle is run over by the vehicle; If not, then determining the boundary position of the vehicle when the vehicle is closest to the suspected obstacle based on the historical driving trajectory of the vehicle, and judging whether the boundary position of the suspected obstacle intersects with the boundary position of the vehicle; If so, it is determined that the suspected obstacle is run over by the vehicle; If not, it is determined that the suspicious obstacle has not been run over by the vehicle.

5. The method according to any one of claims 1 to 4, before obtaining obstacles whose moving speed in the forward direction of the vehicle is less than the speed of the vehicle from the screened obstacles, further comprising: From the obstacle set, obstacles are screened out, which are located in the forward direction of the vehicle, are less than a preset distance from the vehicle, and whose shortest distance from the vehicle predicted trajectory planning line is less than a specified safety threshold.

6. The method according to any one of claims 1 to 4, wherein: The step of comparing the braking moment with the detection moment set to determine whether the braking cause at the braking moment is a false detection of an obstacle comprises: Determine whether there is a target detection moment satisfying a preset condition in the detection moment set; wherein the preset condition includes: the detection moment is less than or equal to the braking moment, and the difference between the braking moment and the detection moment is less than a preset time difference; If yes, determining that the braking reason at the braking moment is a false detection of an obstacle; If not, it is determined that the braking reason at the braking moment is not a false detection of an obstacle.

7. A false brake detection device, comprising: A screening module is used to screen out suspicious obstacles that may have a collision risk with the vehicle from the obstacle set; The obstacle set is obstacles detected during the vehicle's driving process; The screening module is further used to screen out misdetected obstacles with a possibility of misdetection from various suspicious obstacles; An acquisition module, used to respectively acquire the detection time of each misdetected obstacle screened by the screening module to obtain a detection time set; a determination module, configured to obtain a braking moment of the vehicle, and determine whether the braking reason at the braking moment is a false detection of an obstacle by comparing the braking moment with the detection moment set obtained by the acquisition module; The screening module is specifically used for: Obtaining obstacles whose moving speed in the forward direction of the vehicle is less than the speed of the vehicle from among the screened obstacles, wherein the screened obstacles are obstacles other than obstacles that will not collide with the vehicle; For each obstacle acquired, predicting the moment when the vehicle starts braking from the moment of detection and the position of the vehicle in the forward direction is the same as the position of the obstacle as the predicted collision moment; Determining a predicted position of the obstacle at the predicted collision moment according to the moving speed of the obstacle; Determining whether the shortest distance between the predicted position and the vehicle predicted trajectory planning line is less than a specified safety threshold; If so, the obstacle is determined to be a suspicious obstacle; If not, it is determined that the obstacle is a non-suspicious obstacle.

8. The device according to claim 7, wherein: The screening module is specifically used for: For each suspicious obstacle, determine whether the suspicious obstacle is a dynamic obstacle; If the suspected obstacle is a dynamic obstacle, determining whether the detection state of the suspected obstacle is stable detection according to the time when the suspected obstacle is detected; The non-stably detected dynamic obstacle is regarded as the falsely detected obstacle; If the suspected obstacle is a static obstacle, determining whether the suspected obstacle has been run over by the vehicle based on the historical driving trajectory of the vehicle; The static obstacle that is crushed by the vehicle is regarded as the misdetected obstacle.

9. The device according to claim 8, wherein: The screening module is specifically used for: Starting from the detection time of the suspicious obstacle, determine whether the cumulative detection time of the suspicious obstacle exceeds the preset time within the preset time period, and whether the suspicious obstacle is detected at the specified time; If the judgment results are all yes, then the detection state of the suspicious obstacle is determined to be stable detection; If any of the judgment results is no, it is determined that the detection state of the suspicious obstacle is unstable detection.

10. The device according to claim 8, wherein: The screening module is specifically used for: Determine whether the shortest distance between the center point of the suspicious obstacle and the historical driving trajectory of the vehicle is less than or equal to a specified distance; If yes, it is determined that the suspected obstacle is run over by the vehicle; If not, then determining the boundary position of the vehicle when the vehicle is closest to the suspected obstacle based on the historical driving trajectory of the vehicle, and judging whether the boundary position of the suspected obstacle intersects with the boundary position of the vehicle; If so, it is determined that the suspected obstacle is run over by the vehicle; If not, it is determined that the suspicious obstacle has not been run over by the vehicle.

11. The device according to any one of claims 7 to 10, wherein: The screening module is further used for: Before obtaining an obstacle whose moving speed in the forward direction of the vehicle is less than the speed of the vehicle from among the screened obstacles, screen out from the obstacle set obstacles that are located in the forward direction of the vehicle, are less than a preset distance from the vehicle, and whose shortest distance from the predicted trajectory planning line of the vehicle is less than a specified safety threshold.

12. The device according to any one of claims 7 to 10, wherein: The determining module is specifically used for: Determine whether there is a target detection moment satisfying a preset condition in the detection moment set; wherein the preset condition includes: the detection moment is less than or equal to the braking moment, and the difference between the braking moment and the detection moment is less than a preset time difference; If yes, determining that the braking reason at the braking moment is a false detection of an obstacle; If not, it is determined that the braking reason at the braking moment is not a false detection of an obstacle.

13. An electronic device, comprising: at least one processor; as well as 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, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

15. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

16. An autonomous driving vehicle comprising the electronic device as claimed in claim 13.

Images