Method, apparatus, device and storage medium for identifying error takeover cases

Abstract

The present invention relates to the field of R & D management, and discloses a method, device, equipment and storage medium for identifying takeover cases for mistakes. The method includes: simulating historical takeover cases for mistakes in a historical takeover case set for mistakes through a preset simulation algorithm to obtain a predicted trajectory corresponding to each historical takeover case for mistakes; calculating a path difference threshold according to the predicted trajectories and the corresponding actual trajectories corresponding to all historical takeover cases for mistakes, wherein the path difference threshold includes a distance difference threshold and / or an angle difference threshold; and identifying takeover cases for mistakes to be identified according to the path difference threshold. By calculating the predicted trajectory of autonomous driving through simulation technology and based on the path difference between the predicted trajectory and the actual trajectory, a unified judgment standard for judging takeover vehicles due to mistakes is obtained, which can effectively and quickly sort out takeover cases caused by human mistakes, and then can further conduct a detailed study on the cases, solving the technical problem of inconsistent judgment criteria for takeover behaviors in existing autonomous driving vehicles.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and in particular to a method, device, equipment and storage medium for identifying error takeover cases. Background Art

[0002] With the development of science and technology, especially in the field of autonomous driving of vehicles, when evaluating the performance of an autonomous driving system, MPI (total mileage divided by total takeover times) is a relatively mainstream evaluation criterion. To make this criterion correctly reflect the true performance of autonomous driving, it is necessary to ensure that each takeover used to count the total takeover times is caused by a problem with the autonomous driving system. However, there may be multiple reasons for a takeover. Among them, there may be a situation where the driver takes over in advance due to excessive caution, but in fact, if the vehicle continues to drive autonomously, no safety problems will occur. There may also be a situation where the driver subjectively believes that the driving speed of the autonomous driving vehicle is too slow, and thus takes over to accelerate, and so on. The existing methods for screening out such takeovers due to human errors are somewhat subjective and cannot objectively complete the automatic analysis and screening process of cases according to certain criteria or indicators to determine whether a takeover is a human error, which affects the further research on sub-cases. Summary of the Invention

[0003] The main purpose of the present invention is to solve the technical problem that the existing autonomous driving technology cannot determine takeover behaviors according to a unified standard.

[0004] In the first aspect of the present invention, a method for identifying error takeover cases is provided. The method for identifying error takeover cases includes: simulating historical error takeover cases in a historical error takeover case set through a preset simulation algorithm to obtain a predicted trajectory corresponding to each historical error takeover case; calculating a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical error takeover cases, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; and identifying error takeover cases for the takeover cases to be identified according to the path difference threshold.

[0005] Optionally, in the first implementation manner of the first aspect of the present invention, before forcibly occupying the first dynamic memory when the first dynamic memory is released and adding preset object data to the forcibly occupied first dynamic memory to obtain a second dynamic memory, the method further includes: creating an error release function, where the error release function is used to throw an error; performing a Hook operation on a first memory release function, where the first memory release function is a preset basic function in the system for releasing dynamic memory; and writing the error release function into the first memory release function to obtain a second memory release function.

[0006] Optionally, in the second implementation manner of the first aspect of the present invention, simulating the historical takeover cases in the historical takeover case set through a preset simulation algorithm to obtain a predicted trajectory corresponding to each historical takeover case includes: obtaining vehicle state information of each historical takeover case in the historical takeover case set within a preset time period before manual takeover; inputting the vehicle state information of each historical takeover case into the simulation algorithm for simulation to obtain a corresponding predicted trajectory.

[0007] Optionally, in the third implementation manner of the first aspect of the present invention, calculating a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical takeover cases, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold, includes: determining the actual trajectory of each historical takeover case within a preset time period after manual takeover; calculating the path difference between the actual trajectory and the corresponding predicted trajectory of each historical takeover case after manual takeover, where the path difference includes a distance difference and / or an angle difference; statistically analyzing the path differences of all historical takeover cases to obtain the path difference threshold.

[0008] Optionally, in the fourth implementation manner of the first aspect of the present invention, calculating a corresponding path difference threshold based on the first position coordinate point set and the second position coordinate point set of each historical takeover case includes: splitting the first position coordinate point set into a first abscissa point set and a first ordinate point set; splitting the second position coordinate point set into a second abscissa point set and a second ordinate point set; adding the square of the difference between the first abscissa point and the second abscissa point at the same moment to the square of the difference between the first ordinate point and the second ordinate point at the same moment, and then taking the square root of the sum to obtain a square root result; accumulating all the square root results within a preset time period to obtain the distance difference threshold.

[0009] Optionally, in the fifth implementation manner of the first aspect of the present invention, calculating a corresponding path difference threshold based on the first position coordinate point set and the second position coordinate point set of each historical takeover case includes: splitting the first position coordinate point set into a first abscissa point set and a first ordinate point set; splitting the second position coordinate point set into a second abscissa point set and a second ordinate point set; calculating the difference between consecutive second ordinate points within a preset time period divided by the difference between consecutive second abscissa points within a preset time period, and then taking the arctangent of the calculation result to obtain the actual trajectory angle; calculating the difference between consecutive first ordinate points within a preset time period divided by the difference between consecutive first abscissa points within a preset time period, and then taking the arctangent of the calculation result to obtain the predicted trajectory angle; taking the absolute value of the difference between the actual trajectory angle and the predicted trajectory angle to obtain the angle difference threshold.

[0010] Optionally, in the sixth implementation manner of the first aspect of the present invention, the identifying the takeover case with mistakes according to the path difference threshold includes: calculating the distance difference and the angle difference of the takeover case to be identified; determining whether the distance difference of the takeover case to be identified is greater than the distance difference threshold or whether the angle difference of the takeover case to be identified is greater than the angle difference threshold; if the distance difference of the takeover case to be identified is greater than the distance difference threshold or the angle difference of the takeover case to be identified is greater than the angle difference threshold, then identifying the takeover case to be identified as a takeover case with mistakes.

[0011] The second aspect of the present invention provides a device for identifying takeover cases with mistakes, including: a case simulation module, configured to simulate the historical takeover cases with mistakes in a historical takeover case set with mistakes through a preset simulation algorithm to obtain a predicted trajectory corresponding to each historical takeover case with mistakes; a threshold calculation module, configured to calculate a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical takeover cases with mistakes, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; a case identification module, configured to identify the takeover case to be identified as a takeover case with mistakes according to the path difference threshold.

[0012] Optionally, in the first implementation manner of the second aspect of the present invention, the case simulation module is specifically configured to: obtain the vehicle state information of each historical takeover case with mistakes in the historical takeover case set with mistakes within a preset time period before manual takeover; input the vehicle state information of each historical takeover case with mistakes into the simulation algorithm for simulation to obtain a corresponding predicted trajectory.

[0013] Optionally, in the second implementation manner of the second aspect of the present invention, the threshold calculation module is specifically configured to: a trajectory determination unit, determine the actual trajectory of each historical takeover case with mistakes within a preset time period after manual takeover; a path difference calculation unit, calculate the path difference between the actual trajectory and the corresponding predicted trajectory of each historical takeover case with mistakes after manual takeover, where the path difference includes a distance difference and / or an angle difference; a threshold obtaining unit, perform statistics on the path differences of all historical takeover cases with mistakes to obtain a path difference threshold.

[0014] Optionally, in the third implementation manner of the second aspect of the present invention, the path difference calculation unit is specifically configured to: a first set sub-unit, determine a first set of position coordinate points of the predicted trajectory within a preset time period after manual takeover; a second set sub-unit, determine a second set of position coordinate points of the actual trajectory within a preset time period after manual takeover; a path difference calculation sub-unit, calculate a corresponding path difference threshold based on the first set of position coordinate points and the second set of position coordinate points of each historical takeover case with mistakes.

[0015] Optionally, in the fourth implementation manner of the second aspect of the present invention, the path difference calculation sub-unit is specifically configured to: split the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; split the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; add the square of the difference between the first abscissa point and the second abscissa point at the same moment to the square of the difference between the first ordinate point and the second ordinate point, and then take the square root of the sum to obtain a square root result; accumulate all the square root results within a preset time period to obtain the distance difference threshold.

[0016] Optionally, in the fifth implementation manner of the second aspect of the present invention, the path difference calculation sub-unit is specifically configured to: split the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; split the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; calculate the difference between consecutive second ordinate points within a preset time period divided by the difference between consecutive second abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the actual trajectory angle; calculate the difference between consecutive first ordinate points within a preset time period divided by the difference between consecutive first abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the predicted trajectory angle; take the absolute value of the difference between the actual trajectory angle and the predicted trajectory angle to obtain the angle difference threshold.

[0017] Optionally, in the sixth implementation manner of the second aspect of the present invention, the case identification module is specifically configured to: calculate the distance difference and the angle difference of the takeover case to be identified; determine whether the distance difference of the takeover case to be identified is greater than the distance difference threshold or whether the angle difference of the takeover case to be identified is greater than the angle difference threshold; if the distance difference of the takeover case to be identified is greater than the distance difference threshold or the angle difference of the takeover case to be identified is greater than the angle difference threshold, then identify the takeover case to be identified as a misoperation takeover case.

[0018] The third aspect of the present invention provides a misoperation takeover case identification device, including: a memory and at least one processor, instructions are stored in the memory, and the memory and the at least one processor are interconnected through a line; the at least one processor calls the instructions in the memory so that the misoperation takeover case identification device executes the steps of the above misoperation takeover case identification method.

[0019] The fourth aspect of the present invention provides a computer-readable storage medium, instructions are stored in the computer-readable storage medium, and when it runs on a computer, it causes the computer to execute the steps of the above misoperation takeover case identification method.

[0020] In the technical solution of the present invention, historical takeover cases in a historical error takeover case set are simulated through a preset simulation algorithm to obtain a predicted trajectory corresponding to each historical takeover case; according to the predicted trajectories and corresponding actual trajectories corresponding to all historical takeover cases, a path difference threshold is calculated, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; an error takeover case identification is performed on a takeover case to be identified according to the path difference threshold. By calculating the predicted trajectory of autonomous driving through simulation technology and based on the path difference between the predicted trajectory and the actual trajectory, a unified judgment criterion for judging takeover vehicles due to human error is obtained, which can effectively and quickly sort out takeover cases caused by human error, and then can further conduct a sub - research on the cases, solving the technical problem of inconsistent judgment criteria for takeover behaviors in existing autonomous driving vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Schematic diagram of the first embodiment of the error takeover case identification method in the embodiment of the present invention;

[0022] Figure 2 Schematic diagram of the second embodiment of the error takeover case identification method in the embodiment of the present invention;

[0023] Figure 3 Schematic diagram of the third embodiment of the error takeover case identification method in the embodiment of the present invention;

[0024] Figure 4 Schematic diagram of the fourth embodiment of the error takeover case identification method in the embodiment of the present invention;

[0025] Figure 5 Schematic diagram of an embodiment of the error takeover case identification device in the embodiment of the present invention;

[0026] Figure 6 Schematic diagram of another embodiment of the error takeover case identification device in the embodiment of the present invention;

[0027] Figure 7 Schematic diagram of an embodiment of the error takeover case identification device in the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] In the technical solution of the present invention, the historical takeover cases in the historical takeover case set of mistakes are simulated through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical takeover case of mistake; according to the predicted trajectories corresponding to all historical takeover cases of mistake and the corresponding actual trajectories, a path difference threshold is calculated, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; the takeover case to be identified is identified as a takeover case of mistake according to the path difference threshold. By calculating the predicted trajectory of autonomous driving through simulation technology and based on the path difference between the predicted trajectory and the actual trajectory, a unified judgment standard for judging a takeover vehicle due to human error is obtained, which can effectively and quickly sort out the takeover cases caused by human error, and then can further conduct a detailed study on the cases, solving the technical problem of the inconsistent judgment criteria for takeover behaviors in existing autonomous driving vehicles.

[0029] In the description and claims of the present invention and the above drawings, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" or "having" and any variation thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0030] For ease of understanding, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , the first embodiment of the method for identifying a takeover case of mistake in the embodiments of the present invention includes:

[0031] 101. Simulate the historical takeover cases of mistake in the historical takeover case set of mistakes through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical takeover case of mistake;

[0032] Specifically, it is necessary to save the predicted trajectory of the host vehicle in the next 3 seconds given by the algorithm model in the frame before takeover, and this trajectory is composed of 15 position coordinate points every 0.2 seconds; and the trajectory of the host vehicle manually driven by the driver within 3 seconds after takeover, and this trajectory is composed of 30 position coordinate points every 0.1 seconds.

[0033] 102. Calculate a path difference threshold according to the predicted trajectories and the corresponding actual trajectories corresponding to all historical takeover cases of mistake;

[0034] Specifically, traverse each position coordinate point p0=(x0, y0), p1, p2, …, pn in the predicted trajectory of the host vehicle in the next 3 seconds given by the pre-takeover algorithm, and find the position coordinate p'(n*2)=(x'(n*2), y'(n*2)) of the host vehicle during the driver's manual driving at the corresponding moment to calculate the distance difference and the angle difference. Calculate the distance difference and the angle difference in each historical case manually marked as a takeover due to human error, and statistically obtain a reasonable threshold for the distance difference and the angle difference.

[0035] 103. Identify the takeover cases due to human error for the takeover cases to be identified according to the path difference threshold.

[0036] In this embodiment, for the takeover cases to be judged, calculate the distance difference and the angle difference according to the formula in S2. If the distance difference < the threshold or the angle difference < the threshold, then judge this takeover case as a human error.

[0037] Specifically, there may be various reasons for a takeover. Among them, there may be a situation where the driver takes over in advance due to excessive caution, but in fact, if the vehicle continues to drive autonomously, there will be no safety problems. There may also be a situation where the driver subjectively believes that the driving speed of the autonomous vehicle is too slow, and thus takes over to accelerate, etc. The existing method for screening out such takeovers due to human error is to let annotators or engineers manually annotate. The annotator or engineer will judge whether a takeover is a human error according to the video record of the takeover, the driver's recording, and the visual simulation of the vehicle's internal and external states.

[0038] In this embodiment, simulate the historical takeover cases due to human error in the historical takeover case set of human errors through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical takeover case due to human error; calculate the path difference threshold according to the predicted trajectory and the corresponding actual trajectory corresponding to all historical takeover cases due to human error, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; identify the takeover cases due to human error for the takeover cases to be identified according to the path difference threshold. By calculating the predicted trajectory of autonomous driving through simulation technology and based on the path difference between the predicted trajectory and the actual trajectory, a unified judgment standard for judging the takeover of a vehicle due to human error is obtained, which can effectively and quickly sort out the takeover cases caused by human error, and then can further conduct a sub-study on the cases, solving the technical problem of the inconsistent judgment criteria for takeover behaviors in existing autonomous vehicles.

[0039] Please refer to Figure 2 The second embodiment of the method for identifying takeover cases due to human error in the embodiment of the present invention includes:

[0040] 201. Obtain the vehicle state information of each historical takeover case due to human error in the historical takeover case set of human errors within a preset time period before the manual takeover.

[0041] In this embodiment, the collected historical takeover cases of mistakes are all the takeover by the original system. When the vehicle is in the autonomous driving state, the driving data of the vehicle and the operation data of the user during the period when the driving authority is transferred to the user are combined. By obtaining the corresponding historical takeover cases of mistakes, further simulation is carried out.

[0042] 202. Input the vehicle state information of each historical takeover case of mistakes into the simulation algorithm for simulation to obtain the corresponding predicted trajectory.

[0043] In this embodiment, by extracting the specific parameters from the cases that have been manually classified as historical takeover cases of physical objects, and inputting the parameters into the simulation model for simulation, the trajectory that the vehicle will continue to travel under the control of the program if not taken over manually, that is, the predicted trajectory, can be obtained.

[0044] Specifically, the relevant data of autonomous driving, the relevant data of manual driving, and the road surface data before and after takeover are used as the input parameters for the simulation.

[0045] 203. Calculate the path difference threshold according to the predicted trajectory and the corresponding actual trajectory corresponding to all historical takeover cases of mistakes.

[0046] 204. Calculate the distance difference and angle difference of the takeover case to be identified.

[0047] In this embodiment, by calculating the relevant data of autonomous driving, the relevant data of manual driving, and the road surface data before and after takeover in the takeover case to be identified as the input parameters for the simulation, and inputting them into the simulation model, two sets of data of the distance difference and angle difference between the specific manual and autonomous driving are obtained. These two sets of data are used for comparison in the next step.

[0048] 205. Judge whether the distance difference of the takeover case to be identified is greater than the distance difference threshold or whether the angle difference of the takeover case to be identified is greater than the angle difference threshold.

[0049] Specifically, it is also possible to simply judge by only judging one of the distance difference or angle difference data sets, which can realize a relatively simple preliminary classification process. Then, on the basis of the first classification, a complete second classification is carried out. Through the two classification results of the first classification and the second classification, the threshold size in the simulation is further adjusted.

[0050] 206. If the distance difference of the takeover case to be identified is greater than the distance difference threshold or the angle difference of the takeover case to be identified is greater than the angle difference threshold, then the takeover case to be identified is identified as a takeover case of mistake.

[0051] In this embodiment, a comparison result is obtained through step 205. Here, the comparison result can be based on the angle difference, the distance difference, or when both the angle difference and the distance difference exceed the threshold obtained by simulation.

[0052] Specifically, after obtaining the comparison result, classify the takeover cases to be recognized based on the comparison result. After classification, it is used for further refined analysis of the cases.

[0053] Based on the previous embodiment, this embodiment details the calculation of the distance difference and angle difference of the takeover cases to be recognized; determines whether the distance difference of the takeover cases to be recognized is greater than the distance difference threshold or whether the angle difference of the takeover cases to be recognized is greater than the angle difference threshold; if the distance difference of the takeover cases to be recognized is greater than the distance difference threshold or the angle difference of the takeover cases to be recognized is greater than the angle difference threshold, then the process of recognizing the takeover cases to be recognized as misoperation takeover cases. Compared with the traditional method through this embodiment, the specific judgment process in case classification is refined, and the classification process of the takeover cases to be recognized is automated through the judgment method.

[0054] Please refer to Figure 3 , the third embodiment of the misoperation takeover case recognition method in the embodiment of the present invention includes:

[0055] 301. Simulate the historical misoperation takeover cases in the historical misoperation takeover case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical misoperation takeover case;

[0056] 302. Determine the actual trajectory of each historical misoperation takeover case within a preset time period after manual takeover;

[0057] In this embodiment, by extracting the actual running trajectory of the vehicle within a preset time period after manual takeover in the historical misoperation takeover cases, various specific quantifiable analysis parameter values such as path change, speed change, and angle change generated during the vehicle operation are obtained during this time period.

[0058] 303. Determine the first set of position coordinate points of the predicted trajectory within a preset time period after manual takeover;

[0059] In this embodiment, the first set of position coordinate points can be the main vehicle trajectory manually driven by the driver within 3 seconds after takeover, and this trajectory is composed of 30 position coordinate points at every 0.1 second.

[0060] Specifically, due to certain differences in the performance, acquisition frequency, and acquisition settings of the data acquisition devices for different types and models of vehicles, this embodiment does not limit the acquisition time and acquisition frequency of the acquired first set of position coordinate points, and only explains the operation method.

[0061] 304. Determine the second set of position coordinate points of the actual trajectory within a preset time period after manual takeover;

[0062] In this embodiment, the second set of position coordinate points is the predicted trajectory of the host vehicle in the next 3 seconds given by the algorithm model before takeover, and this trajectory is composed of 15 position coordinate points at intervals of 0.2 seconds each.

[0063] Specifically, due to certain differences in performance, output frequency, and output settings among the automatic driving devices of different vehicle types and models, this embodiment does not limit the output time and output frequency of the output second set of position coordinate points, and only explains the operation method.

[0064] 305. Split the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points;

[0065] Specifically, split the obtained set of coordinate points into x1 = (x1 / x2 / x3...xi) and y = (y1 / y2 / y3...yi).

[0066] 306. Split the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; specifically, split the obtained set of coordinate points into x' = (x'1 / x'2 / x'3...x'i) and y' = (y'1 / y'2 / y'3...y'i).

[0067] 307. Add the square of the difference between the first abscissa point and the second abscissa point at the same moment to the square of the difference between the first ordinate point and the second ordinate point at the same moment, and then take the square root of the sum to obtain the square root result;

[0068] Specifically, the distance difference formula is as follows:

[0069]

[0070] 308. Accumulate all the square root results within a preset time period to obtain the distance difference threshold;

[0071] Specifically, by calculating multiple cases that have been classified as failed takeover cases, obtain the threshold that can define whether it is a failed takeover.

[0072] In this embodiment, by statistically analyzing the cases, define the range with commonalities in the cases, extract its parameters, and obtain the distance difference threshold.

[0073] 309. Statistically analyze the path differences of all historical failed takeover cases to obtain the path difference threshold;

[0074] 310. Identify failed takeover cases for the takeover cases to be identified according to the path difference threshold.

[0075] On the basis of the previous embodiment, this embodiment details the process of splitting the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; splitting the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; adding the square of the difference between the first abscissa point and the second abscissa point and the square of the difference between the first ordinate point and the second ordinate point at the same moment, and then taking the square root of the sum to obtain the square root result; and accumulating all the square root results within a preset time period to obtain the distance difference threshold. Compared with the traditional method, this embodiment refines the calculation process of the distance difference, making the identification method for takeover failure cases clearer and easier to understand.

[0076] Please refer to Figure 4 , the fourth embodiment of the takeover failure case identification method in the embodiment of the present invention includes:

[0077] 401. Simulate the historical takeover failure cases in the historical takeover failure case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical takeover failure case;

[0078] 402. Determine the actual trajectory of each historical takeover failure case within a preset time period after manual takeover;

[0079] 403. Determine the first set of position coordinate points of the predicted trajectory within a preset time period after manual takeover;

[0080] 404. Determine the second set of position coordinate points of the actual trajectory within a preset time period after manual takeover;

[0081] 405. Split the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points;

[0082] 406. Split the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points;

[0083] 407. Calculate the difference between consecutive second ordinate points within a preset time period divided by the difference between consecutive second abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the actual trajectory angle;

[0084] 408. Calculate the difference between consecutive first ordinate points within a preset time period divided by the difference between consecutive first abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the predicted trajectory angle;

[0085] 409. Take the absolute value after subtracting the predicted trajectory angle from the actual trajectory angle to obtain the angle difference threshold;

[0086] Specifically, the angle difference formula is as follows:

[0087]

[0088] 410. Statistically analyze the path differences of all historical error takeover cases to obtain a path difference threshold.

[0089] 411. Identify error takeover cases for the takeover cases to be identified according to the path difference threshold.

[0090] Based on the previous embodiments, this embodiment details the process of splitting the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; splitting the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; calculating the difference between consecutive second ordinate points within a preset time period divided by the difference between consecutive second abscissa points within the preset time period, and then taking the arctangent of the calculation result to obtain the actual trajectory angle; calculating the difference between consecutive first ordinate points within a preset time period divided by the difference between consecutive first abscissa points within the preset time period, and then taking the arctangent of the calculation result to obtain the predicted trajectory angle; subtracting the predicted trajectory angle from the actual trajectory angle and taking the absolute value to obtain the angle difference threshold. Compared with traditional methods, this embodiment refines the calculation process of the angle difference, making the identification method for error takeover cases clearer and easier to understand.

[0091] The above describes the error takeover case identification method in the embodiments of the present invention. Next, the error takeover case identification device in the embodiments of the present invention will be described. Please refer to Figure 5 One embodiment of the error takeover case identification device in the embodiments of the present invention includes:

[0092] A case simulation module 501, configured to simulate the historical error takeover cases in the historical error takeover case set through a preset simulation algorithm to obtain a predicted trajectory corresponding to each historical error takeover case.

[0093] A threshold calculation module 502, configured to calculate a path difference threshold according to the predicted trajectories and the corresponding actual trajectories corresponding to all historical error takeover cases.

[0094] A case identification module 503, configured to identify error takeover cases for the takeover cases to be identified according to the path difference threshold.

[0095] In the embodiments of the present invention, the error takeover case recognition device runs the above error takeover case recognition method, including simulating the historical error takeover cases in the historical error takeover case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical error takeover case; calculating a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical error takeover cases, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; and recognizing the takeover case to be recognized as an error takeover case according to the path difference threshold. By calculating the predicted trajectory of autonomous driving through simulation technology and based on the path difference between the predicted trajectory and the actual trajectory, a unified judgment criterion for judging an error takeover vehicle is obtained, which can effectively and quickly sort out the takeover cases caused by human errors, and then can further conduct a detailed study on the cases, solving the technical problem of the inconsistent judgment criteria for takeover behaviors in existing autonomous driving vehicles.

[0096] Please refer to Figure 6 , the second embodiment of the error takeover case recognition device in the embodiments of the present invention includes:

[0097] A case simulation module 501, configured to simulate the historical error takeover cases in the historical error takeover case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical error takeover case;

[0098] A threshold calculation module 502, configured to calculate a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical error takeover cases;

[0099] A case recognition module 503, configured to recognize the takeover case to be recognized as an error takeover case according to the path difference threshold.

[0100] In this embodiment, the case simulation module 501 is specifically configured to:

[0101] Obtain the vehicle state information of each historical error takeover case in the historical error takeover case set within a preset time period before the manual takeover; input the vehicle state information of each historical error takeover case into the simulation algorithm for simulation to obtain the corresponding predicted trajectory.

[0102] In this embodiment, the threshold calculation module 502 is specifically configured to:

[0103] A trajectory determination unit 5021 determines the actual trajectory of each historical error takeover case within a preset time period after the manual takeover; a path difference calculation unit 5022 calculates the path difference between the actual trajectory and the corresponding predicted trajectory of each historical error takeover case after the manual takeover, where the path difference includes a distance difference and / or an angle difference; a threshold obtaining unit 5023 statistically analyzes the path differences of all historical error takeover cases to obtain a path difference threshold.

[0104] In this embodiment, the path difference calculation unit 5022 is specifically configured to: a first set subunit 50221 determines a first set of position coordinate points of the predicted trajectory within a preset time period after manual takeover; a second set subunit 50222 determines a second set of position coordinate points of the actual trajectory within a preset time period after manual takeover; a path difference calculation subunit 50223 calculates a corresponding path difference threshold based on the first set of position coordinate points and the second set of position coordinate points of each historical failed takeover case.

[0105] In this embodiment, the path difference calculation subunit 50223 is specifically configured to:

[0106] Split the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; split the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; add the square of the difference between the first abscissa point and the second abscissa point at the same moment to the square of the difference between the first ordinate point and the second ordinate point, and then take the square root of the sum to obtain a square root result; accumulate all the square root results within the preset time period to obtain the distance difference threshold.

[0107] In this embodiment, the path difference calculation subunit 50223 is specifically configured to:

[0108] Split the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; split the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; calculate the difference between consecutive second ordinate points within the preset time period divided by the difference between consecutive second abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the actual trajectory angle; calculate the difference between consecutive first ordinate points within the preset time period divided by the difference between consecutive first abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the predicted trajectory angle; take the absolute value after subtracting the predicted trajectory angle from the actual trajectory angle to obtain the angle difference threshold.

[0109] In this embodiment, the case recognition module 503 is specifically configured to:

[0110] Calculate the distance difference and the angle difference of the takeover case to be recognized; determine whether the distance difference of the takeover case to be recognized is greater than the distance difference threshold or whether the angle difference of the takeover case to be recognized is greater than the angle difference threshold; if the distance difference of the takeover case to be recognized is greater than the distance difference threshold or the angle difference of the takeover case to be recognized is greater than the angle difference threshold, then recognize the takeover case to be recognized as a failed takeover case.

[0111] Based on the previous embodiment, this embodiment describes in detail the specific functions of each module and the unit composition of some modules. By refining the specific functions of the units and subunits within each module, the operation of the fault takeover case recognition device is improved, its reliability during operation is enhanced, and the actual logic between each step is clarified, thereby improving the practicality of the device.

[0112] Above Figure 5 And Figure 6 The fault takeover case recognition device in the embodiments of the present invention has been described in detail from the perspective of modular functional entities. Next, the fault takeover case recognition device in the embodiments of the present invention will be described in detail from the perspective of hardware processing.

[0113] Figure 7 FIG. is a schematic structural diagram of a fault takeover case recognition device provided by an embodiment of the present invention. The fault takeover case recognition device 700 may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPU) 710 (for example, one or more processors) and a memory 720, and one or more storage media 730 for storing application programs 733 or data 732 (for example, one or more mass storage devices). Among them, the memory 720 and the storage media 730 may be transient storage or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the fault takeover case recognition device 700. Further, the processor 710 may be configured to communicate with the storage media 730 and execute a series of instruction operations in the storage media 730 on the fault takeover case recognition device 700 to implement the steps of the above-mentioned fault takeover case recognition method.

[0114] The fault takeover case recognition device 700 may further include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input / output interfaces 760, and / or one or more operating systems 731, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that Figure 7 The shown structural diagram of the fault takeover case recognition device does not limit the fault takeover case recognition device provided in this application, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0115] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is caused to execute the steps of the error takeover case recognition method described above.

[0116] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, or units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0117] 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 such an understanding, the technical solution of the present invention, in essence, 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 to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0118] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for identifying a failover case, characterized in that, The described failure takeover case recognition method includes: Simulating the historical failure takeover cases in the historical failure takeover case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical failure takeover case; Calculating a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical failure takeover cases, wherein the path difference threshold includes a distance difference threshold and / or an angle difference threshold; Identifying the takeover case to be recognized as a failure takeover case according to the path difference threshold; The identifying the takeover case to be recognized as a failure takeover case according to the path difference threshold includes: Calculating the distance difference and the angle difference of the takeover case to be recognized; judging whether the distance difference of the takeover case to be recognized is greater than the distance difference threshold or whether the angle difference of the takeover case to be recognized is greater than the angle difference threshold; if the distance difference of the takeover case to be recognized is less than the distance difference threshold or the angle difference of the takeover case to be recognized is less than the angle difference threshold, then identifying the takeover case to be recognized as a failure takeover case; The calculating a path difference threshold according to the predicted trajectories and the corresponding actual trajectories of all historical failure takeover cases, wherein the path difference threshold includes a distance difference threshold and / or an angle difference threshold, includes: determining the actual trajectory of each historical failure takeover case within a preset time period after manual takeover; calculating the path difference between the actual trajectory and the corresponding predicted trajectory of each historical failure takeover case after manual takeover, and statistically analyzing the path differences of all historical failure takeover cases to obtain the path difference threshold; The calculating the path difference between the actual trajectory and the corresponding predicted trajectory of each historical failure takeover case after manual takeover, and statistically analyzing the path differences of all historical failure takeover cases to obtain the path difference threshold, wherein the path difference includes a distance difference and / or an angle difference, includes: determining a first set of position coordinate points of the predicted trajectory within a preset time period after manual takeover; determining a second set of position coordinate points of the actual trajectory within a preset time period after manual takeover; calculating the corresponding path difference threshold based on the first set of position coordinate points and the second set of position coordinate points of each historical failure takeover case.

2. The method for identifying a failure takeover case according to claim 1, wherein, The simulating the historical failure takeover cases in the historical failure takeover case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical failure takeover case includes: Obtaining the vehicle state information of each historical failure takeover case in the historical failure takeover case set within a preset time period before manual takeover; Inputting the vehicle state information of each historical failure takeover case into the simulation algorithm for simulation to obtain the corresponding predicted trajectory.

3. The method for identifying a failure takeover case according to claim 1, characterized in that The calculating the corresponding path difference threshold based on the first set of position coordinate points and the second set of position coordinate points of each historical failure takeover case includes: Splitting the first set of position coordinate points into a first set of abscissa points and a first set of ordinate points; Splitting the second set of position coordinate points into a second set of abscissa points and a second set of ordinate points; Add the square of the difference between the first abscissa point and the second abscissa point at the same moment to the square of the difference between the first ordinate point and the second ordinate point, and then take the square root of the sum to obtain the square root result; Accumulate all the square root results within a preset time period to obtain the distance difference threshold.

4. The method for identifying a fault takeover case according to claim 1, characterized in that, Calculating the corresponding path difference threshold based on the first position coordinate point set and the second position coordinate point set of each historical error takeover case includes: Split the first position coordinate point set into a first abscissa point set and a first ordinate point set; Split the second position coordinate point set into a second abscissa point set and a second ordinate point set; Calculate the difference between consecutive second ordinate points within a preset time period divided by the difference between consecutive second abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the actual trajectory angle; Calculate the difference between consecutive first ordinate points within a preset time period divided by the difference between consecutive first abscissa points within the preset time period, and then take the arctangent of the calculation result to obtain the predicted trajectory angle; Take the absolute value after subtracting the predicted trajectory angle from the actual trajectory angle to obtain the angle difference threshold.

5. A failure takeover case recognition device, characterized in that The error takeover case recognition device includes: A case simulation module for simulating the historical error takeover cases in the historical error takeover case set through a preset simulation algorithm to obtain the predicted trajectory corresponding to each historical error takeover case; A threshold calculation module for calculating a path difference threshold according to the predicted trajectory and the corresponding actual trajectory of all historical error takeover cases, where the path difference threshold includes a distance difference threshold and / or an angle difference threshold; A case recognition module for recognizing an error takeover case for the takeover case to be recognized according to the path difference threshold; The case recognition module is specifically used for: calculating the distance difference and the angle difference of the takeover case to be recognized; judging whether the distance difference of the takeover case to be recognized is greater than the distance difference threshold or whether the angle difference of the takeover case to be recognized is greater than the angle difference threshold; if the distance difference of the takeover case to be recognized is less than the distance difference threshold or the angle difference of the takeover case to be recognized is less than the angle difference threshold, then recognize the takeover case to be recognized as an error takeover case; The threshold calculation module includes: a trajectory determination unit for determining the actual trajectory of each historical error takeover case within a preset time period after manual takeover; a path difference calculation unit for calculating the path difference between the actual trajectory and the corresponding predicted trajectory of each historical error takeover case after manual takeover, where the path difference includes a distance difference and / or an angle difference; a threshold obtaining unit for statistically analyzing the path differences of all historical error takeover cases to obtain a path difference threshold; The path difference calculation unit includes: a first set subunit for determining a first position coordinate point set of the predicted trajectory within a preset time period after manual takeover; a second set subunit for determining a second position coordinate point set of the actual trajectory within a preset time period after manual takeover; a path difference calculation subunit for calculating the corresponding path difference threshold based on the first position coordinate point set and the second position coordinate point set of each historical error takeover case.

6. An electronic device, characterized in that, The electronic device includes: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected by a line; The at least one processor calls the instructions in the memory to cause the electronic device to execute each step of the error takeover case recognition method according to any one of claims 1-4.

7. A computer-readable storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by a processor, it implements each step of the error takeover case recognition method according to any one of claims 1-4.

Images