CN116610091A - Simulation test method, electronic equipment and computer storage medium - Google Patents
Simulation test method, electronic equipment and computer storage medium Download PDFInfo
- Publication number
- CN116610091A CN116610091A CN202310461054.7A CN202310461054A CN116610091A CN 116610091 A CN116610091 A CN 116610091A CN 202310461054 A CN202310461054 A CN 202310461054A CN 116610091 A CN116610091 A CN 116610091A
- Authority
- CN
- China
- Prior art keywords
- interaction
- automatic driving
- obstacle
- simulation
- key
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 197
- 238000010998 test method Methods 0.000 title claims abstract description 15
- 238000003860 storage Methods 0.000 title claims abstract description 9
- 230000003993 interaction Effects 0.000 claims abstract description 213
- 238000012360 testing method Methods 0.000 claims abstract description 156
- 230000002452 interceptive effect Effects 0.000 claims abstract description 39
- 238000000034 method Methods 0.000 claims description 58
- 238000011156 evaluation Methods 0.000 claims description 44
- 230000006870 function Effects 0.000 claims description 41
- 238000004891 communication Methods 0.000 claims description 29
- 230000033001 locomotion Effects 0.000 claims description 24
- 230000001133 acceleration Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 5
- 230000006399 behavior Effects 0.000 description 11
- 230000008569 process Effects 0.000 description 10
- 230000008859 change Effects 0.000 description 6
- 238000002372 labelling Methods 0.000 description 6
- 230000004888 barrier function Effects 0.000 description 5
- 238000013102 re-test Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000005070 sampling Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000005520 cutting process Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000013136 deep learning model Methods 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011022 operating instruction Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000010992 reflux Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/24—Pc safety
- G05B2219/24065—Real time diagnostics
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides a simulation test method, electronic equipment and a computer storage medium, wherein the simulation test method comprises the following steps: identifying a key obstacle based on a simulation playback test of drive test data, wherein the key obstacle is a traffic object which generates a driving influence on driving of an automatic driving simulation and is a responsible party of the collision; determining various interaction strategies for the interaction of the key obstacle and the automatic driving simulation equipment based on the scene information corresponding to the driving influence; performing interactive deduction between the key obstacle and the automatic driving simulation equipment by using a plurality of interaction strategies; and obtaining a simulation test result according to the interactive deduction result. According to the embodiment of the application, the simulation test efficiency of the simulation test by using the drive test data is improved on the whole, the participation of a complex prediction model is not needed, and the scheme realization cost is reduced.
Description
Technical Field
The embodiment of the application relates to the technical field of automatic driving, in particular to a simulation test method, electronic equipment and a computer storage medium.
Background
With the development of autopilot technology, more and more industries and fields use equipment (such as vehicles, aircrafts and the like) with autopilot functions to perform corresponding work so as to improve the working efficiency and reduce the burden of manual work. In order to ensure the driving safety of the devices in actual work, automatic driving test becomes an indispensable link.
In the automatic driving test, the road test proves that the safety cost of automatic driving is huge, and the simulation test is a lower-cost and safer test mode. The simulation test relies on the drive test data to refine the key scene, so that the automatic driving simulation is effectively realized, and the iteration update of the simulation test algorithm version can be continuously carried out. However, due to the difference of the automatic driving simulation vehicles in terms of decision planning among different versions of the algorithm, when the positions of the automatic driving simulation vehicles are inconsistent with the positions of the collected vehicles in the road test data, and the environmental obstacle in the road test return data is still reduced according to the original position to test the automatic driving system, unreasonable interaction between the automatic driving simulation vehicles and the road test return environmental obstacle is extremely easy to occur, so that the simulation vehicles and the like generate a large number of test false alarms such as unreasonable sudden braking and collision, and the test efficiency is greatly reduced. For this purpose, in a related approach, the situation is handled in a manner that predicts the collision and uniformly takes over the obstacle. However, the mode needs to predict the prediction model of collision to have higher accuracy on one hand, and on the other hand, the mode of unified take over is single, so that the diversity of interactive tests is difficult to realize, and the actual test scene is covered.
Therefore, how to effectively use drive test data to perform simulation test and improve the efficiency of the simulation test becomes a problem to be solved urgently.
Disclosure of Invention
In view of the above, an embodiment of the present application provides a simulation test scheme to at least partially solve the above-mentioned problems.
According to a first aspect of an embodiment of the present application, there is provided a simulation test method, including: identifying a key obstacle based on a simulation playback test of drive test data, wherein the key obstacle is a traffic object which generates a driving influence on the driving of automatic driving simulation equipment and affects a responsible party; determining various interaction strategies for the interaction of the key obstacle and the automatic driving simulation equipment based on the scene information corresponding to the driving influence; performing interactive deduction between the key obstacle and the automatic driving simulation equipment by using a plurality of interaction strategies; and obtaining a simulation test result according to the interactive deduction result.
According to a second aspect of an embodiment of the present application, there is provided an electronic device including: the device comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete communication with each other through the communication bus; the memory is configured to store at least one executable instruction, where the executable instruction causes the processor to perform operations corresponding to the method according to the first aspect.
According to a third aspect of embodiments of the present application, there is provided a computer storage medium having stored thereon a computer program which, when executed by a processor, implements the method according to the first aspect.
According to the scheme provided by the embodiment of the application, in the process of using the drive test data to carry out simulation playback test, the key obstacle which needs to bear driving influence responsibility is identified, so that multiple interaction strategies are determined according to the interaction between the key obstacle and the automatic driving simulation equipment in the scene, and corresponding interaction deductions are carried out based on the multiple interaction strategies, so that multiple different interaction tests in the scene are realized. Through a plurality of different interaction tests, on one hand, the automatic driving simulation equipment can be effectively tested for a plurality of times from different interaction angles so as to determine the performance of the automatic driving simulation equipment, and the driving scene which can be covered by the test is expanded; on the other hand, the effective utilization of the drive test data is realized; on the other hand, the real reason for the running influence on the running of the automatic driving simulation equipment can be determined according to the simulation test result, for example, whether the automatic driving simulation equipment is defective or the false alarm is caused by the version of the simulation test algorithm, so that the possible simulation playback test problem can be exposed. By the method, the simulation test efficiency of the simulation test by using the drive test data is improved on the whole, the participation of a complex prediction model is not needed, and the scheme implementation cost is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the embodiments of the present application, and other drawings may be obtained according to these drawings for a person having ordinary skill in the art.
FIG. 1 is a schematic diagram of an exemplary system for applying aspects of embodiments of the present application;
FIG. 2A is a flow chart of steps of a simulation test method according to an embodiment of the present application;
FIG. 2B is a schematic diagram of an example of a scenario in the embodiment shown in FIG. 2A;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to better understand the technical solutions in the embodiments of the present application, the following description will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which are derived by a person skilled in the art based on the embodiments of the present application, shall fall within the scope of protection of the embodiments of the present application.
The implementation of the embodiments of the present application will be further described below with reference to the accompanying drawings.
FIG. 1 illustrates an exemplary system to which embodiments of the present application may be applied. As shown in fig. 1, the system 100 may include a cloud service 102, a communication network 104, and/or one or more user devices 106, which are illustrated in fig. 1 as a plurality of user devices. In this example, the user device 106 may be implemented as a test-end device for performing autopilot simulation testing.
Cloud server 102 may be any suitable device for storing information, data, programs, and/or any other suitable type of content, including, but not limited to, distributed storage system devices, server clusters, computing cloud server clusters, and the like. In some embodiments, cloud server 102 may store at least drive test data. However, the present application is not limited thereto, and other data, and process data, result data, etc. in the simulated playback test performed by the user device 106 may also be stored in the cloud server 102.
In some embodiments, the communication network 104 may be any suitable combination of one or more wired and/or wireless networks. For example, the communication network 104 can include any one or more of the following: the internet, an intranet, a wide area network (Wide Area Network, WAN), a local area network (Local Area Network, LAN), a wireless network, a digital subscriber line (Digital Subscriber Line, DSL) network, a frame relay network, an asynchronous transfer mode (Asynchronous Transfer Mode, ATM) network, a virtual private network (Virtual Private Network, VPN), and/or any other suitable communication network. The user device 106 can be connected to the communication network 104 via one or more communication links (e.g., communication link 112), and the communication network 104 can be linked to the cloud service 102 via one or more communication links (e.g., communication link 114). The communication link may be any communication link suitable for transferring data between the user device 106 and the cloud service 102, such as a network link, a dial-up link, a wireless link, a hardwired link, any other suitable communication link, or any suitable combination of such links.
The user device 106 may include any one or more user devices suitable for performing automated driving simulation tests. In some embodiments, the user device 106 may perform a simulated playback test based on the drive test data. Because of the iterative updating of the test algorithm version formed by the simulation playback test based on the drive test data, unreasonable interaction between the automatic driving simulation vehicle and the road test playback environment obstacle is caused, and a large number of abnormal false alarms are caused. Thus, in some embodiments, the user device 106 may identify key obstacles during the simulated playback test based on the drive test data, i.e., those that interact with and have a driving impact on the driving of the automated driving simulation device, and are traffic objects affecting responsible parties (including, but not limited to, motor vehicles or non-motor vehicles, etc.); and further, determining multiple interaction strategies for the interaction of the key obstacle and the automatic driving simulation equipment according to the scene corresponding to the driving influence, and carrying out corresponding multiple interaction deductions under the multiple interaction strategies, so as to find out the problem of the automatic driving simulation equipment or determine the false report of the data according to the interaction deduction result. In some embodiments, user device 106 may comprise any suitable type of device. For example, in some embodiments, user device 106 may include a mobile device, a tablet computer, a laptop computer, a desktop computer, a vehicle platform system, and/or any other suitable type of user device.
Based on the above system, the embodiments of the present application provide a simulation test method, which is described by a plurality of embodiments.
Referring to FIG. 2A, a flowchart of steps of a simulation test method in accordance with an embodiment of the present application is shown.
The simulation test method of the embodiment comprises the following steps:
step S202: and identifying the key obstacle based on the simulation playback test of the drive test data.
The key obstacle is a traffic object which interacts with the automatic driving simulation equipment, generates driving influence (such as collision) on the driving of the automatic driving simulation equipment, and is a responsible party of the influence.
The drive test data can be generally obtained by sampling data of the automatic driving collection vehicle in an actual physical environment, and the drive test data can generally well describe the physical environment in which the automatic driving collection vehicle is located and the interaction behavior of the automatic driving collection vehicle and other traffic objects in the physical environment. In the embodiment of the application, the obtaining mode and the specific data form of the drive test data and the physical environment described by the drive test data are not limited.
On the basis of obtaining the drive test data, the simulation playback test of the automatic driving simulation equipment can be performed by any appropriate mode or algorithm. In the simulation playback test, the automatic driving simulation equipment takes the physical environment of the automatic driving collection vehicle during sampling as a simulation environment, and performs simulation interaction with each traffic object in the environment. However, as previously described, the version of the simulated playback testing algorithm is continually updated iteratively, and in each iterative update, the interaction strategy employed by the autopilot simulation device in interacting with other traffic objects may be different, thereby potentially generating anomaly information.
If the generated abnormal information directly indicates that the automatic driving simulation equipment has a problem or a fault (if the collision occurs and the responsible party for the collision is the automatic driving simulation equipment, or if the automatic driving simulation equipment has a problem such as sudden braking, forced stopping and the like), the problem or the fault, together with scene information and running information (including but not limited to an interaction strategy, a running track and the like) where the problem or the fault is located, is sent to preset analysis equipment, so that the analysis equipment analyzes the problem or the fault, and further performs corresponding optimization on the automatic driving simulation equipment.
If the generated abnormal information indicates that the automatic driving simulation equipment interacts with other traffic objects, and the interaction has bad driving influence such as collision on the driving of the automatic driving simulation equipment, and the responsible party of the collision is the other traffic objects, then there is a possibility that the automatic driving simulation equipment and the other traffic objects have unreasonable interaction due to version updating, for example, the automatic driving simulation equipment is inconsistent with the position of an automatic driving collection vehicle in the drive test data, and the environmental obstacle in the drive test return data still returns to the original test automatic driving system according to the original test, thereby generating test false alarms such as unreasonable emergency brake and collision of the automatic driving simulation equipment. In this case, further screening processing is required for the traveling influence such as collision.
Therefore, in the embodiment of the application, the key barriers in the simulation playback test are identified, namely, those which interact with the automatic driving simulation equipment and generate driving influence on the driving of the automatic driving simulation equipment are identified, and the traffic objects which influence responsible parties are identified for subsequent further processing.
In one possible manner, a simulated playback test may be performed according to the drive test data to identify therefrom key data frames that have a driving impact (e.g., a collision) on the driving of the autopilot simulation device; from the key data frames, key obstructions are identified. By means of the method that the key data frames are identified first and then the key obstacles are identified, the identification of the key obstacles is more efficient, and the data processing burden of the system is smaller. The key data frame may be a data frame corresponding to a time point when a travel influence such as a collision occurs, or a multi-frame data frame related to a time point when a travel influence such as a collision occurs, or the like, for example. However, the method is not limited thereto, and key data frames generating driving influence can be identified by intelligent labeling. The intelligent labeling is a machine learning auxiliary technology, auxiliary labeling is carried out through a machine learning model, a labeling person only needs to label a target object once, and the model can track and predict the position and the driving condition of the target object in a subsequent frame. Based on the above, during the simulation playback test, one or more frames of the automatic driving simulation device, which generate driving influence such as collision, can be identified as key data frames. Further, a plurality of parties that affect traveling such as collision are identified from the key data frame, and two parties are exemplified, one of which is an automatic driving simulation device and the other of which is another traffic object. Further, responsibilities between multiple parties may be determined based on a responsibilities distribution model, such as an RSS (Responsibility Sensitive Safety, responsibilities sensitive security) model. Taking a collision as an example, after a key data frame that the automatic driving simulation device collides with other traffic objects is identified, for the parties in the key data frame, such as the automatic driving simulation device party and the other traffic objects party, the responsibility of the two parties can be determined through an RSS model, and if the responsibility is determined to be the other traffic objects party, the traffic objects party is identified as a key obstacle. If the responsibility is determined to be on the side of the automatic driving simulation equipment, corresponding information is directly sent to the background analysis equipment for analysis as described above.
In order to facilitate the subsequent retest in the same scene to fully expose the problems of the autopilot simulation device or the test data, in one possible manner, after obtaining the key data frame, the motion trail of the key obstacle and the motion trail of the autopilot simulation device can be obtained based on the key data frame; and determining scene information corresponding to the driving influence, such as collision scene information of the collision, according to the obtained motion trail. The scene information includes, but is not limited to: information of a driving route of a key obstacle and an automatic driving simulation device, information of a driving direction, information of a driving behavior, information of a driving speed, information of a relative position, information of a relative speed, information of a surrounding environment and other traffic objects, and the like.
Through the process, the identification of key barriers and the collection of scene information are realized, and conditions are provided for the follow-up interactive deduction.
Step S204: based on the scene information corresponding to the driving impact, a plurality of interaction strategies are determined for the interaction of the key obstacle and the automatic driving simulation equipment.
For the identified scene which is responsible by the key obstacle, there are various possibilities for the occurrence of the scene, such as test false alarms generated by inconsistent positions of the automatic driving simulation equipment and the automatic driving collection vehicle in the drive test data, or driving anomalies such as collision caused by the automatic driving simulation equipment or the key obstacle, etc. For this purpose, multiple tests of different interactions can be performed for this scenario to test possible causes and possible problems.
In one possible manner, multiple interaction strategies may be determined for the interaction of the critical obstacle with the autopilot simulation device through a monte carlo tree search based on the scenario information corresponding to the driving impact. The Monte Carlo tree searching mode is mostly based on the current situation, simulates the situation which possibly happens in the future, and finds out the current feasible action from a plurality of possible actions. In particular, taking a collision scene as an example, the embodiment of the application can determine various interaction strategies between a key obstacle and automatic driving simulation equipment through Monte Carlo tree search based on collision scene information, including the relative position and relative speed between the key obstacle and the automatic driving simulation equipment, the position and speed between the key obstacle and other traffic objects, and the traffic environment (lanes, speed limit, traffic signals, etc.). It should be noted that, the method is not limited to the method of Monte Carlo tree search, and other methods of determining the interaction policy, such as a method of determining the interaction policy by training a completed convolutional neural network model, may be equally applicable to the solution of the embodiment of the present application. But by means of Monte Carlo tree searching, a more reasonable interaction strategy can be obtained, and further retest effect and efficiency are improved.
However, in order to make the interaction and test efficiency higher, in a feasible manner, an interaction target may also be set, specifically, an interaction target of interaction between the key obstacle and the autopilot simulation device in a scene indicated by the scene information may be obtained according to the scene information corresponding to the driving influence; generating a plurality of candidate interaction strategy sequences according to the interaction targets; evaluating a plurality of candidate interaction strategy sequences according to a preset cost evaluation function; and determining a target interaction strategy sequence according to the evaluation result. Wherein the interaction target is used to indicate a target result of the interaction, such as, for example, a critical obstacle cutting into the front of the autopilot simulation device, or keeping the following car behind the autopilot simulation device, or changing lanes to the side of the autopilot simulation device, etc. The interaction target may be set by those skilled in the art according to the actual situation, and in the embodiment of the present application, the specific implementation of the interaction target is not limited.
Based on the interaction targets, a plurality of candidate interaction policy sequences may be generated, for example, by a Monte Carlo tree search. The sequence is a combination of a series of interaction behaviors, such as lane change-acceleration-cutting-in, and the like, so that a candidate interaction strategy sequence can be formed. The interaction policy sequence for realizing a certain interaction target may be various, such as the aforementioned cut-in, which may be realized as a left lane-accelerating-cut-in, a right lane-accelerating-cut-in, or a following lane-cutting-in. However, the purpose of re-interaction and testing is to achieve reasonable and effective interaction while avoiding, so that for multiple candidate interaction strategy sequences generated, evaluation can be performed by a preset cost evaluation function to determine a preferred sequence, namely a target interaction strategy sequence.
Optionally, the cost evaluation function may include at least one of: a first price evaluation function for evaluating the implementation efficiency of the interaction target based on the candidate interaction policy sequence; the second cost evaluation function is used for evaluating the smoothness of the motion trail when the key obstacle moves based on the candidate interaction strategy sequence; and the third price evaluation function is used for evaluating the interaction difficulty of the key obstacle and the automatic driving simulation equipment when the candidate interaction strategy sequence is used for performing movement. For the driving decision, the evaluation can be performed through various aspects such as comfort, safety, stability and the like, and on the basis, the embodiment of the application selects a plurality of dimensions such as the realization efficiency of the interaction target, the smoothness of the motion track, the interaction difficulty and the like to evaluate whether the candidate interaction strategy sequence is reasonable and efficient on the premise of fully considering the interaction target.
The evaluation of the multiple candidate interaction strategy sequences according to the preset cost evaluation function may include at least one of the following: according to the implementation time of the interaction targets, a first price evaluation function is used for evaluating the implementation efficiency of the interaction targets realized by the multiple candidate interaction strategy sequences; according to the acceleration of the key obstacle when the key obstacle moves based on the candidate interaction strategy sequences and the average value of the acceleration, a second cost evaluation function is used for evaluating the smoothness of the moving track of the key obstacle; and according to the shortest interaction distance between the key obstacle and the automatic driving simulation equipment when the key obstacle moves based on the multiple candidate interaction strategy sequences, using a third price evaluation function to evaluate the interaction difficulty between the key obstacle and the automatic driving simulation equipment. It should be noted that, in the embodiment of the present application, specific implementation manners of the first, second and third cost evaluation functions are not limited, and those skilled in the art may implement the cost evaluation functions by using any suitable formulas or algorithms based on consideration factors of the foregoing cost evaluation functions according to actual needs, so as to evaluate multiple candidate interaction policy sequences. For the first price evaluation function, the implementation efficiency level can be determined by comparing the implementation time length of various candidate interaction strategy sequences in the test and evaluating the candidate interaction strategy sequences; alternatively, the implementation time of various candidate interaction strategy sequences in the test is compared with the average implementation time of the interaction targets, so as to compare the various candidate interaction strategy sequences, determine the level of the implementation efficiency, and the like. For example, the first price estimation function includes, but is not limited to, a function that calculates a simulated duration, a simulated average speed, a simulated mileage, a simulated moving average speed, a number of parks, a total parking duration, a maximum parking duration, and the like. The second and third price estimation functions may also be used in a similar manner. Illustratively, the second cost evaluation function may include, but is not limited to: an abrupt brake evaluation function (counting that the x, y value of acceleration x, y of each frame and x, y value of abrupt brake direction exceeds the frame number in a prescribed interval, the frame number exceeds a certain threshold), a large steering (steering wheel angle exceeds a certain threshold), etc. Third generation price estimation functions may include, but are not limited to: a function of distance based on TTC (Time-To-Collision), etc.
Through evaluation, one or more strategy sequences can be selected from a plurality of candidate interaction strategy sequences to serve as target interaction strategy sequences. For the completeness of the test, in general, the target interaction strategy sequence may be multiple.
In addition, in order to enable reasonable driving behavior of the following of the critical obstacle, in one possible manner, after determining the target interaction strategy sequence, a strategy for keeping the current lane driving may be added to the critical obstacle. That is, after a series of interaction actions indicated by the target interaction policy sequence are executed, the key obstacle will adopt a mode of keeping the key obstacle running on the lane after the last interaction action is completed, so as to perform subsequent further operations and ensure running stability.
Step S206: and using various interaction strategies to carry out interaction deduction between the key obstacle and the automatic driving simulation equipment.
In the embodiment of the application, interactive deduction means a test process of enabling a key obstacle and automatic driving simulation equipment to run according to a series of interactive actions indicated by various determined interactive strategies under a scene corresponding to the running influence, such as a collision scene indicated by collision scene information, and obtaining an interactive result.
In order to enable the process to be carried out smoothly, in a feasible manner, track planning can be carried out for various interaction strategies respectively; and carrying out interactive deduction between the key obstacle and the automatic driving simulation equipment according to various interactive strategies and corresponding track planning thereof. In specific implementation, a corresponding track plan can be sampled according to an interaction target, for example, when a lane change decision is made, different lane offsets and lane change times are sampled on a target lane, the fastest lane change track is found under the constraint of acceleration and speed ranges, and the track of continuous motion from the current running state to the target running state is solved by using a five-time polynomial, so that the position, speed and acceleration state of a key obstacle are obtained. For example, a Lattice trajectory planning algorithm may be employed for trajectory planning. The Lattice planning algorithm is a motion planning algorithm based on sampling, and a vehicle coordinate system is converted into a reference line coordinate system, namely a freset coordinate system, then d-axis and s-axis of freset are respectively planned under the freset coordinate system to form a planned track under the freset coordinate system, and then the track under the freset coordinate system is synthesized into a track under the world coordinate system to be restored to the track under the world coordinate system. One specific implementation of the Lattice planning algorithm is to use Lattice Planner, which is a vehicle motion local track Planner, and the output track is directly input to the controller, so that the controller can complete tracking control of the local track. Thus, the track of the Lattice Planner output is a smooth, collision-free, steady and safe local track that satisfies the vehicle kinematics and speed constraints. The input mainly comprises three parts, namely: sensing and obstacle information, reference line information and positioning information; the output of the vehicle tracking control system is a track formed by a series of track points with speed information, so that the stability and the safety of the vehicle controller in the vehicle tracking control process can be ensured. But not limited to, other ways of implementing the track rule according to the interaction strategy, such as a deep learning model, are also applicable to the scheme of the embodiment of the present application.
In the embodiment of the application, a plurality of interaction strategies to be subjected to interaction deduction may be provided, and in order to improve the interaction deduction efficiency, in a feasible manner, a plurality of interaction strategies can be used to parallelly carry out a plurality of interaction deductions between the key obstacle and the automatic driving simulation equipment, wherein the interaction deductions correspond to the plurality of interaction strategies. The parallel mode can be realized by a person skilled in the art in a proper parallel mode according to actual conditions, for example, a plurality of identical collision scenes are started through a plurality of processes, different interaction strategies are used in different collision scenes, and interaction deduction is carried out according to the track rules corresponding to the interaction strategies; or, in the same interface, a plurality of identical collision scenes are presented, but different interaction strategies are used in different collision scenes, and interaction deduction is performed according to the track rules corresponding to the interaction strategies, and the like, which are all within the protection scope of the embodiment of the application.
Step S208: and obtaining a simulation test result according to the interactive deduction result.
After the interaction deduction between the key obstacle and the automatic driving simulation equipment is carried out according to each interaction strategy, a corresponding interaction deduction result is obtained. The interactive deduction results may be deduction results with the influence eliminated, such as eliminating the collision, and then the original collision may be test false alarm caused by the drive test data; but may also be the result of a deduction that a collision has occurred again. In this case, it is required to determine whether a driving impact such as collision occurs when a key obstacle interacts with the autopilot simulation device in the corresponding interaction deduction according to the interaction deduction result, and the responsible party is the interaction of the autopilot simulation device; if the information is available, the information deduced by the interaction corresponding to the interaction is determined to be the auxiliary information of the problem to be processed aiming at the automatic driving simulation equipment. Further, the data can be sent to a background analysis device for targeted analysis. Therefore, the effectiveness of the simulation test can be ensured, and meanwhile, the test diversity is enriched.
According to the embodiment, in the process of performing simulation playback test by using drive test data, a key obstacle needing to bear driving influence responsibility is identified, and then multiple interaction strategies are determined according to interaction between the key obstacle and automatic driving simulation equipment in the scene, and corresponding interaction deductions are performed based on the multiple interaction strategies, so that multiple different interaction tests in the scene are realized. Through a plurality of different interaction tests, on one hand, the automatic driving simulation equipment can be effectively tested for a plurality of times from different interaction angles so as to determine the performance of the automatic driving simulation equipment, and the driving scene which can be covered by the test is expanded; on the other hand, the effective utilization of the drive test data is realized; on the other hand, the real reason for the running influence on the running of the automatic driving simulation equipment can be determined according to the simulation test result, for example, whether the automatic driving simulation equipment is defective or the false alarm is caused by the version of the simulation test algorithm, so that the possible simulation playback test problem can be exposed. By the method, the simulation test efficiency of the simulation test by using the drive test data is improved on the whole, the participation of a complex prediction model is not needed, and the scheme implementation cost is reduced.
Hereinafter, the above-described process will be exemplarily described with a specific example of a collision scenario, as shown in fig. 2B.
As shown in fig. 2B, the process includes:
step S302: and performing simulation playback test according to the drive test data.
In the step, after the drive test data are obtained and the automatic driving algorithm to be tested is determined, the simulation playback test can be performed. Optionally, in this example, scene data before and after the key data frame is also identified and collected as collision scene information.
When the simulation playback test is carried out, an automatic driving algorithm to be tested can be adopted, the road test obstacle is played back by adopting the position, the speed, the acceleration, the course angle and the like in the road test data, and the road test obstacle is used as the perception input of the automatic driving algorithm to be tested to drive the prediction, decision and planning control module of the automatic driving simulation vehicle to run.
When the automatic driving simulation vehicle and the road test data acquisition vehicle make different decisions, such as acceleration and deceleration, road change and the like, part of road test obstacles collide with the automatic driving simulation vehicle due to lack of interaction.
In this example, the data frame corresponding to the time of collision is used as the key data frame, but as described above, the key data frame may be identified by means such as intelligent labeling.
Based on the key data frames, collision responsibility determination can be made according to rules of traffic, and in particular implementations, the collision responsibility can be determined using an RSS model. If the road test obstacle is mainly responsible, such as a rear-end collision of an automatic driving simulation vehicle, the road test obstacle is determined to be a key obstacle, the behavior of the road test obstacle needs to be deduced, and a test is performed again to determine the validity of the collision.
In this example, the motion trajectories of all the road-test obstacles and the motion trajectories of the automatic driving simulation vehicle 10s before and after the collision are also recorded, so as to provide scene information for the effective interaction behavior of the subsequent deduction key obstacle. If the main responsibility of the collision is due to the fact that the autonomous simulated vehicle, for example, is cut into in the face of other vehicles, the autonomous simulated vehicle is completely free of reactions such as deceleration interactions, no deduction is made for the situation, but the result of the simulated test not passing is directly given. Further, the relevant data is handed to a background analysis device for subsequent problem analysis for the automated driving simulation vehicle.
Therefore, the step carries out simulation playback test by playing back the original drive test data, determines the key obstacle through the methods of obstacle collision judgment, intelligent labeling and the like, and collects and records the movement track data of the obstacle and the automatic driving simulation equipment before and after the key data frame. Therefore, the primary simulation is realized, the collision responsibility is determined, and relevant data input and scene information are provided for the subsequent interactive deduction of the key obstacle.
Step S304: an interaction strategy in a collision scenario is determined for the autopilot simulation device and the critical obstacle.
Firstly, restoring a real collision scene by utilizing the motion trail of the road test obstacle including the key obstacle before and after the collision and the motion trail of the automatic driving simulation equipment, namely the collision scene information.
On this basis, a Monte Carlo tree search may be employed to determine interaction strategies. The monte carlo tree search combines the generality of random simulation with the accuracy of the tree search, giving a better decision, i.e. an interaction strategy.
In this example, the determination of the interaction policy may also incorporate interaction targets. For example, given the interaction targets of the key obstacle and the automatic driving simulation equipment, for example, the key obstacle is cut into the front of the automatic driving simulation equipment, the following car is kept behind the automatic driving simulation equipment, or the road change is carried out to the side of the automatic driving simulation equipment, and the like, a better target interaction strategy sequence is determined through a plurality of cost evaluation functions, so that reasonable and effective interaction is achieved while collision is avoided.
Wherein the cost evaluation function comprises at least one of: a cost evaluation function (first price evaluation function) that evaluates the implementation efficiency of the interaction target, which can be implemented by evaluating the implementation time of the interaction target; the cost evaluation function for evaluating the comfort of the trajectory, i.e., the cost evaluation function for evaluating the smoothness of the motion trajectory (second cost evaluation function), may be implemented by evaluating the acceleration of the critical obstacle and the average value of the acceleration; the challenging cost evaluation function of evaluating interactions, i.e., the cost evaluation function of evaluating interaction difficulty (third cost evaluation function), can be achieved by evaluating the shortest interaction distance with a critical obstacle.
Taking the example that a key obstacle cuts into an interactive target of the automatic driving simulation equipment during the simulation playback test, for example, a Monte Carlo tree search can be performed 5s before collision, so as to respectively try to cut into the interactive target right in front of the automatic driving simulation equipment, wait for the automatic driving simulation equipment to pass into the interactive target behind the automatic driving simulation equipment and keep the interactive target at the side of the automatic driving simulation equipment. Through Monte Carlo tree search, the acceleration of a key obstacle in a lane can be obtained, then the lane is changed to the lane where the automatic driving simulation equipment is located so as to realize the interaction strategy sequence cut into the front of the automatic driving simulation equipment, and the interaction strategy sequence is used as a target interaction strategy sequence. Of course, other corresponding interaction policy sequences may also implement the interaction target, and may also be used as the target interaction policy sequence.
Because the present example mainly solves the behavior planning of the key interaction segment of the key obstacle in the collision scene, it may also be set for the key obstacle to increase the interaction policy sequence for keeping the lane running for the key obstacle after the target interaction policy sequence is completed, so as to realize the subsequent reasonable behavior of the key obstacle.
It can be seen that according to the interaction behavior of the key obstacle and the automatic driving simulation equipment in the collision scene, such as the interaction targets of following a car, cutting in, and the like, a reasonable one or more target interaction strategy sequences of the key obstacle and the automatic driving simulation equipment interaction are obtained through Monte Carlo tree search, and interaction deduction is performed. By exploring possible interaction targets and corresponding interaction strategy sequences between the key obstacle and the automatic driving simulation equipment according to the collision scene, a mode of key obstacle interaction deduction is provided for subsequent targeted test as a takeover mode basis of the key obstacle in a new simulation playback test task.
Step S306: and performing interactive deduction of the key obstacle and the automatic driving simulation equipment.
In the example, a Lattice track planning algorithm is adopted to realize the motion planning of the key obstacle and the automatic driving simulation equipment corresponding to the target interaction strategy sequence.
For example, according to a determined target interaction strategy sequence, corresponding planning states, such as lane changing decisions, are sampled, different lane offsets and lane changing times are sampled on a target lane, the fastest lane changing track is found under the constraint of acceleration and speed ranges, and the track from the current state to the target state to continuous motion is solved by using a five-time polynomial, so that the positions, the speeds and the acceleration states of the key obstacle and the automatic driving simulation equipment are obtained.
Illustratively, the tracks output by the Lattice Planner can be used as initial values, and then the Lattice Planner is used for densely sampling near the initial values, so that collision-free tracks meeting kinematic constraints are obtained to drive key barriers in the simulation playback test to interact.
In addition, the present example also provides an expansion of a plurality of track planning methods, and by way of example, a method of combining a learning-based Planner and a Lattice Planner can be supported to carry out track planning, control key obstacles in a simulated playback test, and enrich interactive test methods. By means of a track planning algorithm based on combination of a Lattice Planner and a deep learning model and the like, various types of interactive deductions between key barriers and automatic driving simulation equipment can be achieved.
Therefore, through the step, the movement track planning of the key obstacle is realized by using the Lattice track planning algorithm, and the effective interaction test can be performed while collision with the automatic driving simulation equipment is avoided.
Step S308: and outputting a simulation test result.
In this example, the target interaction policy sequence may include a plurality of corresponding interaction deductions, and the plurality of interaction deductions may be executed in parallel to implement the multi-task parallel test.
After all the interactive deductions are finished, corresponding simulation playback test results are generated, and test scores of different interactive deductions are given by combining various evaluation dimensions of the automatic driving simulation equipment in collision, traffic rule evaluation, driving comfort and the like, namely the influence degree of the different interactive deductions on the automatic driving simulation equipment.
Then, the interactive deduction result with the greatest influence degree on the automatic driving simulation equipment can be selected and presented to the user, and the rest interactive deduction result can be folded under the interactive deduction result, so that the user can conveniently check the interactive deduction result.
Through the step, the presentation of the test result corresponding to the interactive deduction is realized, and the result is based on various interactive deductions, for example, the result of various interactive behaviors such as cutting into the automatic driving simulation equipment by accelerating and changing the lane, giving way to the automatic driving simulation equipment by decelerating and changing the lane, keeping the lane to run and the like, so that the effectiveness of the simulation test is ensured, and meanwhile, the test diversity is enriched.
According to the method and the device, the retest flow of automatic identification of the key obstacle, interactive strategy determination of the key obstacle, various interactive deduction of the key obstacle and comparison and selection of the test result is realized, the collision behavior generated by unreasonable drive test data can be effectively reduced through the retest flow, and the effectiveness of automatic driving algorithm test is improved.
Based on the flow, a test architecture of drive test data playback simulation test-key obstacle deduction (interaction strategy determination and corresponding interaction deduction execution) -simulation retest is realized, and test diversity and effectiveness of the drive test data are improved. Different interaction strategies and interaction deduction of key barriers are realized by adopting Monte Carlo tree searching and Lattice track planning modes, a better target interaction strategy sequence is obtained for simulation test, and interaction test diversity is improved. In addition, by adopting a Monte Carlo tree searching and Lattice track planning mode, on one hand, the track planning can acquire a correct motion track a priori due to a deduction mechanism, and a better planning result is facilitated; on the other hand, compared with the result of model training, the method is more controllable and has the condition of large-scale landing.
After the result of the interactive deduction is obtained, the example also combines the influence degree of the result on the automatic driving simulation equipment to give the result of the interactive deduction with challenges to the automatic driving simulation equipment.
Compared with the traditional mode, such as the interim mode, the method determines whether to take over the obstacle according to the predicted interaction track, the method judges according to the deterministic interaction result (such as collision), and the judging result is not too conservative or aggressive along with the prediction model, so that boundary test conditions are reserved and are not biased along with the model. Compared with the logsim mode, the method solves the problem of unintelligible obstacle interaction of the logsim mode, greatly improves the utilization effectiveness of the drive test data, improves the test value of the automatic drive test data reflux, and plays an important role in accelerating the automatic drive test and improving the safety of algorithms.
It should be noted that, although the collision scenario is taken as an example in this embodiment, it should be apparent to those skilled in the art that other scenarios that affect the running of the autopilot simulation device (such as scenarios that cause the autopilot simulation device to generate sudden braking, sudden change of track, etc.) may refer to this embodiment to implement corresponding simulation tests, which are all within the scope of the present application.
Referring to fig. 3, a schematic structural diagram of an electronic device according to an embodiment of the present application is shown, and the specific embodiment of the present application is not limited to the specific implementation of the electronic device.
As shown in fig. 3, the electronic device may include: a processor 402, a communication interface (Communications Interface) 404, a memory 406, and a communication bus 408.
Wherein:
processor 402, communication interface 404, and memory 406 communicate with each other via communication bus 408.
A communication interface 404 for communicating with other electronic devices or servers.
The processor 402 is configured to execute the program 410, and may specifically perform relevant steps in the above-described simulation test method embodiment.
In particular, program 410 may include program code including computer-operating instructions.
The processor 402 may be a CPU or specific integrated circuit ASIC (Application Specific Integrated Circuit) or one or more integrated circuits configured to implement embodiments of the present application. The one or more processors comprised by the smart device may be the same type of processor, such as one or more CPUs; but may also be different types of processors such as one or more CPUs and one or more ASICs.
Memory 406 for storing programs 410. Memory 406 may comprise high-speed RAM memory or may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
The program 410 may include a plurality of computer instructions, and the program 410 may specifically enable the processor 402 to perform operations corresponding to the simulation test method described in the foregoing method embodiment through the plurality of computer instructions.
The specific implementation of each step in the procedure 410 may refer to the corresponding steps and corresponding descriptions in the units in the above method embodiment, and have corresponding beneficial effects, which are not described herein. It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus and modules described above may refer to corresponding procedure descriptions in the foregoing method embodiments, which are not repeated herein.
The present application also provides a computer storage medium having stored thereon a computer program which, when executed by a processor, implements the method described in any of the preceding method embodiments. The computer storage media includes, but is not limited to: a compact disk read Only (Compact Disc Read-Only Memory, CD-ROM), random access Memory (Random Access Memory, RAM), floppy disk, hard disk, magneto-optical disk, or the like.
The embodiment of the application also provides a computer program product, which comprises computer instructions for instructing a computing device to execute the operations corresponding to the simulation test method in the embodiment of the method.
In addition, it should be noted that, the information related to the user (including, but not limited to, user equipment information, user personal information, etc.) and the data related to the embodiment of the present application (including, but not limited to, sample data for training the model, user data in the collected drive test data, data for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region, and provide a corresponding operation entry for the user to select authorization or rejection.
It should be noted that, according to implementation requirements, each component/step described in the embodiments of the present application may be split into more components/steps, or two or more components/steps or part of operations of the components/steps may be combined into new components/steps, so as to achieve the objects of the embodiments of the present application.
The methods according to embodiments of the present application described above may be implemented in hardware, firmware, or as software or computer code storable in a recording medium such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk, or as computer code originally stored in a remote recording medium or a non-transitory machine-readable medium and to be stored in a local recording medium downloaded through a network, so that the methods described herein may be processed by such software on a recording medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware such as an application specific integrated circuit (Application Specific Integrated Circuit, ASIC) or field programmable or gate array (Field Programmable Gate Array, FPGA). It is understood that a computer, processor, microprocessor controller, or programmable hardware includes a Memory component (e.g., random access Memory (Random Access Memory, RAM), read-Only Memory (ROM), flash Memory, etc.) that can store or receive software or computer code that, when accessed and executed by the computer, processor, or hardware, performs the methods described herein. Furthermore, when a general purpose computer accesses code for implementing the methods illustrated herein, execution of the code converts the general purpose computer into a special purpose computer for performing the methods illustrated herein.
Those of ordinary skill in the art will appreciate that the elements and method steps of the examples described in connection with the embodiments disclosed herein can be implemented as electronic hardware, or as a combination of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the embodiments of the present application.
The above embodiments are only for illustrating the embodiments of the present application, but not for limiting the embodiments of the present application, and various changes and modifications may be made by one skilled in the relevant art without departing from the spirit and scope of the embodiments of the present application, so that all equivalent technical solutions also fall within the scope of the embodiments of the present application, and the scope of the embodiments of the present application should be defined by the claims.
Claims (14)
1. A simulation test method, comprising:
identifying a key obstacle based on a simulation playback test of drive test data, wherein the key obstacle is a traffic object which generates a driving influence on the driving of automatic driving simulation equipment and affects a responsible party;
Determining various interaction strategies for the interaction of the key obstacle and the automatic driving simulation equipment based on the scene information corresponding to the driving influence;
performing interactive deduction between the key obstacle and the automatic driving simulation equipment by using a plurality of interaction strategies;
and obtaining a simulation test result according to the interactive deduction result.
2. The method of claim 1, wherein the drive test data based simulated playback test identifies a critical obstacle comprising:
performing simulation playback test according to the drive test data to identify key data frames which have driving influence on the driving of the automatic driving simulation equipment;
and identifying the key obstacle from the key data frame.
3. The method of claim 2, wherein the method further comprises:
based on the key data frame, obtaining a motion trail of the key obstacle and a motion trail of the automatic driving simulation equipment;
and determining scene information corresponding to the driving influence according to the obtained motion trail.
4. A method according to any one of claims 1-3, wherein said determining a plurality of interaction strategies for the interaction of the critical obstacle with the autopilot simulation device based on the scenario information corresponding to the driving impact comprises:
Based on the scene information corresponding to the driving influence, determining various interaction strategies for the interaction of the key obstacle and the automatic driving simulation equipment through Monte Carlo tree search.
5. A method according to any one of claims 1-3, wherein said determining a plurality of interaction strategies for the interaction of the critical obstacle with the autopilot simulation device based on the scenario information corresponding to the driving impact comprises:
according to the scene information corresponding to the driving influence, an interaction target of interaction between the key obstacle and the automatic driving simulation equipment in the scene indicated by the scene information is obtained;
generating a plurality of candidate interaction strategy sequences according to the interaction targets;
evaluating the multiple candidate interaction strategy sequences according to a preset cost evaluation function;
and determining a target interaction strategy sequence according to the evaluation result.
6. The method of claim 5, wherein the cost evaluation function comprises at least one of:
a first price evaluation function for evaluating the implementation efficiency of the interaction target based on the candidate interaction policy sequence;
the second cost evaluation function is used for evaluating the smoothness of the motion trail when the key obstacle moves based on the candidate interaction strategy sequence;
And a third price evaluation function for evaluating the difficulty of interaction of the key obstacle with the automatic driving simulation equipment when the candidate interaction strategy sequence is used for performing movement.
7. The method of claim 6, wherein the evaluating the plurality of candidate interaction policy sequences according to a preset cost evaluation function comprises at least one of:
when the interaction targets are realized, the first price evaluation function is used for evaluating the realization efficiency of the interaction targets realized by the multiple candidate interaction strategy sequences;
according to the acceleration and the average value of the acceleration of the key obstacle when the key obstacle moves based on a plurality of candidate interaction strategy sequences, the second cost evaluation function is used for evaluating the smoothness of the moving track of the key obstacle;
and according to the shortest interaction distance between the key obstacle and the automatic driving simulation equipment when the key obstacle moves based on a plurality of candidate interaction strategy sequences, using the third price evaluation function to evaluate the interaction difficulty between the key obstacle and the automatic driving simulation equipment.
8. The method of claim 5, wherein after the determining the target interaction policy sequence according to the evaluation result, the method further comprises:
After the target interaction strategy sequence is determined, a strategy for keeping running on the current lane is added to the key obstacle.
9. A method according to any one of claims 1-3, wherein said using a plurality of said interaction strategies for performing an interaction deduction between said critical obstacle and said autopilot simulation device comprises:
track planning is respectively carried out for a plurality of interaction strategies;
and carrying out interaction deduction between the key obstacle and the automatic driving simulation equipment according to the interaction strategies and the corresponding track plans thereof.
10. A method according to any one of claims 1-3, wherein said using a plurality of said interaction strategies for performing an interaction deduction between said critical obstacle and said autopilot simulation device comprises:
and carrying out multiple interaction deductions between the key obstacle and the automatic driving simulation equipment, wherein the multiple interaction deductions correspond to the multiple interaction strategies, by using the multiple interaction strategies.
11. A method according to any one of claims 1-3, wherein said obtaining a simulation test result from the interactive deduction result comprises:
judging whether the corresponding interaction deduction has running influence when the key obstacle interacts with the automatic driving simulation equipment or not according to the interaction deduction result, wherein the influence responsible party is the interaction of the automatic driving simulation equipment;
If so, determining the information deduced by the interaction corresponding to the interaction as the auxiliary information of the problem to be processed aiming at the automatic driving simulation equipment.
12. An electronic device, comprising: the device comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete communication with each other through the communication bus;
the memory is configured to store at least one executable instruction that causes the processor to perform operations corresponding to the method of any one of claims 1-11.
13. A computer storage medium having stored thereon a computer program which, when executed by a processor, implements the method of any of claims 1-11.
14. A computer program product comprising computer instructions that instruct a computing device to perform operations corresponding to the method of any one of claims 1-11.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310461054.7A CN116610091A (en) | 2023-04-23 | 2023-04-23 | Simulation test method, electronic equipment and computer storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310461054.7A CN116610091A (en) | 2023-04-23 | 2023-04-23 | Simulation test method, electronic equipment and computer storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116610091A true CN116610091A (en) | 2023-08-18 |
Family
ID=87677195
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310461054.7A Pending CN116610091A (en) | 2023-04-23 | 2023-04-23 | Simulation test method, electronic equipment and computer storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116610091A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118072133A (en) * | 2024-04-17 | 2024-05-24 | 中国电子科技集团公司信息科学研究院 | Automatic test method and device based on simulation deduction |
-
2023
- 2023-04-23 CN CN202310461054.7A patent/CN116610091A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118072133A (en) * | 2024-04-17 | 2024-05-24 | 中国电子科技集团公司信息科学研究院 | Automatic test method and device based on simulation deduction |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | Av-fuzzer: Finding safety violations in autonomous driving systems | |
CN108919795B (en) | Automatic driving automobile lane change decision method and device | |
CN110197027B (en) | Automatic driving test method and device, intelligent equipment and server | |
Essa et al. | Simulated traffic conflicts: do they accurately represent field-measured conflicts? | |
Furda et al. | Enabling safe autonomous driving in real-world city traffic using multiple criteria decision making | |
JP2020121714A (en) | Learning method and learning device for determining whether vehicle mode is changed from manual travel mode to autonomous travel mode or not by performing trajectory-based behavior analysis for recent travel route | |
CN110047276A (en) | Method and device for determining congestion state of obstacle vehicle and related product | |
CN116610091A (en) | Simulation test method, electronic equipment and computer storage medium | |
Klück et al. | Performance comparison of two search-based testing strategies for ADAS system validation | |
Nguyen et al. | Salvo: Automated generation of diversified tests for self-driving cars from existing maps | |
Tian et al. | MOSAT: finding safety violations of autonomous driving systems using multi-objective genetic algorithm | |
CN115402344A (en) | Parking scene simulation method and device | |
CN116011225A (en) | Scene library generation method, test method, electronic device and storage medium | |
Muzahid et al. | Learning-based conceptual framework for threat assessment of multiple vehicle collision in autonomous driving | |
CN112671487B (en) | Vehicle testing method, server and testing vehicle | |
CN114647954A (en) | Simulation scene generation method and device, computer equipment and storage medium | |
Eggert et al. | The foresighted driver: Future ADAS based on generalized predictive risk estimation | |
Furda et al. | Towards increased road safety: Real-time decision making for driverless city vehicles | |
CN117111578A (en) | Automatic driving system detection blind area guiding fuzzy test method and system | |
CN111983934A (en) | Unmanned vehicle simulation test case generation method and system | |
CN116680979A (en) | Unmanned test scene automatic generation method based on reinforcement learning | |
CN114104005B (en) | Decision-making method, device and equipment of automatic driving equipment and readable storage medium | |
CN115454082A (en) | Vehicle obstacle avoidance method and system, computer readable storage medium and electronic device | |
CN115320581A (en) | Real-time vehicle lane change risk assessment method, computer device and computer storage medium | |
Malayjerdi et al. | A Two-Layered Approach for the Validation of an Operational Autonomous Shuttle |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |