A safety capability detection method and device for an autonomous vehicle

Abstract

A safety capability detection method and device for an autonomous vehicle, the method comprising: when there is a collision risk between the autonomous vehicle and a first vehicle, obtaining the response control force output by the decision-making of the autonomous vehicle, wherein the first vehicle is the vehicle in the current lane where the autonomous vehicle is traveling and closest to the autonomous vehicle in front; at a first moment when the autonomous vehicle makes a braking response, determining the motion change parameters of the autonomous vehicle, as well as the relative speed and relative distance between the autonomous vehicle and the first vehicle; according to the determined motion change parameters, relative speed and relative distance, generating a dual margin spectrum by using a preset dual margin spectrum model, wherein the dual margin spectrum consists of a time margin spectrum and a control margin spectrum; determining the safety capability data of the autonomous vehicle according to the generated dual margin spectrum.

Description

Technical Field

[0001] This document relates to the technical field of vehicle autonomous driving, and particularly to a method and device for detecting the safety capabilities of an autonomous vehicle. Background Art

[0002] Safety is of utmost importance for autonomous vehicles. Their safety evaluation is a crucial link in the iterative development and feedback verification of intelligent vehicles. Inadequate safety evaluation is difficult to detect potential safety hazards and capability deficiencies in intelligent vehicles, and is likely to lead to unexpected collision accidents. In some technologies, the safety capabilities of autonomous driving are only verified based on whether a collision occurs, lacking a deep understanding and analysis of the responses of the autonomous driving system, and lacking an accurate observation model for the differentiation and flexibility of the responses of the autonomous driving system. Summary of the Invention

[0003] This application provides a method and device for detecting the safety capabilities of an autonomous vehicle, which solves the problems of a single verification process for autonomous driving safety capabilities, lack of rationality, and low flexibility.

[0004] This application provides a method for detecting the safety capabilities of an autonomous vehicle, the method comprising:

[0005] When there is a collision risk between the autonomous vehicle and a first vehicle, obtaining the response control force output by the decision-making of the autonomous vehicle, wherein the first vehicle is the vehicle in the current lane where the autonomous vehicle is traveling and closest to the autonomous vehicle in front;

[0006] At a first moment when the autonomous vehicle makes a braking response, determining the motion change parameters of the autonomous vehicle, as well as the relative speed and relative distance between the autonomous vehicle and the first vehicle;

[0007] According to the determined motion change parameters, relative speed and relative distance, generating a dual margin spectrum by using a preset dual margin spectrum model, wherein the dual margin spectrum consists of a time margin spectrum and a control margin spectrum;

[0008] Determining the safety capability data of the autonomous vehicle according to the generated dual margin spectrum.

[0009] In an exemplary embodiment, the time margin spectrum is a spectrum with the scene urgency as the abscissa and the time margin as the ordinate;

[0010] The control margin spectrum is a spectrum with the scene urgency as the abscissa and the operation margin as the ordinate.

[0011] In an exemplary embodiment, at a first moment when the autonomous vehicle makes a braking response, determining the motion change parameters of the autonomous vehicle includes:

[0012] At the first moment when the autonomous vehicle makes a braking response, the speed, position, and actual response time of the autonomous vehicle at the first moment are obtained by using the autonomous vehicle control platform.

[0013] In an exemplary embodiment, the generating of the dual margin spectrum by using the preset dual margin spectrum model according to the determined motion change parameter, relative speed, and relative distance includes:

[0014] Calculating the time response margin of the autonomous vehicle by using the response time margin model;

[0015] Calculating the braking margin of the autonomous vehicle by using the handling margin model;

[0016] Generating a dual margin spectrum according to the obtained response margin spectrum curve and handling margin spectrum curve in the time dimension.

[0017] In an exemplary embodiment, the response time margin model is:

[0018] Tm = (Sd - Sa) / V;

[0019] This response time margin model is the time difference Tm between the first moment Td and the preset theoretical latest response moment Ta;

[0020] Wherein, Td is the moment when the autonomous vehicle makes a braking response, Sd is the relative distance between the autonomous vehicle and the first vehicle corresponding to the first moment, Sa is the relative distance between the autonomous vehicle and the first vehicle corresponding to the theoretical latest response moment, and V is the running speed of the autonomous vehicle at the first moment.

[0021] In an exemplary embodiment, the relative distance Sa between the autonomous vehicle and the first vehicle corresponding to the theoretical latest response moment Ta is the distance required for the autonomous vehicle to brake with the maximum braking intensity and the relative speed to just decrease to 0; the Sa is determined by the following method:

[0022] Sa = (V - V0) * (V - V0) / 2 / a0;

[0023] Wherein, V is the running speed of the autonomous vehicle at the first moment, V0 is the speed of the first vehicle at the first moment, and a0 is the maximum braking intensity of the vehicle on the current road surface.

[0024] In an exemplary embodiment, the handling margin model: am = a0 - a;

[0025] Wherein, a is the response handling force output by the decision of the autonomous vehicle at the first moment Td, a0 is the maximum braking intensity of the vehicle on the current road surface, and am is the braking margin of the collision avoidance ability;

[0026] If am > 0, it means there is a margin in the control margin;

[0027] If am = 0, it means full braking;

[0028] If am < 0, it means that the maximum braking intensity of the vehicle on the current road surface is less than the collision avoidance braking intensity a expected to be applied by the autonomous driving, and a collision will necessarily occur under this working condition.

[0029] In an exemplary embodiment, the scenario urgency is: C = 1 / TTC, TTC = Sd / (V - V0);

[0030] Or C = 1 / HeadWay, HeadWay = Sd / V;

[0031] Wherein, Sd is the relative distance between the autonomous driving vehicle and the first vehicle at the first moment, V is the speed of the autonomous driving vehicle at the first moment, and V0 is the speed of the first vehicle at the first moment.

[0032] In an exemplary embodiment, determining the safety capability data of the autonomous driving vehicle according to the generated dual margin spectrum includes:

[0033] According to the generated dual margin spectrum, determine the interval on the continuous scale of the response capability of the autonomous driving vehicle when the collision avoidance control margin value is zero;

[0034] Calculate the total area of the response time margin corresponding to this interval;

[0035] Wherein, this total area represents the cumulative evaluation value of the safety capability of the autonomous driving vehicle on the continuous scale.

[0036] This application also provides a safety capability detection device for an autonomous driving vehicle. The memory is used to save the program for detecting the safety capability of the autonomous driving vehicle, and the processor is used to read and execute the program for detecting the safety capability of the autonomous driving vehicle, and execute the method described in any one of the above embodiments.

[0037] Compared with the related art, the present application provides a method and a device for detecting the safety ability of an autonomous vehicle. The method includes: when there is a collision risk between the autonomous vehicle and a first vehicle, obtaining the response control force output by the decision-making of the autonomous vehicle, where the first vehicle is the vehicle closest to the autonomous vehicle in the current lane where the autonomous vehicle is traveling; at a first moment when the autonomous vehicle makes a braking response, determining the motion change parameters of the autonomous vehicle, as well as the relative speed and relative distance between the autonomous vehicle and the first vehicle; according to the determined motion change parameters, relative speed and relative distance, generating a dual margin spectrum by using a preset dual margin spectrum model, where the dual margin spectrum is composed of a time margin spectrum and a control margin spectrum; determining the safety ability data of the autonomous vehicle according to the generated dual margin spectrum. Through the technical solution of the present invention, the problem that the verification process of the autonomous driving safety ability is single, lacks rationality and has low flexibility is solved.

[0038] Other features and advantages of the present application will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present application. Other advantages of the present application can be realized and obtained through the solutions described in the specification and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings are used to provide an understanding of the technical solutions of the present application, and constitute a part of the specification. They are used together with the embodiments of the present application to explain the technical solutions of the present application, and do not constitute a limitation to the technical solutions of the present application.

[0040] Figure 1 It is a flowchart of the method for detecting the safety ability of the autonomous vehicle according to the embodiment of the present application;

[0041] Figure 2 It is a schematic diagram of the device for detecting the safety ability of the autonomous vehicle according to the embodiment of the present application;

[0042] Figure 3 It is a structural block diagram of the method for evaluating the autonomous driving safety ability based on random response in some exemplary embodiments;

[0043] Figure 4 It is a schematic flowchart of the method for evaluating the autonomous driving safety ability based on random response in some exemplary embodiments;

[0044] Figure 5 It is a schematic diagram of the safety evaluation test scenario in some exemplary embodiments;

[0045] Figure 6 It is a working condition simulation diagram of the safety evaluation in some exemplary embodiments;

[0046] Figure 7It is a time node diagram of the safety evaluation process in some exemplary embodiments;

[0047] Figure 8 It is a flowchart of a method for evaluating the safety capabilities of an autonomous vehicle in some exemplary embodiments;

[0048] Figure 9 It is a schematic diagram of the single response moment line of an autonomous vehicle in some exemplary embodiments;

[0049] Figure 10 It is an example of the distribution of the actual response lines under multiple random scenarios in some exemplary embodiments;

[0050] Figure 11 It is a schematic illustration in a coordinate system of the physical meanings of several pairs of time margins / maneuver margins in some exemplary embodiments;

[0051] Figure 12 It is a schematic diagram of the double - margin response model of an autonomous vehicle and its safety - capability bandwidth value in some exemplary embodiments. Detailed implementation manners

[0052] This application describes multiple embodiments, but the description is exemplary rather than restrictive, and it will be apparent to those of ordinary skill in the art that there can be more embodiments and implementation solutions within the scope of the embodiments described in this application. Although many possible combinations of features are shown in the drawings and discussed in the detailed implementation manners, many other combinations of the disclosed features are also possible. Unless specifically restricted, any feature or element of any embodiment can be combined with any other feature or element in any other embodiment, or can replace any other feature or element in any other embodiment.

[0053] This application includes and contemplates combinations with features and elements known to those of ordinary skill in the art. The embodiments, features, and elements disclosed in this application can also be combined with any conventional features or elements to form unique inventive solutions defined by the claims. Any feature or element of any embodiment can also be combined with features or elements from other inventive solutions to form another unique inventive solution defined by the claims. Therefore, it should be understood that any feature shown and / or discussed in this application can be implemented alone or in any suitable combination. Therefore, except for the limitations made according to the appended claims and their equivalent replacements, the embodiments are not subject to other limitations. In addition, various modifications and changes can be made within the scope of the protection of the appended claims.

[0054] In addition, when describing representative embodiments, the specification may have presented the method and / or process as a specific sequence of steps. However, to the extent that the method or process does not depend on the specific order of the steps described herein, the method or process should not be limited to the specific order of steps described. As will be understood by those of ordinary skill in the art, other step orders are possible. Therefore, the specific order of steps set forth in the specification should not be construed as a limitation on the claims. In addition, the claims directed to the method and / or process should not be limited to performing their steps in the order written, as those skilled in the art can readily understand that these orders can vary and still remain within the spirit and scope of the embodiments of the present application.

[0055] Embodiments of the present disclosure provide a method for detecting the safety capabilities of an autonomous vehicle, as Figure 1 shown, the method includes steps S100 - S130, specifically as follows:

[0056] S100. When there is a risk of collision between the autonomous vehicle and a first vehicle, obtain the response control force output by the decision of the autonomous vehicle, where the first vehicle is the vehicle in the current lane in front of the autonomous vehicle that is closest to the autonomous vehicle;

[0057] S110. At the first moment when the autonomous vehicle makes a braking response, determine the motion change parameters of the autonomous vehicle, as well as the relative speed and relative distance between the autonomous vehicle and the first vehicle;

[0058] S120. According to the determined motion change parameters, relative speed, and relative distance, generate a dual - margin spectrum using a pre - set dual - margin spectrum model, where the dual - margin spectrum consists of a time - margin spectrum and a control - margin spectrum;

[0059] S130. Determine the safety - capability data of the autonomous vehicle according to the generated dual - margin spectrum.

[0060] In this embodiment, in most autonomous driving scenarios, collision is the safety bottom line of an autonomous vehicle. In this embodiment, it is assumed that this process is a complex non - linear system. Initially, it is in a safe state. After a certain moment of movement, danger is sensed. Within the safe operable space, there are two extreme safety - braking forms: one is to respond immediately at the earliest response - time line, decelerate and brake with a smaller braking force, and then enter a new safe state relative to the vehicle in front; the other is to respond immediately at the latest response - time line, brake with the maximum braking force, and then enter a new safe state relative to the vehicle in front.

[0061] Physically, as Figure 6As shown, the earliest response time line is based on the real-time speed / position at the current calculation time. As long as a collision risk can be perceived, it can be considered as the earliest response time line; while the latest response time line is actually calculated based on physical quantities such as relative distance and relative speed, and is related to the maximum collision avoidance braking ability a0 of the vehicle itself.

[0062] In this embodiment, the first vehicle is the vehicle in the current lane closest to the autonomous vehicle in front of the autonomous vehicle, and the speed of the first vehicle is less than the operating speed of the autonomous vehicle.

[0063] The response control force output by the decision-making of the autonomous vehicle is output from the control platform of the autonomous vehicle. Determining the response control force is determined by the control platform of the autonomous vehicle according to the parameters of the current vehicle operation.

[0064] In an exemplary embodiment, the time amount between the actual response time line and the latest response time line of the autonomous vehicle is defined as the response time margin, which reflects a calm response ability. Under different working conditions, the actual response time line of the autonomous vehicle is different, and at this time, the response time margin is normalized. The time margin spectrum is a spectrum with the scene urgency as the abscissa and the time margin as the ordinate.

[0065] The expected control ability at the actual response time of the autonomous vehicle is a, then the collision avoidance control margin is defined as a0 - a. Among them, if the collision avoidance control margin is positive, it reflects a safe ability with ease; if the margin value is zero, it represents an emergency full braking, with a slightly tense safe ability; if the margin value is negative, it represents that the actual collision avoidance ability of the vehicle is less than the expected collision avoidance control ability applied, and at this time, the actual response time line has moved to the left of the latest response time line, and a collision is inevitable. The control margin spectrum is a spectrum with the scene urgency as the abscissa and the operation margin as the ordinate.

[0066] In an exemplary embodiment, combining the physical definitions of the response time margin and the collision avoidance control ability margin, an offline characterization of the multi-dimensional safety margin can be obtained. Then, combining the numerical interpolation fitting method, the reciprocal of the Headway or the reciprocal of the TTC (Time to Collision) between the autonomous vehicle and the obstacle in front can be used to represent the urgency of the working condition, and using these two margins as the ordinates respectively, an evaluation result of the random response time of the safety ability of the autonomous vehicle can be formed. Among them, the scene urgency is: C = 1 / TTC, TTC = Sd / (V - V0); or C = 1 / HeadWay, HeadWay = Sd / V; where Sd is the relative distance between the autonomous vehicle and the first vehicle at the first moment, V is the speed of the autonomous vehicle at the first moment, and V0 is the speed of the first vehicle at the first moment.

[0067] In an exemplary embodiment, determining the motion change parameters of the autonomous vehicle at the first moment when the autonomous vehicle makes a braking response includes: at the first moment when the autonomous vehicle makes a braking response, using the autonomous vehicle control platform to obtain the speed, position, and actual response time of the autonomous vehicle at the first moment.

[0068] In an exemplary embodiment, generating a dual margin spectrum using a pre-set dual margin spectrum model according to the determined motion change parameters, relative speed, and relative distance includes: calculating the time response margin of the autonomous vehicle using a response time margin model; calculating the braking margin of the autonomous vehicle using a handling margin model; generating a dual margin spectrum based on the obtained response margin spectrum curve and handling margin spectrum curve in the time domain. In this embodiment, for a set of scenario parameters, a point pair of response margin and handling margin is obtained, as Figure 9 shown in the schematic diagram of the single response moment line of the autonomous vehicle. The first line from the left is the latest response line; the third line from the left is the earliest response line; the second line from the left is a certain actual response moment line; in a scenario, a point pair of response margin and handling margin is obtained. The physical meaning of each point pair of time margin / handling margin is shown schematically in Figure 11 as shown. The abscissa represents the scenario urgency: C = 1 / TTC or C = 1 / HeadWay. The part of the vertical axis above the abscissa represents the operation time margin corresponding to the response, and the part of the vertical axis below the abscissa represents the collision avoidance ability margin corresponding to the response. The first point pair on the left in the figure represents that in this working condition, the AD has a good time margin and collision avoidance ability margin; in the same type of working condition, the response ability represented by the second point pair on the left is significantly worse than that of the first point pair of the AD system. The third point pair on the left in the figure represents that at the last moment, using the maximum braking force, the collision can just be avoided. The fourth point pair on the left in the figure represents that even using the maximum braking force, a collision will still occur; but if no collision occurs, an additional virtual braking force (with a negative margin) is required. The expected braking force is very large. For the fifth point pair on the left in the figure, considering that the abscissa is more to the right, the more urgent the working condition is and the more likely a collision is to occur; even if it exceeds the inherent ability of the vehicle, the urgency of the situation will lead to an inevitable collision. In this state of testing, the collision avoidance ability margin must be negative, but the time margin is not necessarily negative.

[0069] As Figure 10The distribution of the actual response lines under multiple random scenarios shown, that is, under multiple scenarios, for each scenario, a pair of points of response margin and maneuver margin is obtained, and multiple pairs of points of response margin and maneuver margin are obtained for multiple scenarios; finally, based on multiple pairs of points of response margin and maneuver margin, a response margin spectrum curve and a maneuver margin spectrum curve in the time dimension are formed, and a dual margin spectrum is generated according to the formed response margin spectrum curve and maneuver margin spectrum curve.

[0070] In an exemplary embodiment, the response time margin model is:

[0071] Tm = (Sd - Sa) / V;

[0072] This response time margin model is the time difference Tm between the first moment Td and the preset theoretical latest response moment Ta;

[0073] wherein, Td is the moment when the autonomous vehicle makes a braking response, Sd is the relative distance between the autonomous vehicle and the first vehicle corresponding to the first moment, Sa is the relative distance between the autonomous vehicle and the first vehicle corresponding to the theoretical latest response moment, and V is the running speed of the autonomous vehicle at the first moment. In this embodiment, the relative distance Sa between the autonomous vehicle and the first vehicle corresponding to the theoretical latest response moment Ta is the distance required for the autonomous vehicle to brake with the maximum braking intensity and the relative speed just reduced to 0; the Sa is determined by the following method:

[0074] Sa = (V - V0)*(V - V0) / 2 / a0;

[0075] wherein, V is the running speed of the autonomous vehicle at the first moment, V0 is the speed of the first vehicle at the first moment, and a0 is the maximum braking intensity of the vehicle on the current road surface.

[0076] In an exemplary embodiment, the maneuver margin model: am = a0 - a;

[0077] wherein, a is the response maneuvering force output by the decision of the autonomous vehicle at the first moment Td, a0 is the maximum braking intensity of the vehicle on the current road surface, and am is the braking margin of the collision avoidance ability;

[0078] If am > 0, it means there is a margin in the maneuver margin;

[0079] If am = 0, it means full braking;

[0080] If am < 0, it means that the maximum braking intensity of the vehicle on the current road surface is less than the collision avoidance braking intensity a expected to be applied by the autonomous driving, and a collision will definitely occur under this working condition.

[0081] In an exemplary embodiment, determining the safety capability data of the autonomous vehicle according to the generated dual margin spectrum includes:

[0082] According to the generated dual margin spectrum, when the collision avoidance control margin value is zero, determine the interval on the continuous scale of the response capability of the autonomous vehicle;

[0083] Calculate the total area of the response time margin corresponding to this interval;

[0084] Wherein, this total area represents the cumulative evaluation value of the safety capability of the autonomous vehicle on the continuous scale.

[0085] The embodiment of the present disclosure also provides a traffic signal control device, as Figure 2 shown, the device includes: a memory 210 and a processor 220; the memory 210 is used to save the program for detecting the safety capability of the autonomous vehicle, and the processor 220 is used to read and execute the program for detecting the safety capability of the autonomous vehicle, and execute the method described in any one of the above embodiments.

[0086] Example 1

[0087] Figure 3 Shows a structural block diagram of an autonomous driving safety capability evaluation method based on random response in this embodiment. The evaluation method includes a natural driving full-scenario domain model, a traffic scenario random excitation model, an autonomous vehicle control platform, and a response equivalent dual margin spectrum model, as well as a continuous scale safety capability evaluation;

[0088] The natural driving full-scenario domain model is used to obtain the scene parameters of the target virtual scene;

[0089] The traffic scenario random excitation model is used to obtain the multi-dimensional motion parameters of the vehicle in the target virtual scene;

[0090] The autonomous vehicle control platform is used to feedback multi-motion change parameters such as the vehicle speed and acceleration of the autonomous vehicle;

[0091] The response equivalent dual margin spectrum model is used to generate a response time margin and a collision avoidance control margin by using the change parameters of the autonomous vehicle control platform.

[0092] According to the response time margin and the collision avoidance control margin generated by the response equivalent dual margin spectrum model, determine the cumulative evaluation value of the safety capability of the autonomous vehicle on the continuous scale to evaluate the safety capability of the autonomous vehicle.

[0093] In this example, as Figure 4As shown, the natural driving full-scenario model includes multi-type scenario screening, a scenario vehicle positioning and detection model, extraction of vehicle multi-dimensional motion parameters, modeling of the probability distribution of motion parameters, and description of the complexity of natural driving scenarios.

[0094] The traffic scenario random excitation model includes the perception, evaluation, and decision-making of an autonomous vehicle in different scenarios, combined with the vehicle system dynamics to form an external motion scenario excitation, as well as an impact response model.

[0095] The autonomous vehicle control platform includes the speed, position, and adjustment of lights, rain, and fog of an autonomous vehicle, a virtual instrument measurement and control platform, a VTEHIL experimental platform, an autonomous driving elastic safety envelope, and a parameter-adjustable autonomous driving model.

[0096] The response equivalent double margin spectrum model includes a response time margin, a collision avoidance control margin, and a working condition urgency. A dual expression model of the time response margin and the collision avoidance control margin at the actual response time point is constructed; combined with continuous excitations in different scenarios and the multi-dimensional margin offline characterization combined with numerical interpolation fitting, a safety capability evaluation index system for autonomous vehicles under continuous scales is obtained.

[0097] Example 2

[0098] The following uses an example to illustrate the process of a safety capability detection method for an autonomous vehicle:

[0099] The first step: Set the test scenario for the safety capability detection of the autonomous vehicle, such as Figure 5 As shown in the schematic diagram of the safety capability detection scenario, it includes a test control platform, a virtual working condition platform, an autonomous vehicle, and a vehicle wheel suspension platform. The test control platform is communicatively connected to the virtual working condition platform, the experimental control platform is communicatively connected to the autonomous vehicle, the autonomous vehicle is placed on the vehicle wheel suspension platform, and the autonomous vehicle identifies the scenario in the virtual working condition platform through its own sensors.

[0100] The second step: Set the working condition simulation diagram for the safety capability detection of the autonomous vehicle, such as Figure 6 As shown, various scenarios such as random deceleration of the vehicle in the same lane ahead, vehicle cut-in / cut-out, lane change in the adjacent lane or ramp / own vehicle lateral lane change are considered. The simulation diagram shows the manifestation forms of multiple parameters such as the latest collision-free response time line, the earliest collision-free response time line, the actual response time line, the response time margin, the collision avoidance control margin, the maximum acceleration, the actual response time acceleration, and the elastic domain constrained by the autonomous driving decision parameters.

[0101] The third step: Set the time node diagram for the safety capability detection of the autonomous vehicle, such as Figure 7As shown in the figure, the abscissa represents time, and there are three modules on the ordinate: the experimental control platform, the virtual working condition platform, and the autonomous driving vehicle. The diamond points on each module represent the response of the module at this time, and the arrows represent the paths of response parameter transmission. For example: the first arrow on the left side of the figure indicates that at the first time point, the experimental control platform A sends the selected working condition and sets parameters such as weather, lighting, and road conditions to the virtual working condition platform B; the second arrow on the left side of the figure indicates that at the second time point, the experimental control platform A sends the set speed V of the autonomous driving vehicle to the autonomous driving vehicle C; the third arrow on the left side of the figure indicates that at the third time point, the autonomous driving vehicle C detects the distance L meters from the target vehicle and sends this parameter to the experimental control platform A. At this time, after receiving this parameter, the experimental control platform A makes a response, and at the fourth time point, the parameter of setting the target vehicle speed and starting to decelerate is sent to the virtual working condition platform B; at the fifth time point, the autonomous driving vehicle C feeds back the actual response moment speed and acceleration to the experimental control platform A to calculate the TTC; at the sixth time point, the autonomous driving vehicle C feeds back the response time margin spectrum and the operation margin to the experimental control platform A, and at the sixth time point, the continuous initial safety ability evaluation is calculated. Figure 7 It can intuitively and clearly show the response points and response order of each module in the test scenario.

[0102] Step 4: Establish and train the double margin spectrum model;

[0103] In this step, a double margin spectrum is established, which consists of a time margin spectrum and an operation margin spectrum; the time margin spectrum is a spectrum with the scene urgency as the abscissa and the time margin as the ordinate;

[0104] The operation margin spectrum is a spectrum with the scene urgency as the abscissa and the operation margin as the ordinate.

[0105] As Figure 8 shown, select the distribution of the actual response lines in multiple scenarios to train the double margin spectrum model. The training process may include the following steps:

[0106] Step 1, the experimental control platform selects the virtual working condition platform (cut in / cut out, decelerate), as well as the scene weather conditions, lighting intensity, road surface conditions, etc.;

[0107] Step 2, the experimental control platform sets the speed of the autonomous driving vehicle on the vehicle wheel suspension platform to V = N km / h (N = 20, 30, 40, 50, 60, 70, 80);

[0108] Step 3, the experimental control platform starts to provide the speed V for the target vehicle in the virtual working condition and starts to decelerate at a distance of L = M m (M = 10, 20, 30, 40, 50, 60) from the target vehicle on the lane;

[0109] Step 4: The virtual working condition platform and the autonomous vehicle feed back multi-motion parameters such as the speed and acceleration at the actual response moment to the test control platform, and the test control platform calculates the TTC (Time to Collision).

[0110] Step 5: The autonomous vehicle measures the response time margin and the collision avoidance maneuver margin at the actual response moment and feeds them back to the test control platform, and a dual expression model of the response time margin and the collision avoidance maneuver margin at the response moment point is constructed.

[0111] Step 6: Repeat steps 1 to 5 to obtain a dual margin expression model of the continuous scale response time margin and the collision avoidance maneuver margin of the autonomous vehicle. As Figure 12 shown, when the value of the collision avoidance maneuver margin is zero, it means that the autonomous vehicle just reaches the vehicle ability boundary (zero crossing point), which represents the cut-off emergency level (bandwidth) that can handle the emergency degree of the emergency working condition. The total area of the response time margin corresponding to the bandwidth represents the total margin of a response ability, and the positive area of the value of the collision avoidance maneuver margin also represents a safety ability. Calculate the area of the dual margin expression model to complete the numerical calculation and evaluation of the safety ability of the autonomous vehicle under the continuous scale.

[0112] In this embodiment, the safety ability detection method of the autonomous vehicle passes the instantaneous braking response of the autonomous vehicle in different traffic scenarios, defines the time margin model in the time dimension and the maneuver margin model in the vehicle control dimension, and through the response observations of multiple random traffic scenarios, a dual margin spectrum model of time margin - maneuver ability margin can be synchronously formed in the double coordinate system, and an ability bandwidth model is defined within this model, so as to quantitatively evaluate the safety ability of autonomous driving on the continuous scale; it solves the problems of single process in the verification of autonomous driving safety ability, lack of rationality and low flexibility. This method integrates interdisciplinary subjects such as vehicle system dynamics / kinematics and classical control theory, considers the full-scenario domain of the autonomous vehicle, and based on multi-dimensional input variables such as relative distance and relative speed, in a certain variable space, combines the response of the input variables to this space to judge the safety attributes and autonomous driving safety ability of the autonomous vehicle.

[0113] Those of ordinary skill in the art will understand that all or some of the steps in the methods disclosed above, and the functional modules / units in systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof. In the hardware implementation, the division of functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be executed by several physical components in cooperation. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software can be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, communication media typically contains computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and can include any information delivery medium.

Claims

1. A safety capability detection method for an autonomous vehicle, characterized in that, The method includes: When there is a risk of collision between the autonomous vehicle and the first vehicle, obtaining the response control force output by the decision-making of the autonomous vehicle, where the first vehicle is the vehicle in the current lane where the autonomous vehicle is traveling and is the closest vehicle in front of the autonomous vehicle; At the first moment when the autonomous vehicle makes a braking response, determining the motion change parameters of the autonomous vehicle, as well as the relative speed and relative distance between the autonomous vehicle and the first vehicle; According to the determined motion change parameters, relative speed and relative distance, generating a dual margin spectrum using a pre-set dual margin spectrum model, where the dual margin spectrum consists of a time margin spectrum and a control margin spectrum; Determining the safety capability data of the autonomous vehicle according to the generated dual margin spectrum; The determining the safety capability data of the autonomous vehicle according to the generated dual margin spectrum includes: According to the generated dual margin spectrum, determining the interval on the continuous scale of the response capability of the autonomous vehicle when the collision avoidance control margin value is zero; Calculating the total area of the response time margin corresponding to this interval; Wherein, this total area represents the cumulative evaluation value of the safety capability of the autonomous vehicle on the continuous scale; The dual margin spectrum model includes: a response time margin model and a control margin model; The response time margin model is: Tm = (Sd - Sa) / V; This response time margin model is the time difference Tm between the first moment Td and the pre-set theoretical latest response moment Ta; Wherein, Td is the moment when the autonomous vehicle makes a braking response, Sd is the relative distance between the autonomous vehicle and the first vehicle corresponding to the first moment, Sa is the relative distance between the autonomous vehicle and the first vehicle corresponding to the theoretical latest response moment, and V is the running speed of the autonomous vehicle at the first moment; The control margin model: am = a0 - a; Wherein, a is the response control force output by the decision-making of the autonomous vehicle at the first moment Td, a0 is the maximum braking intensity that the vehicle has on the current road surface, and am is the braking margin of the collision avoidance capability.

2. The safety capability detection method for an autonomous vehicle according to claim 1, characterized in that The time margin spectrum is a spectrum with the scene urgency as the abscissa and the time margin as the ordinate; The control margin spectrum is a spectrum with the scene urgency as the abscissa and the operation margin as the ordinate.

3. The safety capability detection method for an autonomous vehicle according to claim 2, characterized in that, The determining the motion change parameters of the autonomous vehicle at the first moment when the autonomous vehicle makes a braking response includes: At the first moment when the autonomous vehicle makes a braking response, using the autonomous vehicle control platform to obtain the speed, position, and actual response time of the autonomous vehicle at the first moment.

4. The safety capability detection method of an autonomous vehicle according to claim 3, wherein The generating a dual margin spectrum using a pre-set dual margin spectrum model according to the determined motion change parameters, relative speed and relative distance includes: Calculating the time response margin of the autonomous vehicle using the response time margin model; Calculating the braking margin of the autonomous vehicle using the control margin model; Generating a dual margin spectrum according to the obtained response margin spectrum curve and control margin spectrum curve in the time dimension.

5. The safety ability detection method of an autonomous vehicle according to claim 4, wherein the relative distance Sa between the autonomous vehicle and the first vehicle corresponding to the theoretical latest response time Ta is the distance required for the autonomous vehicle to brake with the maximum braking intensity and the relative speed to just reduce to 0; the Sa is determined by the following method: Sa = (V - V0) * (V - V0) / 2 / a0; wherein, V is the running speed of the autonomous vehicle at the first moment, V0 is the speed of the first vehicle at the first moment, and a0 is the maximum braking intensity of the vehicle on the current road surface.

6. The safety ability detection method of an autonomous vehicle according to claim 5, wherein if am > 0, it means there is a margin in the control margin; if am = 0, it means full braking; if am < 0, it means the maximum braking intensity of the vehicle on the current road surface is less than the collision avoidance braking intensity a expected to be applied by the autonomous driving, and a collision will necessarily occur under this working condition.

7. The safety ability detection method of an autonomous vehicle according to claim 6, wherein the scene urgency is: C = 1 / TTC, TTC = Sd / (V - V0); or C = 1 / HeadWay, HeadWay = Sd / V; wherein, Sd is the relative distance between the autonomous vehicle and the first vehicle at the first moment, V is the speed of the autonomous vehicle at the first moment, and V0 is the speed of the first vehicle at the first moment.

8. A safety capability detection device for an autonomous vehicle, characterized in that, The device includes: a memory and a processor; wherein, the memory is used to save the program for detecting the safety ability of the autonomous vehicle, and the processor is used to read and execute the program for detecting the safety ability of the autonomous vehicle, and execute the method according to any one of claims 1-7.

Images