A performance evaluation method, device and storage medium for a vehicle following control algorithm

Abstract

The present disclosure provides a performance evaluation method for a vehicle following control algorithm, including: constructing first and second evaluation indicators for evaluating the control effect and application characteristics of the following control algorithm, respectively. The first evaluation indicator is a comprehensive indicator obtained by summing the following, comfort, and economy indicators of the following control algorithm obtained based on evaluation data when sensor filtering is used, with weight coefficients determined by the hierarchical analysis method. The second evaluation indicator is composed of real-time and anti-interference evaluation indicators. The real-time evaluation indicator is calculated based on the single-step calculation duration of calling the following control algorithm, and the anti-interference evaluation indicator is obtained based on the growth rate of the first evaluation indicator when no sensor filtering is used. The following control algorithm is screened from all following control algorithms for those that meet the requirements for both real-time and anti-interference evaluation indicators, and then the following control algorithm with the smallest first evaluation indicator is selected. The present disclosure can objectively and effectively evaluate the control effect and application characteristics of the following control algorithm.

Description

Technical Field

[0001] The present disclosure belongs to the technical field of intelligent connected vehicles, and in particular relates to a performance evaluation method, device, and storage medium for a vehicle following control algorithm. Background Art

[0002] Following control is a core function of adaptive cruise control, requiring the vehicle to maintain a reasonable speed and distance while dynamically following the vehicle ahead. However, current domestic and international testing standards for adaptive cruise control, such as the current domestic test standard GBT 20608-2006 Intelligent Transport Systems Adaptive Cruise Control Systems Performance Requirements and Test Methods and the current international test standard ISO 15622-2018 Intelligent transport systems—Adaptive cruise control systems—Performance requirements and test procedures, only set minimum requirements for the completion of following control and do not provide a standardized quantitative evaluation method for algorithm performance. This hinders the comparison and selection of following control algorithms.

[0003] Existing performance evaluation methods for vehicle-following control algorithms mainly appear in academic research, but they have the following two main problems:

[0004] First, the design of evaluation indicators for control effectiveness is incomplete. Some work focuses on functional achievement, without quantitatively evaluating the performance of the algorithm. Most other work tends to only evaluate certain aspects of followability, comfort, and economy, and the definitions of evaluation indicators vary. A quantitative, scientific evaluation system has not been established. For example, some evaluate followability based on the error distribution of macro indicators such as vehicle distance and speed, some evaluate comfort based on the numerical distribution of acceleration and jerk, and some evaluate economy based on actual fuel consumption. However, no work has fully defined these three evaluation indicators and integrated them into a comprehensive evaluation index.

[0005] Second, there's a significant gap in the evaluation of application characteristics. Almost all work focuses solely on evaluating the control effectiveness of vehicle-following control algorithms, while neglecting their application characteristics. However, in actual algorithm deployment, application characteristics such as real-time performance and interference immunity often take precedence over control effectiveness. The lack of application characteristic evaluation in existing methods significantly impacts the assessment of the algorithm's performance in real-world vehicles. Summary of the Invention

[0006] The present disclosure aims to solve at least one of the technical problems existing in the prior art.

[0007] To this end, the first embodiment of the present disclosure establishes a performance evaluation method for a vehicle following control algorithm that can scientifically and effectively assess the algorithm's control effectiveness and application characteristics. Regarding control effectiveness, this method quantitatively evaluates the algorithm's performance in various directions using the following performance evaluation index TRA, the comfort evaluation index COM, and the economy evaluation index ECO, and reflects the algorithm's overall performance using a comprehensive evaluation index CEI. Regarding application characteristics, this method evaluates the algorithm's real-time performance using the average single-step calculation time h, and its anti-interference performance using the CEI index's growth rate α when sensorless filtering is used.

[0008] In order to achieve the above objectives, the present disclosure adopts the following technical solutions:

[0009] A first aspect of the present disclosure provides a performance evaluation method for a vehicle following control algorithm, comprising:

[0010] Several vehicle-following control algorithms are deployed on test vehicles equipped with the same sensors and sensor signal filtering algorithms, and the sensor signal acquisition period T is set. c And the following control algorithm call cycle T s , T c <T s , in the preset vehicle following test scenario, control each test vehicle to reach the set initial speed and initial vehicle distance;

[0011] Control each test vehicle to enter the automatic driving mode, every T c Collect the original data during the test and output them in T s The vehicle controller records the single-step calculation duration of each call to the vehicle-following control algorithm for evaluation data with and without sensor filtering. The evaluation data includes the speed of the leading vehicle, the speed, acceleration, and energy consumption of the test vehicle, and the distance between the two vehicles.

[0012] A first evaluation index for evaluating the control effect of the vehicle-following control algorithm and a second evaluation index for evaluating the application characteristics of the vehicle-following control algorithm are respectively constructed. The first evaluation index is a comprehensive index obtained by summing the following index, comfort index, and economy index of the vehicle-following control algorithm obtained based on evaluation data when sensor filtering is used, and assigning weight coefficients determined by the hierarchical analysis method. The second evaluation index is composed of a real-time evaluation index and an anti-interference evaluation index. The real-time evaluation index is calculated based on the single-step calculation time of calling the vehicle-following control algorithm, and the anti-interference evaluation index is obtained based on the growth rate of the first evaluation index when sensorless filtering is used.

[0013] Calculating the first evaluation index and the second evaluation index of each vehicle following control algorithm based on the evaluation data of the entire vehicle following test process;

[0014] From all the following control algorithms, the following control algorithms that meet both the real-time evaluation index and the anti-interference evaluation index are selected as candidate following control algorithms, and the candidate following control algorithm with the smallest first evaluation index is selected as the optimal following control algorithm.

[0015] Optionally, the first evaluation index is constructed according to the following formula:

[0016]

[0017] ω tra +ω com +ω eco =1

[0018]

[0019]

[0020]

[0021] Among them, CEI is the first evaluation index; TRA, COM and ECO are the followability index, comfort index and economy index of the vehicle following control algorithm respectively; ω tra 、ω com and ω eco are the weight coefficients of followability index, comfort index and economy index respectively; and They are the followability objective function, comfort objective function and economy objective function after standardization respectively.

[0022] Optionally, the standardization processing method used for each objective function is a Z-Score method.

[0023] Optionally, the followability objective function before normalization is assumed to be L tra , whose expression is:

[0024] L tra =ω d Δd 2 +ω v Δv 2

[0025] Wherein, Δd is the inter-vehicle distance error, which is defined as the difference between the actual inter-vehicle distance between the test vehicle and the preceding vehicle and the desired inter-vehicle distance, where the desired inter-vehicle distance uses a fixed inter-vehicle time interval; Δv is the speed error, which is defined as the difference in speed between the test vehicle and the preceding vehicle; ω d and ω v are weight coefficients for vehicle spacing error and vehicle speed error respectively; and / or

[0026] Assume that the comfort objective function before standardization is L com , whose expression is:

[0027]

[0028] Among them, a f To test the acceleration of the vehicle; and / or

[0029] Assume that the economic objective function before standardization is L eco , whose expression is:

[0030]

[0031] Where, P is the instantaneous power consumption of the test vehicle, v f To test the vehicle's speed.

[0032] Optionally, the weight coefficients ω of the followability index, comfort index and economy index in the first evaluation index CEI are determined by the hierarchical analysis method. tra 、ω com and ω eco The steps include:

[0033] Step S311: Establishing a hierarchical structure model, wherein the hierarchical structure model is a single-layer structure model having a followability target, a comfort target, and an economy target in a criterion layer;

[0034] Step S312: construct a judgment matrix for quantifying the relative importance of different targets through the importance scale table. Let the element in the i-th row and j-th column of the judgment matrix be a ij , represents the relative importance of target i to target j, i, j∈[1, 2, 3], and satisfies: a ii =1,

[0035] S313, calculating the maximum eigenvalue λ and the corresponding eigenvector of the judgment matrix, normalizing the eigenvector, and obtaining initial weight coefficients of the followability target, the comfort target, and the economy target;

[0036] S314. Perform a consistency check on the discriminant matrix. If the consistency check passes, the initial weight coefficients of the followability target, comfort target, and economy target obtained in step S313 are used as the final weight coefficients of each target. If the consistency check fails, return to step S312 to reconstruct the discriminant matrix.

[0037] Optionally, the average value of the single-step calculation duration of each call of the vehicle-following control algorithm during the vehicle-following test is used as the real-time evaluation index.

[0038] Optionally, the anti-interference evaluation index is calculated according to the following formula:

[0039]

[0040] Wherein, α is the anti-interference evaluation index, that is, the growth rate of the first evaluation index when there is no sensor filtering; CEI F CEI is the statistical value of the first evaluation index of the whole process of the vehicle following test; NF It is the statistical value of the third evaluation index of the entire vehicle following test process. The third evaluation index is a comprehensive index obtained by summing up the followability index, comfort index and economy index of the vehicle following control algorithm obtained based on the evaluation data during sensorless filtering, and assigning weight coefficients determined by the hierarchical analysis method.

[0041] Optionally, the real-time evaluation index is required to be no more than T s 70%, requiring that the anti-interference evaluation index does not exceed 10%.

[0042] The performance evaluation method of a vehicle following control algorithm provided by the embodiments of the present disclosure has the following characteristics and beneficial effects:

[0043] 1. Scientifically and quantitatively reflect the control effect of the following control algorithm. During the following process, objective functions for following, comfort, and economy are established based on the state variables of the two vehicles, and each objective function is processed using the Z-score standardization method. Then, based on the objective functions, three discrete evaluation indicators are defined: the following evaluation indicator TRA, the comfort evaluation indicator COM, and the economy evaluation indicator ECO, which have the practical advantage of quantitative comparability. Furthermore, the hierarchical analysis method is used to determine the weights of each evaluation indicator, thereby defining the comprehensive evaluation indicator CEI. This definition method has three major advantages: first, the comprehensive evaluation indicator can be used to measure the overall quality of the following control algorithm's control effect; second, the discrete evaluation indicators can be used to observe the specific performance of the algorithm in terms of following, comfort, and economy; and third, the algorithm can reversely determine the objective function based on the evaluation indicators, further improving its overall performance.

[0044] 2. Comprehensively and effectively reflect the application characteristics of the vehicle-following control algorithm. Incorporating real-time performance and anti-interference performance into the evaluation of the vehicle-following control algorithm's application characteristics effectively evaluates its deployment performance in actual applications. The algorithm's real-time performance is measured by the average single-step calculation time, which provides a clear understanding of actual computing resource consumption. The algorithm's anti-interference performance is measured by the growth rate of the comprehensive evaluation indicator CEI (CEI) in the absence of sensor filtering, which determines the algorithm's resistance to sensor errors.

[0045] In summary, the performance evaluation method provided by the embodiment of the present disclosure can effectively evaluate the control effect and application characteristics of the vehicle following control algorithm, and select the algorithm that meets the application requirements and has the best comprehensive performance from a variety of alternative algorithms, thereby realizing reliable intelligent vehicle following control.

[0046] A second aspect of the present disclosure provides a performance evaluation device for a vehicle following control algorithm, comprising:

[0047] The evaluation data acquisition module is configured to acquire evaluation data in the following manner: deploying several vehicle following control algorithms on test vehicles equipped with the same sensors and sensor signal filtering algorithms, setting the sensor signal acquisition period T c And the following control algorithm call cycle T s , T c <T s In the preset vehicle following test scenario, each test vehicle is controlled to reach the set initial speed and initial vehicle distance, and each test vehicle is controlled to enter the automatic driving mode. c Collect the original data during the test and output them in T s The vehicle controller records the single-step calculation duration of each call to the vehicle-following control algorithm for periodic evaluation data with and without sensor filtering. The evaluation data includes the speed of the leading vehicle, the speed, acceleration, and energy consumption of the test vehicle, and the distance between the two vehicles.

[0048] An evaluation index calculation module is configured to calculate a first evaluation index and a second evaluation index for each vehicle following control algorithm based on evaluation data from the entire vehicle following test process; the first evaluation index is used to evaluate the control effect of the vehicle following control algorithm, and is a comprehensive index obtained by summing the followability index, comfort index, and economy index of the vehicle following control algorithm obtained based on the evaluation data when sensor filtering is used, with weight coefficients determined by the hierarchical analysis method; the second evaluation index is used to evaluate the application characteristics of the vehicle following control algorithm, the real-time evaluation index is calculated based on the single-step calculation time of calling the vehicle following control algorithm, and the anti-interference evaluation index is obtained based on the growth rate of the first evaluation index when sensor-free filtering is used;

[0049] The evaluation module is configured to screen out the vehicle following control algorithms that meet the requirements of real-time evaluation index and anti-interference evaluation index from all vehicle following control algorithms as candidate vehicle following control algorithms, and select the candidate vehicle following control algorithm with the smallest first evaluation index as the optimal vehicle following control algorithm.

[0050] A third aspect of the present disclosure provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable the computer to execute the performance evaluation method of the vehicle following control algorithm described in any embodiment of the first aspect of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of a performance evaluation method for a vehicle following control algorithm provided in an embodiment of the first aspect of the present disclosure.

[0052] Figure 2 Schematic diagram of a vehicle following control scenario involved in the performance evaluation method provided by an embodiment of the first aspect of the present disclosure;

[0053] Figure 3 This is a diagram of the actual vehicle test vehicle used in the analysis example of this disclosure;

[0054] Figure 4 is the actual vehicle test road map used in the analysis example of this disclosure;

[0055] Figure 5 This is the speed of the leading vehicle used in the actual vehicle test of the four following control algorithms used in the analysis example of this disclosure.

[0056] Figure 6 It is a structural diagram of an electronic device provided in an embodiment of the third aspect of the present disclosure. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of this application more clearly understood, this application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0058] On the contrary, this application covers any alternatives, modifications, equivalents, and solutions made within the spirit and scope of this application as defined by the claims. Furthermore, to facilitate a better understanding of this application, certain specific details are described in detail below in the detailed description of this application. Those skilled in the art will be able to fully understand this application without these details.

[0059] See also Figure 1 The performance evaluation method of the vehicle following control algorithm provided in the embodiment of the first aspect of the present disclosure can screen out the following control algorithm that meets the application characteristic requirements and is optimal in control effect from a series of following control algorithms, so as to deploy the optimal following control algorithm on the vehicle to achieve stable, robust, close following, safe, comfortable, economical and energy-saving following control. Among them, the control effect evaluation index includes three separate evaluation indicators, namely the following evaluation index TRA, the comfort evaluation index COM and the economy evaluation index ECO, and the comprehensive evaluation index constructed therefrom; the application characteristic evaluation index is composed of the real-time and anti-interference indicators of the algorithm. The performance evaluation method provided in the embodiment of the first aspect of the present disclosure includes:

[0060] Step S1: deploy several vehicle-following control algorithms on test vehicles equipped with the same sensors and sensor signal filtering algorithms, and set the sensor signal acquisition period T c = 20ms and the vehicle following control algorithm call cycle T s = 100ms, in the preset vehicle following test scenario, control each test vehicle to reach the set initial speed and initial vehicle distance;

[0061] Step S2: Control each test vehicle to enter the automatic driving mode, and c Collect the original data during the test and output them in T s The vehicle controller records the single-step calculation duration of each call to the vehicle-following control algorithm for periodic evaluation data with and without sensor filtering. The evaluation data includes the speed of the leading vehicle, the speed, acceleration, and energy consumption of the test vehicle, and the distance between the two vehicles.

[0062] Step S3: constructing a first evaluation index CEI for evaluating the control effect of the vehicle-following control algorithm and a second evaluation index for evaluating the application characteristics of the vehicle-following control algorithm. The first evaluation index CEI is a comprehensive index obtained by summing the followability, comfort, and economy of the vehicle-following control algorithm obtained based on the evaluation data when sensor filtering is used, and assigning weight coefficients determined by the hierarchical analysis method to the indicators. The second evaluation index consists of a real-time evaluation index and an anti-interference evaluation index. The real-time evaluation index is calculated based on the single-step calculation time of calling the vehicle-following control algorithm, and the anti-interference evaluation index is obtained based on the growth rate α of the first evaluation index CEI when sensorless filtering is used.

[0063] Step S4, calculating the first evaluation index CEI and the second evaluation index of each vehicle following control algorithm;

[0064] Step S5: Filter out the following control algorithms that meet both the real-time evaluation index and the anti-interference evaluation index from all the following control algorithms as candidate following control algorithms, and select the candidate following control algorithm with the smallest first evaluation index CEI as the optimal following control algorithm.

[0065] Optionally, in step S1, the preset vehicle following test scenario is a long straight road with a flat road and no other interference factors, and only the longitudinal control of the vehicle is considered (unless otherwise specified, the speed and distance between vehicles recorded in this application refer to the longitudinal speed and longitudinal distance between vehicles), which can be used Figure 2 Description. The leading vehicle is the target vehicle and operates according to its set driving mode. The ego vehicle (i.e., the following test vehicle) uses sensor information such as the speed and distance between the two vehicles to output a control variable to achieve the following task. The control variable can be derived from the vehicle's desired acceleration or throttle / brake settings. Figure 2 In, v fand v p Represents the speed of the vehicle and the preceding vehicle respectively; d is the actual distance between the two vehicles, d des is the desired vehicle spacing; Δd = dd des , is the inter-vehicle distance error, which reflects the difference between the actual inter-vehicle distance and the expected inter-vehicle distance.

[0066] The expected vehicle distance d in the above following test scenario des The desired inter-vehicle distance (DIT) is a key parameter in following vehicle control, impacting vehicle safety, following smoothness, and road capacity. Therefore, designing a reasonable desired inter-vehicle distance is crucial. A too small DIT leaves little margin for the preceding vehicle. This, combined with the inherent latency of the vehicle control system, can lead to frequent acceleration and deceleration during following, increasing the probability of accidents. On the other hand, an excessively large DIT reduces lane space utilization, easily causing vehicles on either side to change lanes and squeeze in, reducing road capacity.

[0067] Expected vehicle spacing d des There are two main ways of definition: fixed vehicle spacing and fixed time headway. Fixed vehicle spacing means that the vehicle and the preceding vehicle maintain a constant safe distance, which weakens the specific control requirements for the vehicle and is suitable for application in macro and complex traffic environments. The fixed time headway adds the vehicle speed factor, which can dynamically adjust the size of the expected vehicle spacing according to the vehicle speed, and is suitable for application in micro and dynamic traffic environments. Since the fixed time headway is more similar to human driving habits, and the more complex definition of the expected vehicle spacing is not very helpful for improving the actual following effect, the embodiment of the present disclosure adopts the fixed time headway to define the expected vehicle spacing d des , the expression is as follows:

[0068] d des =d0+τ0v f

[0069] Where d0 is the safe stationary vehicle distance, representing the desired distance between the two vehicles when the ego vehicle's speed is zero. A smaller value indicates a closer distance to the preceding vehicle when the ego vehicle is stationary. τ0 is the time-to-distance constant, representing the multiplication factor of the desired distance as the ego vehicle's speed increases. A smaller value indicates a smaller change in the desired distance with increasing speed. Different parameter values ​​for d0 and τ0 can define different driving styles, such as conservative, aggressive, and general. Referring to existing models of human driving habits, this embodiment adopts the general driving style following parameters, selecting d0 and τ0 as 1.98m and 1.36s, respectively.

[0070] Optionally, in step S1, the general setting requirement of the sensor is to deploy speed, acceleration and energy consumption sensors on the vehicle. In addition, there are two typical solutions for obtaining signals from the vehicle in front. One is the vehicle-cloud collaborative solution, in which both the front and rear vehicles need to be equipped with speed, GPS sensors and vehicle-cloud signal interaction devices, and the position and speed information of the front vehicle are shared with the vehicle in front through the vehicle-cloud interaction mechanism. The second is the single-vehicle intelligent solution, in which the vehicle in front obtains the position and speed information of the vehicle in front with the help of millimeter-wave radar, camera, lidar and other devices. Under the framework of this embodiment, the sensor signal acquisition period of the two solutions needs to be T c =20ms.

[0071] For the sensor signal filtering algorithm installed on the test vehicle, common filtering algorithms such as mean filtering, median filtering, first-order filtering and Kalman filtering can be used to help reduce the measurement error caused by environmental noise to the sensor. You can choose according to actual needs. Taking mean filtering as an example, due to the acquisition period T of the sensor signal c Less than the calling cycle T of the vehicle following control algorithm s , T s / T c =5, so the average value of the first five sensor signals at the time of calling the vehicle following control algorithm can be used as the sensor signal input of the vehicle following control algorithm. In this embodiment, the period of the filtered data is T s , used as signal input for the following control algorithm and for accessing evaluation data. The initial speed of both vehicles is typically set to 0, making it easy to start the following test at any time. The initial distance between the two vehicles can be set between 2m and 20m. This is done to ensure that the initial distance is greater than the safe stationary distance d0 and to help the two vehicles establish a more stable vehicle-cloud collaboration relationship or single-vehicle sensor connection.

[0072] Optionally, in step S2, the autonomous driving mode of the leading vehicle needs to ensure that the leading vehicle maintains a consistent speed trajectory in multiple repeated experiments, and a fixed input method can be selected (i.e., inputting fixed throttle, brake, and steering wheel values ​​according to time); the autonomous driving mode of the own vehicle needs to be able to follow the leading vehicle stably, and any autonomous driving control algorithm can be selected under the premise of satisfying the input of the state quantities of the two vehicles and the output of the control quantity of the own vehicle.

[0073] After completing the sensor setup, signal filtering algorithm deployment, and vehicle initialization in step S1, c The original data of the test process is collected periodically and the filtered data is output as T s The evaluation data with and without sensor filtering are periodic, including the speed of the preceding vehicle, the speed, acceleration and energy consumption of the vehicle itself, and the distance between the two vehicles. The vehicle controller is used to record the distance every T sThe difference between the start and end times of the vehicle-following control algorithm call is used to calculate the single-step calculation duration of the vehicle-following control algorithm. The evaluation data and the single-step calculation duration of the vehicle-following control algorithm can be used to calculate subsequent evaluation indicators.

[0074] Optionally, in step S3, a first evaluation index CEI is constructed according to the following formula:

[0075]

[0076]

[0077] ω tra +ω com +ω eco =1 (3)

[0078]

[0079]

[0080] L tra =ω d Δd 2 +ω v Δv 2 (6)

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087] in,

[0088] L tra is the objective function used to evaluate the followability of the following control algorithm, which can reflect the following situation of the vehicle to the distance and speed of the preceding vehicle. Preferably, the inter-vehicle error Δd=dd is used in this embodiment. des and vehicle speed error Δv = v p -v f Quantitatively describe the followability objective function L tra , see formula (6), ω d and ω v are the weight coefficients of the vehicle distance error and vehicle speed error, respectively. Referring to the definition of tracking error indicators in relevant literature, distance tracking has ten times the weight of speed tracking, so we take ωd =10 / 11,ω v =1 / 11.

[0089] L com The objective function used to evaluate the comfort of the following control algorithm can reflect the comfort of the passengers in the following process of the ego vehicle. It is usually described by two variables: acceleration or jerk. Acceleration is generally used to quantitatively measure comfort. By analyzing its numerical distribution or square sum, the overall acceleration and deceleration of the vehicle in the following process is evaluated; jerk is generally used to qualitatively evaluate the comfort limit and measure the maximum acceleration change. In this embodiment, only the ego vehicle acceleration a is used. f A quantity, namely formula (9) to construct the comfort objective function L com On the one hand, since the following vehicle scenario is a single longitudinal movement of the vehicle, acceleration can accurately and quantitatively evaluate ride comfort, while jerk can only be used qualitatively as a judgment of acceleration change; on the other hand, since the following vehicle control scenario is usually relatively smooth, the overall vehicle jerk is small and will not exceed the human body's tolerance limit for jerk, so considering jerk constraints is not very meaningful.

[0090] L eco The objective function for evaluating the economic performance of the following control algorithm can reflect the energy consumption of the vehicle during the following process. Currently, the following control work only uses acceleration or a combination of acceleration and speed to approximate energy consumption. As an evaluation of economic performance, this method is obviously inaccurate for the approximation of vehicle energy consumption. The vehicle economic target should be considered as the overall energy consumption under a certain distance. Therefore, this embodiment uses the economic objective function L eco Designed as the vehicle energy consumption per unit distance described in formula (12), P is the instantaneous power consumption of the vehicle.

[0091] and They are the followability objective function, comfort objective function and economy objective function after standardization. It is preferred to use the Z-Score method to standardize each objective function so as to adjust each objective function to the same order of magnitude. It should be noted that for any evaluation index, the input of the objective function is a time-varying signal, so the standardized objective function The value of t is also different at each moment; μ tra and σ tra are the standardized mean and variance of the followability objective function, μ com and σ com are the standardized mean and variance of the comfort objective function, μ eco and σ ecoThey are the standardized mean and variance of the economic objective function respectively. The standardized mean and variance of each objective function are determined by the distribution estimated by the simulation experiment. The simulation experiment process is as follows: the leading vehicle drives according to the Chinese light vehicle driving condition CLTC, and the self-vehicle uses the PID algorithm to control the following. The sensor results at each moment are recorded and the corresponding followability, comfort, and economic objective functions are calculated. After collecting the complete simulation experiment data, the mean and variance of these three objective functions are calculated to obtain the parameter values ​​of the standardized mean and variance of each objective function. In this embodiment, μ is taken as tra =-0.10, μ com =-0.29, μ eco =-0.53, σ tra =0.50,σ com =0.27,σ eco =1.22.

[0092] L is a comprehensive objective function constructed based on the standardized followability objective function, comfort objective function and economy objective function. Specifically, the linear weighted method is used to organize the standardized objective functions into a comprehensive objective function. tra 、ω com and ω eco They are the weight coefficients of followability, comfort and economy respectively. The higher the weight coefficient, the more attention is paid to the performance.

[0093] CEI is the comprehensive objective function L used in the whole test with the vehicle following control algorithm calling cycle T s The first evaluation index is obtained by adding up .

[0094] TRA, COM and ECO are the followability index, comfort index and economy index of the following control algorithm respectively. The corresponding objective function can be used to call the following control algorithm during the whole test. s The cumulative sum of the values ​​is obtained separately. Due to standardized operations, the individual evaluation indicators TRA, COM, and ECO are essentially on the same order of magnitude. During use, each individual evaluation indicator can be observed separately to quantitatively analyze the specific performance of the vehicle following control algorithm in terms of the corresponding performance. It should be noted that the individual evaluation indicators in this embodiment only play a supporting role in performance evaluation. Typically, only the first evaluation indicator, CEI, is used to measure the control performance of the vehicle following control algorithm.

[0095] T s The smaller the value, the more frequently the control instructions are issued and the finer the control effect. On the basis of ensuring stable coupling with the sensor and electronic control system, T s It can be selected as any value between 10ms and 200ms. Preferably, in this embodiment, T s=100ms, which is 5 times the sensor signal acquisition period T c , which is conducive to the effective deployment of filtering algorithms.

[0096] Furthermore, the multi-objective weight coefficients are usually defined manually based on common sense judgment. However, in order to better quantify and objectify human subjectivity, this embodiment uses the hierarchical analysis method to determine the value of each weight coefficient in the comprehensive objective function L. The specific steps are as follows:

[0097] Step S311: Establish a hierarchical model. Since the following performance, comfort, and economy objectives involved in the vehicle-following control problem in this embodiment are at the same level (i.e., they all meet the requirements of the criterion layer and are not subordinate to each other, allowing subsequent steps to determine weight coefficients), analysis only needs to be performed at a single level. Specifically, a single-layer structural model is established with the following performance objective C1, comfort objective C2, and economy objective C3 at the criterion layer.

[0098] Step S312: Construct a judgment matrix. The judgment matrix quantifies each target using an importance scale (the importance scale is used to represent the quantitative value of the importance between two targets. The larger the quantitative value, the more important target i is than target j). This reflects the relative importance of targets at the same level.

[0099] Table 1 Importance scale table

[0100] Target i is better than target j Quantized value Equally important 1 Slightly important 3 Strong and important 5 Strongly important 7 Extremely important 9 Adjacent judgment middle value 2，4，6，8

[0101] Table 2 Judgment matrix of three goals

[0102] Followability Comfort Economical Followability 1 3 6 Comfort 1 / 3 1 2 Economical 1 / 6 1 / 2 1

[0103] According to the importance scale table in Table 1, construct the judgment matrix of the three objectives of followability, comfort, and economy. Use human experience to determine the quantitative value of the importance comparison between each two, and complete the upper right part of the judgment matrix. The element in the i-th row and j-th column of the judgment matrix is ​​represented as a ij , i, j∈[1,2,3], due to the duality relation, there is a ii =1, Determine the lower left remaining part of the judgment matrix. The final result of the judgment matrix is ​​shown in Table 2.

[0104] Step S313: Calculate the weight coefficients. When calculating the weight coefficients using the judgment matrix, first calculate the maximum eigenvalue and corresponding eigenvector of the judgment matrix, then normalize the eigenvector to obtain the initial weight coefficients of the three targets.

[0105] After calculation, the maximum eigenvalue of the above judgment matrix is ​​λ=3, and the normalized eigenvector is (2 / 3, 2 / 9, 1 / 9) T, the initial weight coefficients of the corresponding three objectives are ω tra0 =2 / 3,ω com0 =2 / 9,ω eco0 =1 / 9.

[0106] Step S314: consistency check. The purpose of the consistency check is to ensure that the various targets in the judgment matrix maintain logical consistency during the process of converting qualitative subjective judgments into quantitative parameters. If the judgment matrix fails the consistency check, it means that there is a logical conflict in the parameters in the judgment matrix and adjustment is required. The consistency index CI calculated in this process is:

[0107]

[0108] Wherein, n is the order of the judgment matrix, λ is the maximum eigenvalue of the judgment matrix. In this embodiment, n=3, λ=3, thereby obtaining CI=0.

[0109] The judgment matrix must satisfy the constraints to pass the consistency test:

[0110]

[0111] Where CR is the test coefficient and RI is the average random consistency index, which is determined by referring to the standard table of average random consistency index in Table 3 in the hierarchical analysis method. For the consistency test of three targets m = 3, RI = 0.58.

[0112] Table 3 Average random consistency index standard table

[0113] n 1 2 3 4 5 6 RI 0 0 0.58 0.90 1.12 1.24

[0114] Substituting the calculation result of step S313, we can obtain CI = 0, CR = 0 < 0.1. Therefore, the judgment matrix passes the consistency test, and the initial weight coefficients of the three targets obtained in step S313 are used as the final weight coefficients of each target.

[0115] Optionally, in step S3, the second evaluation index constructed for evaluating the application characteristics of the vehicle following control algorithm is composed of a real-time evaluation index and an anti-interference evaluation index, wherein:

[0116] Real-time performance refers to the time it takes for a vehicle-following control algorithm to produce a calculation result during its application. The average single-step calculation time h can be used as a real-time evaluation metric. The calculation time for each step is typically calculated by recording the start and end times of the vehicle-following control algorithm on the onboard controller and then subtracting them. The average single-step calculation time h is calculated by averaging the calculation times for each step. The shorter the average single-step calculation time h required by each vehicle-following control algorithm, the better the real-time performance of the vehicle-following control algorithm.

[0117] Interference immunity refers to the ability of a vehicle-following control algorithm to resist environmental interference. This can be assessed using the growth rate α of the first evaluation indicator, CEI, when no sensor filtering is used. In vehicle-following control problems, sensor error is the primary manifestation of environmental error and is affected by the use of sensor signal filtering. Therefore, this example sets up the group using sensor signal filtering as the control group, and the group without sensor signal filtering as the experimental group. With all other conditions remaining the same, the CEI indicators of the two vehicle-following control algorithm groups are tested throughout the entire test, and the CEI growth rate α is calculated:

[0118]

[0119] Among them, CEI NF Represents the CEI index of the entire test without sensor filtering, CEI F Represents the CEI for the entire test with sensor filtering. A larger α value indicates a greater increase in CEI without the sensor, a more pronounced deterioration in control effectiveness due to sensor error, and poorer anti-interference performance.

[0120] Optionally, in step S4, based on the evaluation data recorded in step S2 (including the speed of the leading vehicle and the speed, acceleration, energy consumption and single-step calculation time of the following control algorithm of the own vehicle), according to the formula defined in step S3, the first evaluation index CEI is calculated, the following evaluation index TRA, the comfort evaluation index COM and the economy evaluation index ECO are calculated, the real-time evaluation index (average single-step calculation time h) is calculated, and the anti-interference evaluation index (CEI index growth rate α when there is no sensor filtering) is calculated.

[0121] Optionally, in step S5, the real-time evaluation index requirement is defined as: the average single-step calculation time h of the controlled vehicle cannot exceed the fixed following control algorithm call period T s 70% of the real-time requirement. Once this value is exceeded, the vehicle following control algorithm is prone to significant delays, resulting in a decrease in control timeliness, which is considered to be a failure in the real-time requirement. For this embodiment, T s is 100ms, then h cannot exceed 70ms.

[0122] The anti-interference evaluation index requirement is defined as: when there is no sensor filtering algorithm, the growth rate α of the controlled vehicle's first evaluation index CEI cannot exceed 10%. Once this value is exceeded, the following control algorithm is considered to have a significant performance degradation when facing sensor interference, and in this case it is considered to have failed the anti-interference requirement. For example, when a control algorithm uses sensor filtering, the CEI index calculated during the entire test is CEI F =1000, then the CEI index calculated during the entire test for the same algorithm without sensor filtering is NFIt should not exceed 1100 (that is, the growth rate α of the CEI indicator should not exceed 10%).

[0123] From all the following control algorithms, the following control algorithms that meet both the real-time evaluation index and the anti-interference evaluation index are selected as candidate following control algorithms, and the candidate following control algorithm with the smallest first evaluation index CEI is selected as the optimal following control algorithm.

[0124] Verification of the effectiveness of the disclosed embodiment:

[0125] Under the same system, this analysis example implemented the following control function using four following control algorithms: a traditional control algorithm—Proportional Integral Derivative (PID) control, an optimal control algorithm—Model Predictive Control (MPC), and two reinforcement learning algorithms—Adaptive Dynamic Programming (ADP) and Deep Deterministic Policy Gradient (DDPG). In actual vehicle testing, this application demonstrated the versatility and effectiveness of the performance evaluation of various following control algorithms.

[0126] 1. Experimental Preparation

[0127] This analysis example uses two wire-controlled smart cars to conduct a real-car verification experiment of the car-following control algorithm. Figure 3 As shown in the figure, the BAIC EX5 serves as the leading vehicle, following a fixed driving pattern, while the BAIC EU260 serves as the autonomous vehicle. Both vehicles utilize a vehicle-cloud collaborative mechanism to share status information and dynamically respond to follow traffic. Both vehicles are equipped with speed and GPS sensors, as well as vehicle-cloud signal interaction devices, and utilize NXP I.MX6 series main control chips as onboard controllers.

[0128] The four following control algorithms are integrated into the complete underlying algorithm of the controlled vehicle EU260. By compiling and uploading, the following control algorithm can be deployed on the actual vehicle. The signal acquisition cycle of each sensor of EU260 is T c The four following control algorithms call cycles are 20ms and T s The time interval is 100ms, so there are five sensor signals between each call to the vehicle-following control algorithm. Therefore, to ensure the accuracy and stability of the sensor signals, the program performs a mean filter with a step size of 5 on the sensor signals. In other words, the average of the five sensor signals before the algorithm call is calculated and used as the sensor signal input for the vehicle-following control algorithm.

[0129] In order to meet the basic requirements of the vehicle-following test, this analysis example selected Boxing Eighth Road in the Beijing Yizhuang Autonomous Driving Demonstration Zone as the test section. Figure 4 This section of road is 750 meters long and passes through one intersection. It has no traffic lights, is straight, has low traffic flow, and is in good condition, making it suitable for testing the vehicle-following control algorithm. At the beginning of each experiment, the two vehicles are parked 2.5 meters apart at the starting position of the road, waiting for the test to begin.

[0130] 2. Evaluation Data Collection

[0131] To verify the performance differences between various following control algorithms, it's necessary to operate under identical operating conditions, meaning the leading vehicle's driving state must remain consistent. Therefore, this analysis example designed a fixed throttle / brake command in the leading vehicle controller to ensure stable longitudinal motion. Furthermore, a fixed steering wheel command was designed to maintain the leading vehicle's straight trajectory and limit lateral drift.

[0132] The four following control algorithms, PID, MPC, ADP, and DDPG, were deployed on the ego vehicle in sequence, and tests were completed under two conditions: with and without sensor filtering. The two vehicles were controlled to start at the same time. The leading vehicle operated the throttle / brake and steering wheel commands to maintain a straight line, while the trailing vehicle entered the following driving mode. The T-values ​​generated by the two vehicles in the above 4×2=8 groups of experiments were recorded. s The evaluation data is periodic, including the speed of the leading vehicle and the speed, acceleration, energy consumption of the own vehicle, and the single-step calculation time of the following control algorithm.

[0133] Under the condition of sensor filtering, the speed of the leading vehicle in the verification experiment of the four groups of vehicle following control algorithms is as follows: Figure 5 The total length of this actual vehicle test condition is 633m, the duration is 83s, the maximum speed of the leading vehicle is 14.6m / s, and it starts and ends at a standstill. Figure 5 The black thin line in the middle is the mean speed of the front vehicle in each group of experiments, and the shaded area is the error range of the front vehicle speed. Figure 5 The figure shows that the speed of the leading vehicle in the four groups of experiments is basically the same, which can be used to compare the actual performance of each following control algorithm.

[0134] 3. Calculation of control effect evaluation indicators

[0135] The first evaluation index CEI of the control effect of the four following control algorithms, PID, MPC, ADP and DDPG, is calculated under the condition of sensor filtering. The results are shown in Table 3.

[0136] Table 4 The first evaluation index of the control effect of each vehicle following control algorithm

[0137] algorithm CEI TRA COM ECO MPC 1321.4 1863.0 345.4 23.7 ADP 1430.7 2067.6 228.4 13.7 DDPG 1879.3 2750.9 196.9 14.5 PID 2169.9 3125.2 373.0 32.0

[0138] Judging from the first evaluation index CEI, the performance ranking of each vehicle-following control algorithm is MPC>ADP>DDPG>PID, which is consistent with the academic community's judgment on the control effects of the optimal control algorithm and the reinforcement learning algorithm. Looking closely at the three separate evaluation indicators TRA, COM, and ECO, it is found that MPC, ADP, and DDPG each have their own expertise in followability, economy, and comfort, which also happens to reflect the differences in vehicle-following control among various vehicle-following control algorithms. As a traditional algorithm, PID is often used as a baseline for algorithm performance comparison. Its control effect is significantly worse than that of the other three vehicle-following control algorithms, which is also in line with the expectations of this application.

[0139] 4. Calculation of Application Characteristics Evaluation Index

[0140] (1) Real-time

[0141] The average single-step calculation time required for the operation and processing of the above four vehicle-following control algorithms under the condition of sensor filtering is calculated, and the results are shown in Table 4.

[0142] Table 5 Real-time performance indicators of each algorithm application characteristics

[0143]

[0144]

[0145] The time consumption ranking of the following control algorithms in Table 4 is: MPC > ADP = DDPG > PID. MPC requires a longer average single-step calculation time. ADP and DDPG have identical application structures, so their calculation times are essentially the same, and PID still takes the least time.

[0146] (2) Anti-interference

[0147] The CEI indicators of the four car-following control algorithms with and without sensor filtering are calculated respectively, and the growth rate α of the CEI indicator is calculated. The results are shown in Table 5.

[0148] Table 6 Anti-interference index of each algorithm application characteristics

[0149]

[0150] The results in Table 5 demonstrate that applying mean filtering to sensor signals can indeed suppress noise and improve the performance of the following control algorithm. Furthermore, the CEI growth rate α varies across the following control algorithms for varying levels of sensor noise: MPC shows the smallest increase, followed by ADP, DDPG, and PID. This indicates that in real-vehicle testing, the interference resistance ranking of the following control algorithms follows the order: MPC > ADP > DDPG > PID. This ranking provides valuable insights into algorithm application.

[0151] 5. Algorithm Screening

[0152] Summarizing the above calculation process, we can produce a performance evaluation table for the four vehicle-following control algorithms shown in Table 6. The ranking in this table collectively reflects the application results of the evaluation method of this embodiment.

[0153] Table 7 Performance evaluation of four vehicle following control algorithms in this embodiment

[0154]

[0155] First, according to the real-time evaluation index requirements, the average single-step calculation time h of the controlled vehicle cannot exceed the fixed following control algorithm call cycle T s For this example, T s The average single-step duration h cannot exceed 70ms. Therefore, the MPC algorithm does not meet the real-time requirements of this example.

[0156] Secondly, according to the anti-interference evaluation index requirements, when using the sensorless filtering algorithm, the growth rate α of the controlled vehicle's first evaluation index CEI cannot exceed 10%. Inspection shows that the remaining three vehicle-following control algorithms all meet the anti-interference requirements of this example.

[0157] Finally, the candidate following control algorithm with the smallest first evaluation index CEI is selected as the optimal following control algorithm. Therefore, the ADP algorithm is selected as the optimal following control algorithm among the remaining three following control algorithms.

[0158] A performance evaluation device for a vehicle following control algorithm provided by an embodiment of a second aspect of the present disclosure includes:

[0159] The first module is configured to obtain evaluation data in the following manner: deploy several vehicle-following control algorithms on test vehicles equipped with the same sensors and sensor signal filtering algorithms, set the sensor signal acquisition period T c And the following control algorithm call cycle T s , T c <T s In the preset vehicle following test scenario, each test vehicle is controlled to reach the set initial speed and initial vehicle distance, and each test vehicle is controlled to enter the automatic driving mode. c Collect the original data during the test and output them in T s The vehicle controller records the single-step calculation duration of each call to the vehicle-following control algorithm for periodic evaluation data with and without sensor filtering. The evaluation data includes the speed of the leading vehicle, the speed, acceleration, and energy consumption of the test vehicle, and the distance between the two vehicles.

[0160] An evaluation index calculation module is configured to calculate a first evaluation index and a second evaluation index for each vehicle following control algorithm based on evaluation data from the entire vehicle following test process; the first evaluation index is used to evaluate the control effect of the vehicle following control algorithm, and is a comprehensive index obtained by summing the followability index, comfort index, and economy index of the vehicle following control algorithm obtained based on the evaluation data when sensor filtering is used, with weight coefficients determined by the hierarchical analysis method; the second evaluation index is used to evaluate the application characteristics of the vehicle following control algorithm, the real-time evaluation index is calculated based on the single-step calculation time of calling the vehicle following control algorithm, and the anti-interference evaluation index is obtained based on the growth rate of the first evaluation index when sensor-free filtering is used;

[0161] The evaluation module is used to screen out the vehicle following control algorithms that meet the requirements of real-time evaluation index and anti-interference evaluation index from all vehicle following control algorithms as candidate vehicle following control algorithms, and select the candidate vehicle following control algorithm with the smallest first evaluation index as the optimal vehicle following control algorithm.

[0162] In order to implement the above embodiment, the embodiment of the present disclosure further proposes a computer-readable storage medium on which a computer program is stored. The program is executed by a processor to execute the performance evaluation method of the vehicle following control algorithm of the above embodiment.

[0163] Reference below Figure 6 , which shows a schematic structural diagram of an electronic device suitable for implementing the embodiments of the present disclosure. It should be noted that the electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs, desktop computers, and servers. Figure 6 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0164] like Figure 6 As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 101, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 102 or a program loaded from a storage device 108 into a random access memory (RAM) 103. Various programs and data required for the operation of the electronic device are also stored in the RAM 103. The processing device 101, the ROM 102, and the RAM 103 are connected to each other via a bus 104. An input / output (I / O) interface 105 is also connected to the bus 104.

[0165] Typically, the following devices may be connected to the I / O interface 105: an input device 106 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, etc.; an output device 107 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 108 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 109. The communication device 109 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Figure 6 The electronic device is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0166] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, this embodiment includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication device 109, or installed from the storage device 108, or installed from the ROM 102. When the computer program is executed by the processing device 101, the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.

[0167] It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component. In the present disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0168] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0169] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device executes the performance evaluation method of the vehicle following control algorithm.

[0170] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, Python, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0171] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0172] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. Throughout the description of this application, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0173] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0174] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection having one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or otherwise processing it in a suitable manner if necessary, and then storing it in a computer memory.

[0175] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0176] Those skilled in the art will understand that all or part of the steps carried out in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the developed program can be stored in a computer-readable storage medium. When the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0177] In addition, the functional units in the various embodiments of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into a module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0178] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A performance evaluation method for a vehicle following control algorithm, characterized in that: include: Several vehicle-following control algorithms are deployed on test vehicles equipped with the same sensors and sensor signal filtering algorithms, and the sensor signal acquisition period T is set. c And the following control algorithm call cycle T s , T c <T s , in the preset vehicle following test scenario, control each test vehicle to reach the set initial speed and initial vehicle distance; Control each test vehicle to enter the automatic driving mode, every T c Collect the original data during the test and output them in T s The vehicle controller records the single-step calculation duration of each call to the vehicle-following control algorithm for evaluation data with and without sensor filtering. The evaluation data includes the speed of the leading vehicle, the speed, acceleration, and energy consumption of the test vehicle, and the distance between the two vehicles. A first evaluation index for evaluating the control effect of the vehicle-following control algorithm and a second evaluation index for evaluating the application characteristics of the vehicle-following control algorithm are respectively constructed. The first evaluation index is a comprehensive index obtained by summing the following index, comfort index, and economy index of the vehicle-following control algorithm obtained based on evaluation data when sensor filtering is used, and assigning weight coefficients determined by the hierarchical analysis method. The second evaluation index is composed of a real-time evaluation index and an anti-interference evaluation index. The real-time evaluation index is calculated based on the single-step calculation time of calling the vehicle-following control algorithm, and the anti-interference evaluation index is obtained based on the growth rate of the first evaluation index when sensorless filtering is used. Calculating the first evaluation index and the second evaluation index of each vehicle following control algorithm based on the evaluation data of the entire vehicle following test process; From all the following control algorithms, the following control algorithms that meet both the real-time evaluation index and the anti-interference evaluation index are selected as candidate following control algorithms, and the candidate following control algorithm with the smallest first evaluation index is selected as the optimal following control algorithm.

2. The performance evaluation method according to claim 1, wherein: The first evaluation index is constructed according to the following formula: oh tra +oh com +oh eco =1 Among them, CEI is the first evaluation index; TRA, COM and ECO are the followability index, comfort index and economy index of the vehicle following control algorithm respectively; ω tra 、ω com and ω eco are the weight coefficients of followability index, comfort index and economy index respectively; and They are the followability objective function, comfort objective function and economy objective function after standardization respectively.

3. The performance evaluation method according to claim 2, wherein: The standardization method used for each objective function is the Z-Score method.

4. The performance evaluation method according to claim 2, wherein The followability objective function before normalization is L tra , whose expression is: L tra =ω d Δd 2 +oh v Δv 2 Wherein, Δd is the inter-vehicle distance error, which is defined as the difference between the actual inter-vehicle distance between the test vehicle and the preceding vehicle and the desired inter-vehicle distance, where the desired inter-vehicle distance uses a fixed inter-vehicle time interval; Δv is the speed error, which is defined as the difference in speed between the test vehicle and the preceding vehicle; ω d and ω v are weight coefficients for vehicle spacing error and vehicle speed error respectively; and / or Assume that the comfort objective function before standardization is L com , whose expression is: Among them, a f To test the acceleration of the vehicle; and / or Assume that the economic objective function before standardization is L eco , whose expression is: Where, P is the instantaneous power consumption of the test vehicle, v f To test the vehicle's speed.

5. The performance evaluation method according to claim 2, wherein: The weight coefficient ω of the followability index, comfort index and economy index in the first evaluation index CEI is determined by the hierarchical analysis method tra 、ω com and ω eco The steps include: Step S311: Establishing a hierarchical structure model, wherein the hierarchical structure model is a single-layer structure model having a followability target, a comfort target, and an economy target in a criterion layer; Step S312: construct a judgment matrix for quantifying the relative importance of different targets through the importance scale table. Let the element in the i-th row and j-th column of the judgment matrix be a ij , represents the relative importance of target i to target j, i, j∈[1, 2, 3], and satisfies: a ii =1, S313, calculating the maximum eigenvalue λ and the corresponding eigenvector of the judgment matrix, normalizing the eigenvector, and obtaining initial weight coefficients of the followability target, the comfort target, and the economy target; S314. Perform a consistency check on the discriminant matrix. If the consistency check passes, the initial weight coefficients of the followability target, comfort target, and economy target obtained in step S313 are used as the final weight coefficients of each target. If the consistency check fails, return to step S312 to reconstruct the discriminant matrix.

6. The performance evaluation method according to claim 1, wherein: The average value of the single-step calculation time of each call of the vehicle-following control algorithm during the vehicle-following test is used as the real-time evaluation index.

7. The performance evaluation method according to claim 1, wherein: The anti-interference evaluation index is calculated according to the following formula: Wherein, α is the anti-interference evaluation index, that is, the growth rate of the first evaluation index when there is no sensor filtering; CEI F CEI is the statistical value of the first evaluation index of the whole process of the vehicle following test; NF It is the statistical value of the third evaluation index of the entire vehicle following test process. The third evaluation index is a comprehensive index obtained by summing up the followability index, comfort index and economy index of the vehicle following control algorithm obtained based on the evaluation data during sensorless filtering, and assigning weight coefficients determined by the hierarchical analysis method.

8. The performance evaluation method according to claim 1, wherein: The real-time evaluation index is required to be no more than T s 70%, requiring that the anti-interference evaluation index does not exceed 10%.

9. A performance evaluation device for a vehicle following control algorithm, characterized in that: include: The evaluation data acquisition module is configured to acquire evaluation data in the following manner: deploying several vehicle following control algorithms on test vehicles equipped with the same sensors and sensor signal filtering algorithms, setting the sensor signal acquisition period T c And the following control algorithm call cycle T s , T c <T s In the preset vehicle following test scenario, each test vehicle is controlled to reach the set initial speed and initial vehicle distance, and each test vehicle is controlled to enter the automatic driving mode. c Collect the original data during the test and output them in T s The vehicle controller records the single-step calculation duration of each call to the vehicle-following control algorithm for periodic evaluation data with and without sensor filtering. The evaluation data includes the speed of the leading vehicle, the speed, acceleration, and energy consumption of the test vehicle, and the distance between the two vehicles. An evaluation index calculation module is configured to calculate a first evaluation index and a second evaluation index for each vehicle following control algorithm based on evaluation data from the entire vehicle following test process; the first evaluation index is used to evaluate the control effect of the vehicle following control algorithm, and is a comprehensive index obtained by summing the followability index, comfort index, and economy index of the vehicle following control algorithm obtained based on the evaluation data when sensor filtering is used, and assigning weight coefficients determined by the hierarchical analysis method; the second evaluation index is used to evaluate the application characteristics of the vehicle following control algorithm, and is composed of a real-time evaluation index and an anti-interference evaluation index, the real-time evaluation index being calculated based on the single-step calculation time of calling the vehicle following control algorithm, and the anti-interference evaluation index being obtained based on the growth rate of the first evaluation index when sensor-free filtering is used; The evaluation module is configured to screen out the vehicle following control algorithms that meet the requirements of real-time evaluation index and anti-interference evaluation index from all vehicle following control algorithms as candidate vehicle following control algorithms, and select the candidate vehicle following control algorithm with the smallest first evaluation index as the optimal vehicle following control algorithm.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the performance evaluation method of the vehicle following control algorithm according to any one of claims 1 to 8.

Images