Automatic driving decision scene recognition method, device, equipment and storage medium

Abstract

The present invention relates to the field of autonomous driving control and discloses a method, apparatus, device, and storage medium for identifying autonomous driving decision scenarios. The method comprises: identifying the takeover moment for an autonomous driving vehicle, and judging, based on the takeover moment, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition; if the autonomous driving system satisfies the preset trigger condition, then, based on the takeover moment, calculating the displacement information of the autonomous driving vehicle in a preset time interval; based on the displacement information, identifying whether the autonomous driving system satisfies a preset decision misjudgment condition, and, based on the identification result, determining the lane change decision scenario corresponding to the autonomous driving system. The present invention improves the accuracy of marking the autonomous driving system in decision-making errors.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving control, and in particular to an autonomous driving decision-making scene recognition method, device, equipment and storage medium. Background Art

[0002] Lane changing and lane borrowing are common scenarios in autonomous driving. Lane changing involves the vehicle leaving its current lane for an adjacent lane, while lane borrowing involves the vehicle leaving its current lane for an adjacent lane, then bypassing an obstacle ahead and returning to its original lane. In driving scenarios where lane changing or lane borrowing is not appropriate, if the autonomous driving system executes the corresponding lane change or lane borrowing decision, the driver will take over and prevent the autonomous driving system from changing or borrowing lanes. In driving scenarios where lane changing or lane borrowing is appropriate, if the autonomous driving system fails to execute the corresponding lane change or lane borrowing decision, the driver will also take over and manually change lanes or lane borrow.

[0003] The current method for identifying such decision-making errors is to have the driver manually report the problem during road testing, thereby marking the driving segment as a lane change or lane borrowing decision error. The marked content is then used to collect driving data of the erroneous decision, providing an optimization basis for the lane change and lane borrowing decision algorithms. However, the driving data collected by this method is often not comprehensive. Therefore, the existing methods for identifying lane change or lane borrowing decision errors have low accuracy in identifying decision-making error scenarios, such as the decision error category and the decision error time range. Summary of the Invention

[0004] The main purpose of the present invention is to solve the technical problem that the existing methods for identifying lane change or lane borrowing decision errors have low accuracy in identifying decision error scenarios.

[0005] A first aspect of the present invention provides a method for identifying autonomous driving decision scenarios, comprising: identifying a takeover moment for an autonomous driving vehicle, and judging, based on the takeover moment, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition; if the autonomous driving system satisfies the preset trigger condition, calculating, based on the takeover moment, the displacement information of the autonomous driving vehicle in a preset time interval; based on the displacement information, identifying whether the autonomous driving system meets a preset decision misjudgment condition, and determining, based on the identification result, the lane change decision scenario corresponding to the autonomous driving system.

[0006] Optionally, in a first implementation manner of the first aspect of the present invention, the preset trigger condition includes a first trigger condition and a second trigger condition, and judging whether the control of the autonomous driving system over the autonomous driving vehicle meets the preset trigger condition based on the takeover moment includes: judging whether the autonomous driving system makes a lane change decision for the autonomous driving vehicle within a preset first time before the takeover moment; if the autonomous driving system does not make a lane change decision for the autonomous driving vehicle, determining that the autonomous driving system meets the first trigger condition; if the autonomous driving system makes a lane change decision for the autonomous driving vehicle, judging whether the takeover moment is within the time period of the lane change decision execution process; if the takeover moment is within the time period of the lane change decision execution process, determining that the autonomous driving system meets the second trigger condition; if the takeover moment is not within the time period of the lane change decision execution process, determining that the autonomous driving system does not meet the preset trigger condition.

[0007] Optionally, in a second implementation manner of the first aspect of the present invention, if the autonomous driving system meets a preset trigger condition, then according to the takeover moment, calculating the displacement information of the autonomous driving vehicle in a preset time interval includes: if the autonomous driving system meets the first trigger condition, then within the preset first time interval after the takeover moment, calculating the first parallel displacement and the first vertical displacement of the autonomous driving vehicle along the lane line, wherein the displacement information includes the first parallel displacement and the first vertical displacement, and the preset time interval includes the first time interval.

[0008] Optionally, in a third implementation manner of the first aspect of the present invention, identifying whether the automatic driving system meets a preset decision misjudgment condition based on the displacement information includes: determining whether the first parallel displacement is greater than a preset first parallel displacement threshold, and determining whether the first vertical displacement is greater than a preset first vertical displacement threshold; if the first parallel displacement is greater than the preset first parallel displacement threshold, and the first vertical displacement is greater than the preset first vertical displacement threshold, determining that the automatic driving system meets the preset decision misjudgment condition, otherwise determining that the automatic driving system does not meet the preset decision misjudgment condition.

[0009] Optionally, in a fourth implementation manner of the first aspect of the present invention, if the automatic driving system meets a preset trigger condition, then according to the takeover moment, calculating the displacement information of the automatic driving vehicle in a preset time interval includes: if the automatic driving system meets the second trigger condition, then calculating the second parallel displacement and second vertical displacement of the automatic driving vehicle in a preset second time interval before the takeover moment, and calculating the third parallel displacement and third vertical displacement of the automatic driving vehicle in a preset third time interval after the takeover moment, wherein the displacement information includes the second parallel displacement and the second vertical displacement, and the third parallel displacement and the third vertical displacement, and the preset time interval includes the second time interval and the third time interval.

[0010] Optionally, in a fifth implementation manner of the first aspect of the present invention, identifying whether the automatic driving system meets a preset decision misjudgment condition based on the displacement information includes: calculating the relative displacement between the second vertical displacement and the third vertical displacement; judging whether the relative displacement is less than a preset second vertical displacement threshold, judging whether the second parallel displacement is greater than a preset second parallel displacement threshold, and judging whether the third parallel displacement is greater than a preset third parallel displacement threshold; if the relative displacement is less than the preset second vertical displacement threshold, the second parallel displacement is greater than the preset second parallel displacement threshold, and the third parallel displacement is greater than the preset third parallel displacement threshold, then it is determined that the automatic driving system meets the preset decision misjudgment condition; otherwise, it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

[0011] Optionally, in a sixth implementation manner of the first aspect of the present invention, after determining the lane change decision scenario corresponding to the autonomous driving system based on the identification result, it also includes: if the autonomous driving system meets a preset decision misjudgment condition, then when the lane change decision is a lane-borrowing decision, within the second time interval and the third time interval, detecting obstacles in the driving scene where the autonomous driving vehicle is located; identifying the time-series change trajectory of the relative position between the autonomous driving vehicle and the obstacle, and adjusting the lane change decision scenario according to the time-series change trajectory.

[0012] The second aspect of the present invention provides an autonomous driving decision-making scenario recognition device, comprising: an identification module for identifying the takeover moment of the autonomous driving vehicle, and judging whether the autonomous driving system's control of the autonomous driving vehicle meets a preset trigger condition based on the takeover moment; a calculation module for calculating the displacement information of the autonomous driving vehicle in a preset time interval based on the takeover moment if the autonomous driving system meets the preset trigger condition; a determination module for identifying whether the autonomous driving system meets a preset decision misjudgment condition based on the displacement information, and determining the lane change decision scenario corresponding to the autonomous driving system based on the identification result.

[0013] Optionally, in a first implementation manner of the second aspect of the present invention, the preset trigger condition includes a first trigger condition and a second trigger condition, and the identification module includes: a first trigger unit, used to determine whether the automatic driving system makes a lane change decision for the automatic driving vehicle within a preset first time before the takeover moment; if the automatic driving system does not make a lane change decision for the automatic driving vehicle, it is determined that the automatic driving system meets the first trigger condition; a second trigger unit, used to determine whether the takeover moment is within the time period of the lane change decision execution process if the automatic driving system makes a lane change decision for the automatic driving vehicle; if the takeover moment is within the time period of the lane change decision execution process, it is determined that the automatic driving system meets the second trigger condition; if the takeover moment is not within the time period of the lane change decision execution process, it is determined that the automatic driving system does not meet the preset trigger condition.

[0014] Optionally, in a second implementation of the second aspect of the present invention, the calculation module includes a first calculation unit, which is used to: if the autonomous driving system meets the first trigger condition, calculate the first parallel displacement and the first vertical displacement of the autonomous driving vehicle along the lane line within a preset first time interval after the takeover moment, wherein the displacement information includes the first parallel displacement and the first vertical displacement, and the preset time interval includes the first time interval.

[0015] Optionally, in a third implementation manner of the second aspect of the present invention, the determination module includes a first identification unit, used to: determine whether the first parallel displacement is greater than a preset first parallel displacement threshold, and determine whether the first vertical displacement is greater than a preset first vertical displacement threshold; if the first parallel displacement is greater than the preset first parallel displacement threshold, and the first vertical displacement is greater than the preset first vertical displacement threshold, it is determined that the automatic driving system meets the preset decision misjudgment condition, otherwise it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

[0016] Optionally, in a fourth implementation manner of the second aspect of the present invention, the calculation module also includes a second calculation unit, which is used to: if the automatic driving system meets the second trigger condition, calculate the second parallel displacement and the second vertical displacement of the automatic driving vehicle within a preset second time interval before the takeover moment, and calculate the third parallel displacement and the third vertical displacement of the automatic driving vehicle within a preset third time interval after the takeover moment, wherein the displacement information includes the second parallel displacement and the second vertical displacement, and the third parallel displacement and the third vertical displacement, and the preset time interval includes the second time interval and the third time interval.

[0017] Optionally, in a fifth implementation manner of the second aspect of the present invention, the determination module also includes a second identification unit, used to: calculate the relative displacement between the second vertical displacement and the third vertical displacement; determine whether the relative displacement is less than a preset second vertical displacement threshold, determine whether the second parallel displacement is greater than a preset second parallel displacement threshold, and determine whether the third parallel displacement is greater than a preset third parallel displacement threshold; if the relative displacement is less than the preset second vertical displacement threshold, the second parallel displacement is greater than the preset second parallel displacement threshold, and the third parallel displacement is greater than the preset third parallel displacement threshold, then it is determined that the automatic driving system meets the preset decision misjudgment condition; otherwise, it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

[0018] Optionally, in a sixth implementation manner of the second aspect of the present invention, the autonomous driving decision scene recognition device further includes a detection module, which is used to: if the autonomous driving system meets a preset decision misjudgment condition, then when the lane change decision is a lane-borrowing decision, detect obstacles in the driving scene where the autonomous driving vehicle is located within the second time interval and the third time interval; identify the time-series change trajectory of the relative position between the autonomous driving vehicle and the obstacle, and adjust the lane change decision scene according to the time-series change trajectory.

[0019] The third aspect of the present invention provides an autonomous driving decision-making scene recognition device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the autonomous driving decision-making scene recognition device executes the above-mentioned autonomous driving decision-making scene recognition method.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium, which stores instructions that, when run on a computer, enable the computer to execute the above-mentioned autonomous driving decision-making scenario recognition method.

[0021] In the technical solution provided by the present invention, by identifying the decision-making moment of the autonomous driving vehicle when changing lanes or borrowing lanes through autonomous driving, or the takeover moment when changing lanes or borrowing lanes through takeover, the trigger conditions are used to preliminarily determine from the time whether there are decision errors in the two situations. Then, through the decision misjudgment conditions, the accurate decision scenarios of the two situations are further determined from the displacement, and the accurate decision error categories and decision error time ranges are determined. The decision scenarios with decision errors are intelligently marked, so that the marking accuracy of decision error scenarios is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a schematic diagram of a first embodiment of the method for identifying autonomous driving decision-making scenarios according to an embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram of a second embodiment of the method for identifying autonomous driving decision-making scenarios according to an embodiment of the present invention;

[0024] Figure 3 This is a schematic diagram of a third embodiment of the method for identifying autonomous driving decision-making scenarios according to an embodiment of the present invention;

[0025] Figure 4 This is a schematic diagram of an embodiment of an autonomous driving decision-making scene recognition device according to an embodiment of the present invention;

[0026] Figure 5 This is a schematic diagram of another embodiment of the autonomous driving decision-making scene recognition device according to an embodiment of the present invention;

[0027] Figure 6 Schematic diagram of an embodiment of an autonomous driving decision-making scenario recognition device in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] Embodiments of the present invention provide a method, apparatus, device, and storage medium for identifying autonomous driving decision scenarios. These methods identify the takeover moment for an autonomous driving vehicle and, based on the takeover moment, determine whether the autonomous driving system's control of the autonomous driving vehicle meets preset trigger conditions. If the autonomous driving system meets the preset trigger conditions, the method calculates the displacement information of the autonomous driving vehicle within a preset time interval based on the takeover moment. Based on the displacement information, the method identifies whether the autonomous driving system meets preset decision misjudgment conditions and, based on the identification results, determines the lane change decision scenario corresponding to the autonomous driving system. This invention improves the accuracy of identifying scenarios in which the autonomous driving system makes incorrect decisions.

[0029] The terms "first," "second," "third," "fourth," and the like (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatus.

[0030] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the autonomous driving decision-making scene recognition method in the embodiments of the present invention includes:

[0031] 101. Identify a takeover time for the autonomous driving vehicle, and determine, based on the takeover time, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition;

[0032] It is understood that the execution subject of the present invention can be an autonomous driving decision scene recognition device, or a terminal or a server, and the specific implementation is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example.

[0033] In this embodiment, the autonomous driving system identifies decision-making scenarios for the autonomous vehicle in lane change scenarios, such as lane changes or lane borrowing. If the driver takes over control of the autonomous vehicle, this indicates a possible decision error or vehicle failure in the autonomous driving system, and the corresponding decision scenario is classified as an erroneous control scenario. The takeover time is recorded. Subsequently, based on the autonomous driving system's control of the autonomous vehicle, including at least the decision type and time, it is determined whether preset trigger conditions are met. The trigger conditions are used to screen whether the erroneous control scenario occurs during a lane change or lane borrowing decision, thereby triggering subsequent misjudgment identification of this decision scenario.

[0034] In this embodiment, the autonomous driving system may encounter two types of erroneous decision-making scenarios when controlling the autonomous vehicle: making a lane change / borrowing decision when it should not have, and failing to make a lane change / borrowing decision when it should have. For the first type of erroneous decision, which indicates a lane change / borrowing decision was made before the takeover time, at least one trigger condition may be set: the lane change / borrowing decision must have been made within a certain time period before the takeover time. For the second type of erroneous decision, which indicates a lane failure / borrowing decision was made before the takeover time, at least one additional trigger condition may be set: the lane change / borrowing decision must have been made within a certain time period before the takeover time.

[0035] 102. If the automatic driving system meets a preset trigger condition, calculate the displacement information of the automatic driving vehicle in a preset time interval according to the takeover time;

[0036] In this embodiment, if the autonomous vehicle does not meet the triggering conditions, it indicates that the decision scenario is not a lane change decision scenario. Otherwise, it is a lane change decision scenario. Further determination is then required to determine whether the lane change decision scenario is an erroneous decision scenario and the type of erroneous decision scenario. Within the preset time interval corresponding to the takeover moment, the autonomous vehicle's displacement information is used to further determine whether the decision scenario is an erroneous lane change decision scenario. For example, this determination includes determining whether the driver's takeover of the autonomous vehicle resulted from executing a lane change, making or not making a lane change decision, or whether it was a non-erroneous decision scenario resulting from a vehicle stall or emergency braking, but requiring a takeover.

[0037] Specifically, the four aforementioned principles for determining whether an erroneous decision scenario is a lane change decision require calculating the parallel and vertical displacements of the autonomous vehicle after the takeover moment for the first and second principles, and further calculating the parallel and vertical displacements before the takeover moment for the third principle. That is, the preset time interval includes the time interval before and after the takeover moment, and the corresponding calculated displacement information includes the parallel and vertical displacements for the time intervals before and after the takeover moment.

[0038] Among them, parallel displacement refers to the displacement caused by moving in the direction parallel to the lane line, that is, the displacement caused by moving in the direction parallel to the direction of travel of the vehicle body; vertical displacement refers to the displacement caused by moving in the direction perpendicular to the lane line (approaching / moving away from the lane line), that is, the displacement caused by moving left and right in the direction parallel to the direction of travel of the vehicle body.

[0039] 103. Identify, based on the displacement information, whether the autonomous driving system meets a preset decision misjudgment condition, and determine, based on the identification result, a lane change decision scenario corresponding to the autonomous driving system.

[0040] In this embodiment, according to the four aforementioned principles for determining whether an erroneous decision scenario is a lane change decision scenario, corresponding misjudgment sub-conditions are formulated respectively. Then, according to lane change decision scenarios such as lane change decision error and lane borrowing decision error, one or more misjudgment sub-conditions are used to generate misjudgment conditions corresponding to each lane change decision scenario for identification based on displacement information.

[0041] Specifically, the parallel displacement of the autonomous driving vehicle before and / or after the takeover moment can be used to determine whether special situations such as vehicle jamming or emergency braking have occurred; the parallel displacement and vertical displacement after the takeover moment can be used to determine whether the driver's takeover control of the autonomous driving vehicle is to execute a lane change, and to determine the type of lane change, whether it is a lane change or a borrowed lane.

[0042] In an embodiment of the present invention, by identifying the decision moment of the autonomous driving vehicle when changing lanes or borrowing lanes through autonomous driving, or the takeover moment when changing lanes or borrowing lanes through takeover, the trigger conditions are used to preliminarily determine from the time whether there are decision errors in the two situations. Then, the decision misjudgment conditions are used to further determine the accurate decision scenarios of the two situations from the displacement, and the accurate decision error categories and decision error time ranges are used to intelligently mark the decision scenarios with decision errors, thereby improving the marking accuracy of decision error scenarios.

[0043] See also Figure 2 A second embodiment of the method for identifying an autonomous driving decision-making scenario in an embodiment of the present invention includes:

[0044] 201. Identify a takeover time for the autonomous driving vehicle, and determine whether the autonomous driving system makes a lane change decision for the autonomous driving vehicle within a preset first time before the takeover time;

[0045] 202. If the autonomous driving system does not make a lane change decision for the autonomous driving vehicle, determining that the autonomous driving system meets the first triggering condition;

[0046] 203. If the autonomous driving system makes a lane change decision for the autonomous driving vehicle, determining whether the takeover moment is within a time period of the lane change decision execution process;

[0047] 204. If the takeover moment is within the time period of the lane change decision execution process, determining that the automatic driving system meets the second triggering condition;

[0048] 205. If the takeover moment is not within the time period of the lane change decision execution process, determining that the automatic driving system does not meet the preset triggering condition;

[0049] In this embodiment, the takeover moment for the autonomous vehicle can be identified in real time while the autonomous vehicle is driving, or it can be identified later when the autonomous vehicle's driving data is collated. First, by determining whether the autonomous vehicle system made a lane change decision within a preset first time period before the takeover moment, it is determined whether the decision scenario resulted in the autonomous vehicle system making a lane change decision or not making a lane change decision.

[0050] Specifically, for example, after identifying the takeover time t1, the lane change decision made by reverse time search is determined at t i The automated driving system (planning and control module) is checked for a lane change decision, preferably within 10 seconds. If the automated driving system does not execute a lane change decision within 10 seconds, it is determined that the first trigger condition is met, and the corresponding misjudgment condition is subsequently determined. If the automated driving system was executing a lane change decision within 10 seconds, it is further determined whether the takeover occurred while the automated driving system was executing the lane change decision, or after the lane change decision was completed. The former can be considered a lane change decision error and the second trigger condition is met. The latter cannot be considered a lane change decision and can be set as not meeting the trigger condition and not belonging to an erroneous decision scenario.

[0051] 206. If the autonomous driving system satisfies the first trigger condition, then, within a preset first time interval after the takeover moment, calculating a first parallel displacement and a first vertical displacement of the autonomous driving vehicle along the lane line, wherein the displacement information includes the first parallel displacement and the first vertical displacement, and the preset time interval includes the first time interval;

[0052] In this embodiment, if the autonomous driving system meets the first trigger condition, it determines that the current decision-making scenario may be a takeover caused by the autonomous vehicle failing to make a lane change decision when a lane change or lane change is required. In this case, it is necessary to verify whether the driver manipulates the autonomous vehicle to change lanes after taking over to determine whether a lane change or lane change is actually required in the current driving scenario. As previously mentioned, the displacement information required in this case includes the first parallel displacement and the first vertical displacement after the takeover moment, which are the displacements caused by the driver's manual driving after the takeover.

[0053] Specifically, for example, the displacement of the autonomous driving vehicle along the lane line (first parallel displacement) human_drive_s and the displacement human_drive_l in the direction perpendicular to the lane line within 5 seconds (within a preset first time interval) after the driver takes over (after the takeover time t1), then the absolute value of the displacement human_drive_l between [t1, t1+5s] is taken, and the displacement with the largest absolute value human_drive_l_max between [t1, t1+5s] is selected as the first perpendicular displacement.

[0054] 207. Determine whether the first parallel displacement is greater than a preset first parallel displacement threshold, and determine whether the first vertical displacement is greater than a preset first vertical displacement threshold;

[0055] 208. If the first parallel displacement is greater than a preset first parallel displacement threshold, and the first vertical displacement is greater than a preset first vertical displacement threshold, it is determined that the autonomous driving system meets a preset decision misjudgment condition; otherwise, it is determined that the autonomous driving system does not meet the preset decision misjudgment condition, and based on the recognition result, the lane change decision scenario corresponding to the autonomous driving system is determined.

[0056] In this embodiment, a first parallel displacement threshold value, human_drive_0, is set to determine special situations such as vehicle jamming and emergency braking. A first vertical displacement threshold value, human_drive_1_max_0, is set to determine whether the driver has changed lanes for the autonomous vehicle. This is used to set the decision misjudgment condition and determine the lane change decision scenario corresponding to the autonomous driving system.

[0057] Specifically, if human_drive_0 is set to 0.5m and human_drive_1_max_0 is set to 1.5m, then when human_drive_s > 0.5m and human_drive_l_max > 1.5m, it means that the driver has changed lanes or borrowed lanes after taking over, and the current autonomous driving system has not made a lane change or borrowing decision. This is a lane change or borrowing decision error, and the corresponding lane change decision scenario is an incorrect decision scenario.

[0058] In addition, when human_drive_s is less than 0.5m, it is determined that the current driving scenario is a special scenario such as vehicle jamming or emergency braking, and the autonomous driving system has not made an incorrect decision. When human_drive_l_max is less than 1.5m, it means that the driver did not change lanes or borrow lanes after taking over. It may be that the driver took over the autonomous driving vehicle due to other reasons. The current autonomous driving system did not make a lane change or borrow lane decision. The lane change or borrow lane decision is correct, and the corresponding lane change decision scenario is a correct decision scenario.

[0059] When an incorrect decision is detected, the data acquisition module is triggered to collect data and mark this data as an incorrect lane change or lane borrowing decision. Note that the thresholds for human_drive_0 and human_drive_1_max_0 are set to 0.5m and 1.5m, respectively, for typical urban roads. For other road conditions, the thresholds can be adjusted accordingly.

[0060] See also Figure 3 A third embodiment of the method for identifying an autonomous driving decision-making scenario in an embodiment of the present invention includes:

[0061] 301. Identify a takeover time for the autonomous driving vehicle, and determine, based on the takeover time, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition.

[0062] 302. If the autonomous driving system satisfies the second trigger condition, calculating a second parallel displacement and a second vertical displacement of the autonomous driving vehicle within a preset second time interval before the takeover moment, and calculating a third parallel displacement and a third vertical displacement of the autonomous driving vehicle within a preset third time interval after the takeover moment, wherein the displacement information includes the second parallel displacement and the second vertical displacement, and the third parallel displacement and the third vertical displacement, and the preset time interval includes the second time interval and the third time interval.

[0063] In this embodiment, if the autonomous driving system meets the second trigger condition, it determines that the current decision scenario may be a takeover caused by the autonomous vehicle making a lane change decision when a lane change or lane borrowing operation is not required. Alternatively, the autonomous vehicle making a lane change decision may be different from the lane change or lane borrowing operation required by the current decision scenario, resulting in the need for driver takeover. In this case, it is necessary to verify whether the driver controls the autonomous vehicle to change lanes after taking over to determine whether the vehicle actually needs to change lanes or borrow lanes for the current driving scenario. As described above, the displacement information required in this case includes the second parallel displacement and the second vertical displacement before the takeover moment, and the third parallel displacement and the third vertical displacement after the takeover moment.

[0064] Specifically, for example, the displacement of the autonomous vehicle along the lane line (second parallel displacement) auto_s and the displacement perpendicular to the lane line (second vertical displacement) auto_l are calculated within 3 seconds (preset second time interval) before the driver takes over (before takeover time t1). Also, the displacement of the autonomous vehicle along the lane line (third parallel displacement) human_drive_s and the displacement perpendicular to the lane line (third vertical displacement) human_drive_l are calculated within 10 seconds (preset third time interval) after the driver takes over (after takeover time t1).

[0065] 303. Calculate the relative displacement between the second vertical displacement and the third vertical displacement;

[0066] In this embodiment, the relative displacement between the second and third vertical displacements represents the relative displacement perpendicular to the lane line before and after the driver takes over. In one calculation method, the relative displacement between the second and third vertical displacements is expressed by calculating the absolute value of the sum of the two vertical displacements, i.e., relative displacement human_drive_auto_1 = |(auto_1 + human_drive_1)|.

[0067] 304. Determine whether the relative displacement is less than a preset second vertical displacement threshold, determine whether the second parallel displacement is greater than a preset second parallel displacement threshold, and determine whether the third parallel displacement is greater than a preset third parallel displacement threshold;

[0068] 305. If the relative displacement is less than a preset second vertical displacement threshold, the second parallel displacement is greater than the preset second parallel displacement threshold, and the third parallel displacement is greater than the preset third parallel displacement threshold, then it is determined that the autonomous driving system meets a preset misjudgment condition. Otherwise, it is determined that the autonomous driving system does not meet the preset misjudgment condition, and a lane change decision scenario corresponding to the autonomous driving system is determined based on the recognition result.

[0069] In this embodiment, a second parallel displacement threshold value auto_0 and a third parallel displacement pre-threshold value human_drive_0 are set to determine special situations such as vehicle jamming and emergency braking. A second vertical displacement threshold value human_drive_auto_0 is set to determine whether the current driving scenario is a vehicle lane change scenario or a lane borrowing scenario. This is used to set the decision misjudgment condition and determine the lane change decision scenario corresponding to the automatic driving system.

[0070] Specifically, for example, if the second parallel displacement threshold auto_0 is set to 0.5m, the third parallel displacement threshold human_drive_0 is set to 5m, and the second vertical displacement value human_drive_auto_0 is set to 1.5m, then when auto_s>0.5m, human_drive_s>5m, and |(auto_l+human_drive_l)|<1.5m, it means that the driver operated the vehicle to maintain the original lane after taking over, and the current driving scene has not suddenly changed. The lane change or borrowing decision made by the current automatic driving system is an incorrect lane change or borrowing decision, and the corresponding lane change decision scenario is an incorrect decision scenario.

[0071] In addition, when auto_s is less than 0.5m, or human_drive_s is less than 0.5m, it is determined that the current driving scene is a special scene such as vehicle stuck or emergency braking, and the autonomous driving system has not made an incorrect decision; and when |(auto_l+human_drive_l)| is greater than 1.5m, it is determined that the autonomous driving vehicle needs to change lanes in the current driving scene, and the lane change or borrowing decision made by the autonomous driving system is correct, and the corresponding lane change decision scenario may be a correct decision scenario.

[0072] At the same time, when the above-mentioned incorrect decision scenario is identified, the data collection module is triggered to collect data and mark this data as a lane change or lane borrowing decision error. It should be noted that the settings of auto_0 = 0.5m, human_drive_0 = 5m, and human_drive_auto_0 = 1.5m in the above steps correspond to typical urban roads. The thresholds can be adjusted accordingly for different roads.

[0073] 306. If the automated driving system satisfies a preset decision misjudgment condition, then when the lane change decision is a lane borrowing decision, detecting obstacles in the driving scene of the automated driving vehicle within the second time interval and the third time interval;

[0074] 307. Identify a temporal change trajectory of the relative position between the autonomous driving vehicle and the obstacle, and adjust the lane change decision scenario based on the temporal change trajectory.

[0075] In this embodiment, if the autonomous driving system makes a lane-borrowing decision before taking over and meets the decision-making error conditions, it can check whether the autonomous driving vehicle has bypassed any obstacles in front by identifying the obstacles in the driving scene, screening the obstacles directly in front, and identifying the time-series transformation trajectory between the autonomous driving vehicle and the obstacles directly in front. If no obstacles in front are bypassed, it is determined that the current autonomous driving system has made an incorrect lane-borrowing decision.

[0076] Specifically, for the obstacle directly ahead, if the autonomous driving vehicle borrows lanes to the left and overtakes the obstacle directly ahead, the identified time-series transformation trajectory of the relative positions between (autonomous driving vehicle, obstacle) is (back, front; left rear, right front; left, right; left front, right rear; front, back). If the autonomous driving vehicle borrows lanes to the right and overtakes the obstacle directly ahead, the identified time-series transformation trajectory of the relative positions between (autonomous driving vehicle, obstacle) is (back, front; right rear, left front; right, left; right front, left rear; front, back). If it is one of the two aforementioned time-series transformation trajectories, it is determined that the obstacle ahead has been bypassed, and the current autonomous driving system has made a correct decision to borrow lanes, and the category of the originally identified wrong decision to borrow lanes is eliminated. Otherwise, it is determined that the autonomous driving system has made an incorrect decision to borrow lanes, and the category of the originally identified wrong decision to borrow lanes is retained.

[0077] The above describes the automatic driving decision scene recognition method in the embodiment of the present invention. The following describes the automatic driving decision scene recognition device in the embodiment of the present invention. Figure 4 In one embodiment of the present invention, an apparatus for identifying a scene for autonomous driving decision-making includes:

[0078] an identification module 401 for identifying a takeover moment for the autonomous driving vehicle and determining, based on the takeover moment, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition;

[0079] a calculation module 402 configured to calculate the displacement information of the autonomous driving vehicle within a preset time interval based on the takeover time if the autonomous driving system meets a preset trigger condition;

[0080] The determination module 403 is used to identify whether the autonomous driving system meets a preset decision misjudgment condition based on the displacement information, and determine a lane change decision scenario corresponding to the autonomous driving system based on the identification result.

[0081] In an embodiment of the present invention, by identifying the decision moment of the autonomous driving vehicle when changing lanes or borrowing lanes through autonomous driving, or the takeover moment when changing lanes or borrowing lanes through takeover, the trigger conditions are used to preliminarily determine from the time whether there are decision errors in the two situations. Then, the decision misjudgment conditions are used to further determine the accurate decision scenarios of the two situations from the displacement, and the accurate decision error categories and decision error time ranges are used to intelligently mark the decision scenarios with decision errors, thereby improving the marking accuracy of decision error scenarios.

[0082] See also Figure 5 Another embodiment of the autonomous driving decision-making scene recognition device in the embodiment of the present invention includes:

[0083] an identification module 401 for identifying a takeover moment for the autonomous driving vehicle and determining, based on the takeover moment, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition;

[0084] a calculation module 402 configured to calculate the displacement information of the autonomous driving vehicle within a preset time interval based on the takeover time if the autonomous driving system meets a preset trigger condition;

[0085] The determination module 403 is used to identify whether the autonomous driving system meets a preset decision misjudgment condition based on the displacement information, and determine a lane change decision scenario corresponding to the autonomous driving system based on the identification result.

[0086] Specifically, the preset trigger condition includes a first trigger condition and a second trigger condition, and the identification module 401 includes:

[0087] a first triggering unit 4011 configured to determine whether the autonomous driving system makes a lane change decision for the autonomous driving vehicle within a preset first time before the takeover moment; and if the autonomous driving system does not make a lane change decision for the autonomous driving vehicle, determining that the autonomous driving system meets the first triggering condition;

[0088] The second trigger unit 4012 is used to determine whether the takeover moment is within the time period of the lane change decision execution process if the autonomous driving system makes a lane change decision for the autonomous driving vehicle; if the takeover moment is within the time period of the lane change decision execution process, it is determined that the autonomous driving system meets the second trigger condition; if the takeover moment is not within the time period of the lane change decision execution process, it is determined that the autonomous driving system does not meet the preset trigger condition.

[0089] Specifically, the calculation module 402 includes a first calculation unit 4021, which is configured to:

[0090] If the autonomous driving system meets the first trigger condition, then within a preset first time interval after the takeover moment, the first parallel displacement and the first vertical displacement of the autonomous driving vehicle along the lane line are calculated, wherein the displacement information includes the first parallel displacement and the first vertical displacement, and the preset time interval includes the first time interval.

[0091] Specifically, the determining module 403 includes a first identifying unit 4031, which is configured to:

[0092] determining whether the first parallel displacement is greater than a preset first parallel displacement threshold, and determining whether the first vertical displacement is greater than a preset first vertical displacement threshold;

[0093] If the first parallel displacement is greater than a preset first parallel displacement threshold, and the first vertical displacement is greater than a preset first vertical displacement threshold, it is determined that the automatic driving system meets the preset decision misjudgment condition; otherwise, it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

[0094] Specifically, the calculation module 402 further includes a second calculation unit 4022, configured to:

[0095] If the autonomous driving system meets the second trigger condition, the second parallel displacement and the second vertical displacement of the autonomous driving vehicle are calculated within a preset second time interval before the takeover moment, and the third parallel displacement and the third vertical displacement of the autonomous driving vehicle are calculated within a preset third time interval after the takeover moment, wherein the displacement information includes the second parallel displacement and the second vertical displacement, and the third parallel displacement and the third vertical displacement, and the preset time interval includes the second time interval and the third time interval.

[0096] Specifically, the determination module 403 further includes a second identification unit 4032, configured to:

[0097] calculating a relative displacement between the second vertical displacement and the third vertical displacement;

[0098] determining whether the relative displacement is less than a preset second vertical displacement threshold, determining whether the second parallel displacement is greater than a preset second parallel displacement threshold, and determining whether the third parallel displacement is greater than a preset third parallel displacement threshold;

[0099] If the relative displacement is less than the preset second vertical displacement threshold, the second parallel displacement is greater than the preset second parallel displacement threshold, and the third parallel displacement is greater than the preset third parallel displacement threshold, it is determined that the automatic driving system meets the preset decision misjudgment condition; otherwise, it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

[0100] Specifically, the autonomous driving decision scene recognition device further includes a detection module 404, which is used to:

[0101] If the automated driving system satisfies a preset decision misjudgment condition, then when the lane change decision is a lane borrowing decision, detecting obstacles in the driving scene of the automated driving vehicle within the second time interval and the third time interval;

[0102] Identify a temporal change trajectory of the relative position between the autonomous driving vehicle and the obstacle, and adjust the lane change decision scenario based on the temporal change trajectory.

[0103] above Figure 4 and Figure 5 The autonomous driving decision-making scene recognition device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The autonomous driving decision-making scene recognition device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0104] Figure 6 6 is a schematic structural diagram of an autonomous driving decision-making scene recognition device provided by an embodiment of the present invention. The autonomous driving decision-making scene recognition device 600 may vary significantly due to different configurations or performance, and may include one or more processors (central processing units, CPUs) 610 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more mass storage devices) storing application programs 633 or data 632. The memory 620 and storage medium 630 may be either short-term storage or persistent storage. The program stored in the storage medium 630 may include one or more modules (not shown), each of which may include a series of instruction operations in the autonomous driving decision-making scene recognition device 600. Furthermore, the processor 610 may be configured to communicate with the storage medium 630 to execute the series of instruction operations in the storage medium 630 on the autonomous driving decision-making scene recognition device 600.

[0105] The autonomous driving decision-making scene recognition device 600 may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 6 The structure of the autonomous driving decision-making scenario recognition device shown does not constitute a limitation of the autonomous driving decision-making scenario recognition device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0106] The present invention also provides an autonomous driving decision-making scene recognition device, wherein the computer device includes a memory and a processor, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes the steps of the autonomous driving decision-making scene recognition method in the above-mentioned embodiments.

[0107] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the autonomous driving decision-making scenario recognition method.

[0108] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0109] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or 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: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., various media that can store program code.

[0110] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. 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 invention.

Claims

1. A method for identifying autonomous driving decision-making scenarios, characterized in that: The autonomous driving decision-making scenario recognition method includes: Identifying a takeover moment for the autonomous driving vehicle, and determining, based on the takeover moment, whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition; If the automatic driving system meets the preset trigger condition, the displacement information of the automatic driving vehicle in the preset time interval is calculated according to the takeover time; identifying, based on the displacement information, whether the autonomous driving system satisfies a preset decision misjudgment condition, and determining, based on a result of the identification, a lane change decision scenario corresponding to the autonomous driving system; The preset trigger condition includes a first trigger condition. If the automatic driving system meets the preset trigger condition, then according to the takeover moment, the displacement information of the automatic driving vehicle in the preset time interval is calculated, including: if the automatic driving system meets the first trigger condition, then within the preset first time interval after the takeover moment, the first parallel displacement and the first vertical displacement of the automatic driving vehicle along the lane line are calculated, wherein the displacement information includes the first parallel displacement and the first vertical displacement, and the preset time interval includes the first time interval.

2. The autonomous driving decision-making scene recognition method according to claim 1, characterized in that: The preset trigger condition also includes a second trigger condition, and judging whether the control of the autonomous driving system over the autonomous driving vehicle satisfies the preset trigger condition according to the takeover moment includes: determining whether the autonomous driving system makes a lane change decision for the autonomous driving vehicle within a preset first time before the takeover moment; If the autonomous driving system does not make a lane change decision for the autonomous driving vehicle, determining that the autonomous driving system meets the first triggering condition; If the autonomous driving system makes a lane change decision for the autonomous driving vehicle, determining whether the takeover moment is within a time period of the lane change decision execution process; If the takeover moment is within the time period of the lane change decision execution process, determining that the autonomous driving system meets the second triggering condition; If the takeover moment is not within the time period of the lane change decision execution process, it is determined that the automatic driving system does not meet the preset trigger condition.

3. The autonomous driving decision-making scene recognition method according to claim 2, characterized in that: The identifying, based on the displacement information, whether the automatic driving system satisfies a preset decision misjudgment condition includes: determining whether the first parallel displacement is greater than a preset first parallel displacement threshold, and determining whether the first vertical displacement is greater than a preset first vertical displacement threshold; If the first parallel displacement is greater than a preset first parallel displacement threshold, and the first vertical displacement is greater than a preset first vertical displacement threshold, it is determined that the automatic driving system meets the preset decision misjudgment condition; otherwise, it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

4. The autonomous driving decision-making scene recognition method according to claim 2, characterized in that: If the automatic driving system satisfies a preset trigger condition, calculating the displacement information of the automatic driving vehicle in a preset time interval according to the takeover time includes: If the autonomous driving system meets the second trigger condition, the second parallel displacement and the second vertical displacement of the autonomous driving vehicle are calculated within a preset second time interval before the takeover moment, and the third parallel displacement and the third vertical displacement of the autonomous driving vehicle are calculated within a preset third time interval after the takeover moment, wherein the displacement information includes the second parallel displacement and the second vertical displacement, and the third parallel displacement and the third vertical displacement, and the preset time interval includes the second time interval and the third time interval.

5. The autonomous driving decision-making scene recognition method according to claim 4, characterized in that: The identifying, based on the displacement information, whether the automatic driving system satisfies a preset decision misjudgment condition includes: calculating a relative displacement between the second vertical displacement and the third vertical displacement; determining whether the relative displacement is less than a preset second vertical displacement threshold, determining whether the second parallel displacement is greater than a preset second parallel displacement threshold, and determining whether the third parallel displacement is greater than a preset third parallel displacement threshold; If the relative displacement is less than the preset second vertical displacement threshold, the second parallel displacement is greater than the preset second parallel displacement threshold, and the third parallel displacement is greater than the preset third parallel displacement threshold, it is determined that the automatic driving system meets the preset decision misjudgment condition; otherwise, it is determined that the automatic driving system does not meet the preset decision misjudgment condition.

6. The autonomous driving decision scene recognition method according to claim 4 or 5, characterized in that: After determining the lane change decision scenario corresponding to the autonomous driving system based on the recognition result, the method further includes: If the automated driving system satisfies a preset decision misjudgment condition, then when the lane change decision is a lane borrowing decision, detecting obstacles in the driving scene of the automated driving vehicle within the second time interval and the third time interval; Identify a temporal change trajectory of the relative position between the autonomous driving vehicle and the obstacle, and adjust the lane change decision scenario based on the temporal change trajectory.

7. An autonomous driving decision scene recognition device, characterized in that: The autonomous driving decision-making scene recognition device includes: an identification module, configured to identify a takeover moment for taking over the autonomous driving vehicle and, based on the takeover moment, determine whether the autonomous driving system's control of the autonomous driving vehicle satisfies a preset trigger condition; a calculation module, configured to calculate the displacement information of the autonomous driving vehicle in a preset time interval according to the takeover moment if the autonomous driving system meets a preset trigger condition; a determination module, configured to identify, based on the displacement information, whether the autonomous driving system satisfies a preset decision misjudgment condition, and determine, based on the identification result, a lane change decision scenario corresponding to the autonomous driving system; The preset trigger condition includes a first trigger condition. If the automatic driving system meets the preset trigger condition, then according to the takeover moment, the displacement information of the automatic driving vehicle in the preset time interval is calculated, including: if the automatic driving system meets the first trigger condition, then within the preset first time interval after the takeover moment, the first parallel displacement and the first vertical displacement of the automatic driving vehicle along the lane line are calculated, wherein the displacement information includes the first parallel displacement and the first vertical displacement, and the preset time interval includes the first time interval.

8. An autonomous driving decision-making scene recognition device, characterized in that: The autonomous driving decision-making scene recognition device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory so that the autonomous driving decision-making scene recognition device performs the steps of the autonomous driving decision-making scene recognition method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the autonomous driving decision-making scene recognition method as described in any one of claims 1 to 6 are implemented.

Images