Accelerated evaluation method for credibility of human-vehicle-road collaborative autonomous driving

Abstract

The present invention proposes a method for accelerated evaluation of the credibility of human-vehicle-road collaborative unmanned driving, which includes the following steps: based on human-vehicle-road collaborative unmanned driving data and typical unmanned driving operation scenarios, test scenario use cases are generated to build a black box system test platform for unmanned driving; then, single test scenario evaluation is implemented to train a variety of classic unmanned driving black box algorithms; secondly, system accelerated test evaluation is performed, and subset simulation combined with adaptive Markov chain Monte Carlo method is used to iteratively update the sample set to achieve accelerated test evaluation of unmanned driving failure scenarios; finally, the acceleration effect is evaluated and optimized from the perspective of algorithm effectiveness and algorithm efficiency, and the accelerated evaluation results for the credibility evaluation of human-vehicle-road collaborative unmanned driving are output. The present invention is applicable to unmanned driving algorithms with black box characteristics, which improves the evaluation efficiency by 20-200 times, and has the advantages of being easy to expand, generalizable, and applicable to multiple types of typical unmanned driving evaluation tasks.

Description

Technical Field

[0001] The present invention relates to the field of unmanned driving testing and evaluation, and in particular to a method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving. Background Art

[0002] With the development of autonomous driving technology, highly automated vehicles are being implemented in real-world transportation systems to provide more convenient and safe travel options. In recent years, black-box algorithms based on deep learning and reinforcement learning have made significant progress in autonomous driving applications. While end-to-end black-box algorithms offer greater autonomous learning capabilities and the ability to process high-dimensional data compared to traditional algorithms based on predefined rules, which suffer from poor scalability and efficiency, they come at the cost of verifiability and can lead to unpredictable failures in long-tail scenarios.

[0003] Due to the huge harm caused by the effectiveness of the algorithm, it is crucial to quantify and evaluate the safety performance of the black box algorithm for autonomous driving in a human-vehicle-road collaborative environment. The performance of autonomous vehicles in boundary scenarios is an important manifestation of their credibility. However, the current evaluation methods for autonomous driving black box algorithms are still inefficient, becoming a key obstacle to the development and deployment of autonomous driving systems. In high-dimensional space, the essence of verifying the safety performance of an algorithm is a rare event estimation problem. According to a 2021 report from California, the failure rate of the most advanced autonomous vehicles has reached about 2.10 per mile. -5 Algorithm failure becomes a rare event with low probability. The key to solving the problem of autonomous driving credibility evaluation lies in overcoming the "curse of dimensionality" caused by spatiotemporal complexity and the "sparseness problem" caused by a low proportion of useful information. According to recent statistical data, autonomous vehicles require at least 11 billion miles of mileage testing to clearly demonstrate their safety. Therefore, traditional mileage-based real-vehicle testing methods are long, inefficient, and costly. This inspires us to identify more critical test scenarios in simulation environments to improve testing efficiency.

[0004] From a mathematical perspective, solutions to evaluation problems can be roughly divided into three categories: verification, falsification, and probability estimation. Verification is characterized by its rigor, proving theoretically that the system has no faults. However, this method is not applicable at a time when existing autonomous driving manufacturers mostly use end-to-end black-box algorithms. Falsification methods aim to find all failure scenarios in the parameter space. However, this method can provide less information about the test system and is only used to confirm whether the system meets the requirements. When the system operates in an unstructured environment, it is not enough to simply confirm whether a failure has occurred. Research indicates that in this case, it is more important to understand the overall frequency of failures and the relative likelihood of different failure modes, that is, the probability estimation method, which is generally more portable and scalable.

[0005] In summary, autonomous driving credibility assessment is a prerequisite for commercial application. However, autonomous driving testing faces challenges such as the curse of dimensionality, the curse of sparsity, and black-box nature. Rapidly testing and evaluating autonomous driving black-box algorithms in complex traffic environments involving human-vehicle-road collaboration is a critical challenge for autonomous driving development and commercial application. Summary of the Invention

[0006] In response to the defects of existing technologies in practical applications, the present invention provides a method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving.

[0007] The process of the present invention is as follows Figure 1 As shown, it includes the construction of typical unmanned driving test scenarios, single test scenario evaluation, system acceleration test evaluation, acceleration effect evaluation and tuning. The specific technical solution is as follows:

[0008] (1) Construction of typical test scenarios for autonomous driving

[0009] Based on real data sets (real vehicle operation data sets), this paper builds typical application scenarios of unmanned driving, such as following a vehicle, changing lanes, and mixed traffic flows; and constructs a test platform for unmanned driving black box systems, including parsing scene composition, mining scene parameters, extracting vehicle-specific behaviors, and fitting the basic distribution of parameters to generate test scenario use cases.

[0010] 1) Black box system testing platform for unmanned driving

[0011] a. Analyze scene composition

[0012] Based on selected typical autonomous driving application scenarios, the scenario elements are analyzed, considering a seven-layer scenario element model: 1-Road structure; 2-Road facilities; 3-Temporary changes to roads and facilities; 4-Traffic participants; 5-Climate environment; 6-Communication status; and 7-Vehicle status and behavior. The scenario composition is analyzed from functional scenarios to abstract scenarios.

[0013] b. Mining scene parameters

[0014] Based on the abstract scenario, we further explore scenario parameters and construct logical scenarios. We replace scenario elements with parameters that vary within a certain range. We use parameter ranges and distributions to describe certain influencing factors in the scenario, achieving parametric expression. This provides a foundation for subsequent steps to accurately describe a specific scenario or event sequence using fixed parameters.

[0015] c. Extract vehicle-specific behaviors

[0016] Specifically, data preprocessing is performed based on the actual vehicle operation data set to extract the vehicle's driving behavior in specific application scenarios as the data basis.

[0017] 2) Construction of parameter distribution of typical operation scenarios

[0018] a. Fitting parameters basic distribution

[0019] Based on the extracted vehicle driving data and scenario parameter division, combined with distribution fitting or alternative model knowledge, the basic distribution of parameters is fitted to obtain the test parameter distribution function and probability density function for typical test scenarios. Common distribution forms include multi-dimensional Gaussian distribution, Poisson distribution, gamma distribution, Weibull distribution, etc.

[0020] b. Generate test scenario use cases

[0021] Based on the obtained basic distribution and its probability density function, specific test scenarios can be generated through distribution sampling technology as input variables for subsequent tests.

[0022] (2) Single test scenario evaluation

[0023] The present invention takes into account the widespread application of end-to-end black box algorithms in the commercialization of unmanned driving, uses the industry's advanced deep reinforcement learning algorithm to replace the unmanned driving black box algorithm as the evaluation object, and designs a reward function in combination with the operating requirements of the unmanned driving system in specific application scenarios, so that the algorithm converges and can better complete the unmanned driving operation task; further, in combination with the requirements of unmanned driving credibility evaluation, the failure evaluation function and failure judgment threshold of the unmanned driving black box algorithm are designed. Ultimately, a single test scenario evaluation is achieved: the unmanned driving black box algorithm is connected to the test platform, its simulation data over a period of time is output, its failure evaluation function value is calculated, and it is compared with the failure judgment threshold, so that the evaluation result of the unmanned driving in this specific scenario is divided into "failure" or "safe".

[0024] 1) Autonomous driving black box algorithm training

[0025] Deep reinforcement learning is a popular algorithm for commercial applications of autonomous driving. Compared with traditional algorithms, it has stronger high-dimensional data processing capabilities and autonomous learning capabilities. The core purpose of reinforcement learning is to learn the optimal strategy by trying different behaviors so that the agent can obtain the maximum cumulative reward. In reinforcement learning, at time t, the agent will observe the environment state s t , and according to its existing strategy Π(a t |s t ) Select action a from state space A t At the same time, the model system will give the agent a reward r t , and then advance the state to s t+1This process is repeated in each training until the end of training. In this process, the main goal of the agent is to obtain the maximum cumulative reward, which is formulated as follows, where the discount factor α∈(0,1].

[0026]

[0027] Designing an appropriate reward function based on the operational needs of autonomous driving can improve algorithm performance and thus more accurately simulate actual commercial algorithms. For example, in the autonomous driving decision-making algorithm, the three dimensions of safety, efficiency, and comfort can be integrated; and the design can be combined with indicators such as TTC (Time to Collision), headway, and acceleration rate of change.

[0028] The data extracted in step 1 is used to train the autonomous driving black box algorithm. The training goal is to achieve high scores and convergence for each model.

[0029] 2) Failure assessment function design

[0030] In order to reliably evaluate the performance of autonomous driving black box algorithms in a human-vehicle-road collaborative environment, a credibility evaluation index system is introduced. Specifically, the credibility principles of autonomous driving include safety, efficiency, comfort, and compliance, which can specifically take into account TTC, collision frequency, operation delays, acceleration change rate, etc.

[0031] Based on the above trustworthy evaluation indicators, a function is further designed to measure the performance of the algorithm in a specific parameter scenario. After the simulation is completed, the function value is calculated based on the simulation output data and compared with the preset failure judgment threshold to determine the performance score of the algorithm in this scenario and whether it has failed.

[0032] (3) System acceleration test evaluation

[0033] The present invention adopts a method based on probability estimation to accelerate the evaluation of the failure probability of the unmanned driving black box algorithm in a human-vehicle-road collaborative environment in a simulation environment. Specifically, a Subset Simulation method (hereinafter referred to as SuS) is proposed in combination with an Adaptive Markov chain Monte Carlo method (hereinafter referred to as ApMCMC). The latter is used for intermediate process sampling, and the sampled samples are iteratively updated to achieve rapid conversion from initial high-scoring test samples to low-scoring test samples. The computational complexity is reduced by converting high-dimensional problems into multiple low-dimensional sub-problems through an iterative method, and the failure cause analysis is further realized based on the iterative process.

[0034] 1) Subset simulation combined with adaptive Markov chain Monte Carlo method

[0035] SuS is an adaptive Monte Carlo method for estimating rare failure probabilities in high-dimensional problems. It reduces computational complexity by iteratively transforming the high-dimensional problem into multiple low-dimensional subproblems. In the SuS algorithm's implementation, each subproblem requires generating conditional samples under given conditions, which can be achieved using MCMC algorithms. The specific algorithm principles are outlined below:

[0036] Performing SuS requires transforming the underlying distribution into a multidimensional standard normal distribution via a one-to-one mapping U = T(X). If the joint probability density function of X is known, the mapping should be defined using the Rosenblatt transformation. If only the marginal distributions and correlations are known, the joint PDF should be modeled using the Nataf distribution.

[0037] In SuS, a rare failure event that satisfies the condition P0(f(X)<γ) is defined as event F. The basic concept of SuS is to represent event F as M intermediate events F j The intersection of:

[0038]

[0039] Intermediate events are nested within each other, which means:

[0040]

[0041] The final failure probability is estimated as the product of the conditional probabilities:

[0042]

[0043] By selecting intermediate events so that the conditional probability P r (F j |F j-1 ) as large as possible, avoiding the need to directly deal with rare event estimation. The original problem of estimating the probability of rare failures is reduced to a series of intermediate problems corresponding to the estimation of larger conditional probabilities, starting from the higher probability level of the sub-events and gradually approaching the lower probability level of the target event.

[0044] Specifically, the probability of intermediate events is increased by adjusting the threshold of each level through scaling. Let the threshold of the intermediate event at level j be γ j , F j It can be expressed as:

[0045] F j ={X∈P0:P0(f*(X)<γ j )} (5)

[0046] γ j Satisfies the following relationship

[0047] γ0>γ1...>γ M =γ (6)

[0048] For each intermediate event, γ j The value of needs to be adaptively selected so that the conditional probability P r (F j |F j-1 ) corresponds to the selected value p0. This step is performed by continuously sampling each conditional probability density function To achieve this: that is, generate n samples {u (k) j-1 : k = 1, ..., N} And γ j Set to sample u j-1 The p0 percentile of the f(x) value.

[0049] For the subsamples in the iterative process, the MCMC method is used to analyze the samples. Sampling is performed by using the subset level j-1 that falls in F j-1 The sample {u (k) j-1 : k = 1, ..., N s} as the starting seed to simulate the state of the Markov chain. Since all seeds are based on distributed, so the chains have already reached their steady state at the beginning and no burn-in period is needed (a transition period is required before a Markov chain approximately reaches its steady state).

[0050] Based on the MCMC algorithm, the ApMCMC algorithm is proposed. The ApMCMC algorithm mainly introduces a scaling parameter β, which is updated according to the difference between the average acceptance rate and the target acceptance rate. The idea of ​​the adaptive process is to perform the simulation in steps. In each step, the conditional sampling algorithm with the same parameter β is applied to simulate N s The fraction of chains (parameter N a is defined as the update frequency of the scaling parameter). For the next N a The chain is simulated, adjusting the parameter β based on the estimated acceptance probability of the previous chain. For each subsequent level, the initial value of the scaling parameter β should be chosen to be the final value obtained in the previous level to ensure that the acceptance rate remains close to the optimal value.

[0051] Repeat the sampling and iterative process until the percentile of p0 is less than or equal to γ, which means that F M and F are already in the same failure level. In this level, the failure probability The formula can be updated as follows, where I represents an indicator function used to determine whether the sample function value is less than a threshold.

[0052]

[0053] u (k) M-1 It is from Finally, we can obtain an estimate of the failure probability by sampling the samples as follows:

[0054]

[0055] 2) Failure cause analysis

[0056] By observing the size and direction of changes in each dimension of the sample parameters during the sample iteration process, we can explore the key factors that cause the failure of the unmanned driving black box algorithm, and further analyze the failure causes of the unmanned driving black box algorithm.

[0057] (4) Acceleration effect evaluation and optimization

[0058] The present invention evaluates and optimizes the acceleration effect of the credibility acceleration evaluation algorithm; and comprehensively considers the two aspects of algorithm effectiveness and algorithm efficiency to evaluate the performance of the acceleration evaluation algorithm used in (3).

[0059] 1) Evaluation of algorithm effectiveness

[0060] The effectiveness of the algorithm is evaluated by calculating the relative error between the evaluation results of the selected method and the true probability value. First, the failure judgment threshold is selected and the true value is approximated through a large number of random Monte Carlo simulations. The SuS combined with the ApMCMC algorithm is then used to repeatedly estimate the failure probability of the algorithm multiple times (for example, 30 times), and the average value is calculated as the final evaluation probability of the algorithm. The relative error between the evaluation probability and the true probability is calculated to evaluate the effectiveness of the algorithm. The calculation formula is as follows:

[0061]

[0062] 2) Evaluation of algorithm efficiency

[0063] The efficiency of the algorithm can be divided into two main aspects:

[0064] a. Compare the selected method with the traditional Monte Carlo method in terms of estimation accuracy, estimation variance, confidence interval, and computation time.

[0065] For estimation accuracy, the relative errors of the two methods were calculated and compared under the same test samples and different failure judgment thresholds;

[0066] For the estimation variance, the dispersion of data and the stability of the estimated value can be measured. The estimation variance of the two methods under the same sample and different failure judgment thresholds is calculated and compared;

[0067] The confidence interval can reflect the uncertainty of probability estimation. The confidence interval sizes of the two methods under the same sample and different failure judgment thresholds are calculated and compared.

[0068] The calculation time can reflect the efficiency of the algorithm and calculate the time ratio of the two methods required to achieve the same level of sampling accuracy under different failure judgment thresholds.

[0069] b. Compare the selected method with other rare event simulation algorithms

[0070] In addition to the traditional Monte Carlo method, this method is compared with the SuS algorithm and other rare event simulation algorithms such as importance sampling and sequence importance sampling in various dimensions, and the efficiency of this method relative to other methods can be analyzed.

[0071] Compared with the existing technology, the trust perception method and system provided by the present invention has the following key technical points and advantages:

[0072] (1) This paper takes into account the black-box characteristics of the algorithms used in actual commercial unmanned vehicles and provides an end-to-end evaluation method.

[0073] (2) This invention addresses the dimensionality disaster and sparsity disaster and realizes the accelerated evaluation of unmanned driving black box algorithms.

[0074] (3) The method and system for accelerated evaluation of the credibility of human-vehicle-road collaborative unmanned driving proposed in the present invention are scalable and can be easily transplanted into various typical unmanned driving scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0076] Figure 1 Flowchart of the method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving provided by the present invention;

[0077] Figure 2 A schematic diagram of the vehicle following test scenario elements and their corresponding parameters provided by an embodiment of the present invention;

[0078] Figure 3 An integrated theoretical framework for unmanned driving system evaluation provided for the examples of the present invention;

[0079] Figure 4 Schematic diagram of the preceding vehicle speed dimension in a basic distribution P0 corresponding to the test sample X;

[0080] Figure 5(a) to Figure 5(b)Schematic diagram of experimental verification of the effectiveness of the acceleration evaluation algorithm provided by the example of the present invention;

[0081] Figure 6(a) to Figure 6(b) This is a schematic diagram of the experimental verification of the efficiency of the acceleration evaluation algorithm provided by the example of the present invention;

[0082] Figure 7(a) to Figure 7(b) The second schematic diagram of the experimental verification of the efficiency of the acceleration evaluation algorithm provided by the example of the present invention;

[0083] Figure 8 The accelerated evaluation results of the credibility evaluation of the human-vehicle-road collaborative autonomous driving decision-making algorithm provided by the example of the present invention;

[0084] Figure 9 for Figure 4 Schematic diagram of the initial sample u0 corresponding to the basic distribution P0 for the following vehicle velocity and leading vehicle acceleration slices. DETAILED DESCRIPTION

[0085] The present invention relates to a method for accelerating the evaluation of the credibility of human-vehicle-road collaborative unmanned driving. The following figure further illustrates the specific implementation of the present invention in conjunction with the accompanying drawings and specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be pointed out that for those of ordinary skill in the art, several changes and improvements can be made without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0086] This embodiment is based on the real data set NGSim-I80. Taking the decision-making part as the "brain" of the three major components of the unmanned driving system, namely perception, decision-making, and control, as an example, this embodiment realizes the accelerated evaluation of the credibility evaluation of the intelligent decision-making algorithm for human-vehicle-road collaborative unmanned driving, carries out the construction of vehicle-following test scenarios, black box decision algorithm training based on deep reinforcement learning, accelerated test evaluation of decision algorithms, algorithm performance tuning, etc., and builds a theoretical system framework for unmanned driving evaluation. Figure 3 As shown, the specific implementation method includes the following steps:

[0087] S1: Construction of typical autonomous driving test scenarios

[0088] like Figure 2 As shown in the figure, the classic following scene is selected as the test scene, in which the front car simulates other vehicles in the actual road environment, and the rear car simulates an autonomous driving vehicle, which is controlled by a black box decision algorithm based on deep reinforcement learning.

[0089] The key test parameters of the scene are four-dimensional, including the initial speed LvSpeed0 of the front car, the initial speed FvSpeed0 of the rear car, the initial distance Distance0 between the two cars, and the acceleration a of the front car. Lv.

[0090] During the simulation, the black box decision algorithm of the following vehicle takes the current four-dimensional parameters as input and outputs the acceleration a of the following vehicle at the next time step. RL , and the remaining parameters are changed according to physical formulas.

[0091] Assume that the scene is updated at intervals of Δt, and a total of i time intervals have passed. The total length of the simulation is ΔT, and the speed difference between the two vehicles in the i-th time interval is relSpd i , so the update formula of the scene parameters is as follows:

[0092]

[0093] When sampling the test scene, the test sample X It can be expressed as:

[0094] X=(Distance0, LvSpeed0, FvSpeed0, a Lv ) (11)

[0095] This step is based on the real dataset NGSim-I80. A total of 1341 car-following data and 323739 real-time data on the speed of the front and rear vehicles and the following distance are extracted. The distributed environment of human-vehicle-road collaboration is constructed by the distribution fitting method, and the basic distribution P0 is obtained. The distribution fitting diagram of the front vehicle speed is shown in the figure below. Figure 4 The specific test scenario is obtained by sampling parameters from the distribution.

[0096] S2: Single test scenario evaluation

[0097] This step first trains four classic deep reinforcement learning models to simulate the black-box decision-making algorithm of autonomous vehicles, which will serve as the basis for subsequent credibility acceleration testing and evaluation. The reinforcement learning reward function is designed by comprehensively considering the three major indicators of safety, efficiency, and comfort, as follows:

[0098] 1) Security

[0099] Safety is the most critical factor that autonomous vehicles consider in their subsequent decision-making process and is one of the main training objectives of reinforcement learning. The time-to-collision (TTC) metric is used to describe safety, and penalties should be imposed on the following decision-making behaviors with TTC values ​​less than 4 seconds. TTC is a type of safety surrogate measure (SSM) and is the most common indicator for quantifying traffic conflict risk. It represents the remaining time before two vehicles collide. TTC and safety indicator F TTC The formula is as follows:

[0100]

[0101] In addition, considering the serious negative impact of vehicle collision accidents on safety, a collision penalty term is introduced. The penalty coefficient is represented by λ, which is 100. Whether a collision accident occurs in a single simulation process is represented by the Boolean value judgecollision. If a collision occurs, the value is 1, otherwise it is 0.

[0102] 2) Efficiency

[0103] Headway is an important indicator of road capacity. Efficient following is achieved by maintaining a shorter following distance or headway while ensuring safety. A logarithmic distribution of headway intervals based on the NGSIM-I80 dataset is constructed, and a probability density function is used as the efficiency metric. Given the priority of safety, efficiency rewards are only designed when the safety metric is zero. Headway h t and efficiency index F headway The calculation is as follows:

[0104]

[0105] Among them, formula (14) is a theoretical formula, t represents the current time, that is, the headway at the current time is equal to the distance between the two vehicles at the current time divided by the speed of the rear vehicle at the current time, and the speed of the rear vehicle at the current time FvSpeed t The value is the speed of the following vehicle corresponding to Train in formula (10), that is, the speed of the following vehicle in the real dataset NGSim-I80.

[0106] About formula (15)f lognorm The explanation is as follows: Based on the headway data of 1341 car-following data in the real data set, the lognormal distribution is fitted. The lognormal distribution is a probability distribution whose logarithm is normally distributed. headway According to the current time headway h t And the lognormal distribution f lognorm μ, σ are the parameters of the exponential normal distribution, with values ​​of μ = 0.42 and σ = 0.44.

[0107] 3) Comfort

[0108] Acceleration rate is a key indicator that significantly affects passenger comfort. When subjected to a significant acceleration rate, passengers will experience noticeable discomfort. Using it as an indicator to quantify comfort, the comfort index can be expressed as follows:

[0109]

[0110] The final expression of the reward function is as follows:

[0111] reward = 10 FTTC +F headway +F jerk -λ·judgecollision (17)

[0112] During training, the vehicle-following data extracted from NGSim-I80 is used as the training dataset, so that the RollingScore of the model to be evaluated converges and shows satisfactory performance in the given test scenario.

[0113] Finally, to evaluate the safety performance of the intelligent decision-making algorithm in a specific test scenario, a failure assessment function is designed. In a single test scenario, a safety assessment is performed on the autonomous driving black box decision-making algorithm connected to the test scenario to obtain the safety assessment function value. This step can extend the safety assessment function to other evaluation indicators such as the interaction complexity function. The safety assessment function is designed as follows:

[0114] Assume the simulation interval is 0.1s, or 10Hz, and the simulation duration is 10s. The simulation terminates when the total duration is reached or when the two vehicles collide. Scenario samples sampled from the basic distribution are input into the simulation environment. After the simulation test begins, the simulated unmanned vehicles interact through an intelligent decision-making algorithm and output a time series of the following scenario parameters during the simulation. The TTC value sequence for the entire process is calculated and its minimum value is set as the safety assessment function value during the simulation. The safety assessment function formula is as follows:

[0115]

[0116] At this point, single test scenario evaluation is achieved, and the black-box decision algorithm is connected to the vehicle-following simulation environment. The specific test scenario is obtained by sampling the basic distribution of parameters fitted in the vehicle-following simulation environment and input into the black-box decision algorithm and simulation environment. The black-box decision algorithm generates the acceleration of the following vehicle based on the scenario, and runs the simulation process until it stops according to the pre-defined scenario parameter update formula to obtain the simulation output result.

[0117] The safety assessment function is calculated and compared with the failure judgment threshold. If it is less than, the black box decision algorithm fails in the sampling test scenario, and the scenario is the "failure scenario" of the algorithm; if it is greater than or equal to, the black box decision algorithm makes a safe decision in the sampling test scenario.

[0118] S3: System acceleration test evaluation

[0119] This step uses subset simulation combined with the adaptive Markov chain Monte Carlo method to effectively estimate the failure rate of the autonomous driving black box decision algorithm in the simulation scenario. It can be mainly divided into the following three steps:

[0120] S31: Initialization settings

[0121] 1) Initialization parameters, where the number of samples in each intermediate step is set to N = 1000, the subset threshold probability is set to p0 = 0.1, the number of simulation rounds is set to NSIM = 30, and the MCMC method scaling factor is β.

[0122] 2) Design a distribution transformation function to convert samples from any distribution into samples from a standard normal distribution, and vice versa. Specifically, the base distribution P0 is first converted into a standard distribution for updating samples. When calculating the value of the security assessment function or generating the final output sample, the converted sample is reconverted back to a sample from the original distribution.

[0123] S32: Monte Carlo method to generate initial samples

[0124] Generate an initial sample u0 from the basic distribution P0. Use the security evaluation function f(x) to calculate the corresponding function value, denoted as geval0. In addition, calculate its percentile γ of p0(0.1) j Based on the initial sample u0, define the event F0.

[0125] S33: Iteration

[0126] When γ j >γ, iterate

[0127] 1) From the previous level sample u j Select N s = p0N low-value samples, and then use adaptive MCMC to iteratively generate new samples u for each sample j+1 , and update the value of the MCMC scaling β.

[0128] 2) Recalculate geval j+1 and γ j+1 .

[0129] S34: Estimating Failure Probability

[0130] The failure probability is estimated according to the following formula Finally, the number of samples whose security assessment function value is less than γ in the last sampling is recorded as N f

[0131]

[0132] Ultimately, we average the probabilities obtained from the 30 simulations as our final failure probability estimate.

[0133] S4: Acceleration effect evaluation and optimization

[0134] This step evaluates the algorithm performance from two aspects: the effectiveness and efficiency of the rare event simulation algorithm.

[0135] S41: Algorithm Validity

[0136] The effectiveness of the algorithm is proved by calculating the relative error between the evaluation results of the SuS method combined with the ApMCMC method and the true probability value.

[0137] The fault detection threshold is set to γ ​​= 1, and the true value is approximated by 5 million random Monte Carlo simulation iterations. The SuS algorithm is used to repeatedly estimate the algorithm failure probability 30 times, and each evaluation result is recorded. The results are as follows Figure 5(a) to Figure 5(b) As shown in the figure, the final estimated relative errors were 0.65% and 2.84%, respectively, both within a relatively small range. This demonstrates that the example algorithm can accurately estimate the probability of rare events, effectively evaluating the failure probability of the intelligent decision-making algorithm in the vehicle-following scenario.

[0138] Figure 5(a) to Figure 5(b) In the figure, the horizontal axis represents the number of repeated experiments; the vertical axis represents the estimated failure probability of this time; Pf-episode represents the estimated failure probability in each repeated experiment; Pf-mean represents the average value of the estimated failure probability of thirty repeated experiments; Pf-real represents the true value of the failure probability of the algorithm.

[0139] S41: Algorithm Efficiency

[0140] The effectiveness of the algorithm is demonstrated by comparing it with the traditional probabilistic Monte Carlo (MC) method, using the same test samples. This example compares SuS+ApMCMC with the MC method in terms of estimation accuracy, estimation variance, and confidence intervals. The comparison of estimation error and confidence intervals is shown in Figure 6. The experiments demonstrate that the proposed method achieves several to dozens of times the advantages of the traditional Monte Carlo method in all of these indicators.

[0141] In addition, by comparing the confidence interval and variance ratio of MC, SuS, SuS+ApMCMC, and another classic rare event simulation algorithm, Sequential Importance Simulation (SiS), at different thresholds, as shown in Figure 7, it can be found that the SiS algorithm underestimates the failure probability at low thresholds, and ApMCMC+SuS exhibits a significantly smaller variance among the other three algorithms, indicating its excellent evaluation performance and proving the algorithm's high efficiency.

[0142] Figures 6(a) to 7(b) In the figure, the horizontal axis represents the value of the failure probability γ; the vertical axis in Figure 6(a) represents the estimated error value under each failure threshold; Figure 6(b) to Figure 7(a)The vertical axis represents the estimated value of failure probability at each failure threshold; the vertical axis of Figure 7(b) represents the estimated variance ratio of the Monte Carlo algorithm and the experimental algorithm.

[0143] In Figure 6(a), “MC for DDPG” represents the experimental results of applying the MC algorithm to the DDPG algorithm; “SuS for DDPG” represents the experimental results of applying the SuS algorithm to the DDPG algorithm.

[0144] In Figure 6(b), “99% Confidence Interval for MC” represents the 99% confidence interval estimated by the MC algorithm; “99% Confidence Interval for SuS” represents the 99% confidence interval estimated by the SuS algorithm; “Probalility Real” represents the true probability value; γ test That is, it represents the horizontal axis.

[0145] In Figure 7(a), “SuS with Adaptive MCMC” represents the estimation result using the ApMCMC+SuS algorithm; “SuS with MCMC” represents the estimation result using the MCMC+SuS algorithm; “SiS withAdaptive MCMC” represents the estimation result using the ApMCMC+SiS algorithm; “Probability Real forDDPG” represents the true probability value of the DDPG algorithm; γ test That is, it represents the horizontal axis.

[0146] In Figure 7(b), “MC / SuS with Adaptive MCMC” represents the ratio of the estimated variances of the MC algorithm to that of the ApMCMC+SuS algorithm; “MC / SuS with MCMC” represents the ratio of the estimated variances of the MC algorithm to that of the MCMC+SuS algorithm; and “SiS with Adaptive MCMC” represents the ratio of the estimated variances of the MC algorithm to that of the ApMCMC+SiS algorithm.

[0147] So far, we have completed the accelerated evaluation of the credibility evaluation of the decision-making algorithm for human-vehicle-road collaborative autonomous driving. Specifically, we use the classic deep reinforcement learning DDPG (Deep Deterministic Policy Gradient) and A3C (Asynchronous Advantage Actor Critic) as examples. The evaluation results are as follows: Figure 8 shown.

[0148] Figure 8In the figure, “99% Confidence Interval for DDPG MC” represents the 99% confidence interval of the failure probability of the DDPG algorithm using the MC algorithm;

[0149] "99% Confidence Interval for DDPG SuS" indicates the 99% confidence interval of the DDPG algorithm failure probability estimated using the SuS algorithm. "99% Confidence Interval for A3C MC" indicates the 99% confidence interval of the A3C algorithm failure probability estimated using the MC algorithm. "99% Confidence Interval for A3C SuS" indicates the 99% confidence interval of the A3C algorithm failure probability estimated using the SuS algorithm. "Probability Real for A3C" indicates the actual failure probability of the A3C algorithm. "Probability Real for DDPG" indicates the actual failure probability of the DDPG algorithm.

Claims

1. A method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving, characterized by: The following steps are involved: 1) Construction of typical autonomous driving test scenarios: 1.1) Building a black-box system test platform for autonomous driving: Based on real datasets, we build typical autonomous driving application scenarios. Based on the selected typical autonomous driving application scenarios, we analyze the scenario elements. Based on the abstract scenarios, we further explore the scenario parameters and construct logical scenarios. Based on the real datasets, we perform data preprocessing and extract the vehicle's driving behavior in specific application scenarios as the data foundation. 1.2) Construction of parameter distributions for typical operating scenarios: Based on the extracted vehicle driving data and scenario parameter divisions, the basic distribution of parameters is fitted. Specific test scenarios are generated through distribution sampling techniques and serve as input variables for subsequent tests. 2) Single test scenario evaluation: 2.2) Autonomous driving black box algorithm training: Use deep reinforcement learning algorithms instead of autonomous driving black box algorithms as the evaluation object, and design reward functions based on the operating requirements of autonomous driving systems in specific application scenarios to achieve algorithm convergence; 2.2) Failure assessment function design: The performance of autonomous driving black-box algorithms in a human-vehicle-road collaborative environment is credibly evaluated, introducing a credibility evaluation index system. Based on the credibility evaluation indicators, a function is further designed to measure the algorithm's performance in specific parameter scenarios. After the simulation is completed, the function value is calculated based on the simulation output data and compared with the preset failure threshold to determine the algorithm's performance score and failure in the scenario. 3) System acceleration test evaluation: 3.1) Subset simulation combined with adaptive Markov chain Monte Carlo method: Based on a probability estimation method, a subset simulation combined with an adaptive Markov chain Monte Carlo method is proposed to accelerate the evaluation of the failure probability of autonomous driving black box algorithms in a human-vehicle-road collaborative environment in a simulation environment. The subset simulation combined with the adaptive Markov chain Monte Carlo method comprises the following steps: 4.1) Initialization settings: including initializing parameters and designing distribution transformation functions; 4.2) Generate initial samples: Generate initial samples using random Monte Carlo method; 4.3) Iterative update: Iteratively update samples based on pre-defined evaluation criteria; 4.4) Estimation of failure probability: Calculate the failure probability according to the formula; 3.2) Failure cause analysis: By observing the size and direction of changes in each dimension of sample parameters during sample iteration, failure cause analysis of the autonomous driving black box algorithm can be achieved; 4) Acceleration effect evaluation and optimization: 4.1) Evaluation of algorithm effectiveness: Evaluate and optimize the acceleration effect of the credibility acceleration evaluation algorithm from the perspective of algorithm effectiveness, including calculating the relative error between the evaluation results of the selected method and the true probability value for evaluation; 4.2) Evaluation of algorithm efficiency: The acceleration effect evaluation and tuning of the credibility acceleration evaluation algorithm are carried out from the perspective of algorithm efficiency, including calculating the differences between the selected method and the traditional Monte Carlo method in estimation accuracy, estimation variance, confidence interval, and calculation time, as well as the differences between the selected method and other rare event simulation algorithms in the above indicators.

2. The method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving according to claim 1, characterized in that: The construction of the unmanned driving black box system test platform includes the following steps: 2.1) Analyze scenario composition: Analyze the scenario elements based on the selected typical autonomous driving application scenarios; 2.2) Mining scenario parameters: Based on the abstract scenario, further mining scenario parameters is performed to construct a logical scenario. 2.3) Extracting vehicle-specific behaviors: Based on real-vehicle operation datasets, extract the vehicle’s driving behaviors in specific application scenarios.

3. The method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving according to claim 1, characterized in that: The unmanned driving black box algorithm training includes the following steps: 3.1) Training data selection: Extract vehicle behaviors based on real datasets as training dataset input; 3.2) Reward Function Design: Design a reward function based on safety, efficiency, and comfort metrics. 3.3) Score Convergence: Modify hyperparameters and train the model until the score converges.

4. The method for accelerating the credibility evaluation of human-vehicle-road collaborative unmanned driving according to claim 1, characterized in that: The algorithm efficiency evaluation includes the following steps: 5.1) Comparison with traditional Monte Carlo methods: Comparison in terms of estimation accuracy, estimation variance, confidence interval, and computation time; 5.2) Comparison with other rare event simulation algorithms: Compare with the SuS algorithm and other rare event simulation algorithms such as importance sampling and sequence importance sampling methods in various dimensions.

Images