Electric vehicle bench test method and system based on actual traffic scene

Abstract

The invention relates to the technical field of vehicle test analysis, in particular to an electric vehicle bench test method and system based on an actual traffic scene, and the method comprises the steps: obtaining the map information and driving scene data of a target path; constructing a road scene of the target path according to the map information, constructing a traffic flow scene of the target path according to the map information and the driving scene data, and generating an actual traffic scene of the target path according to the road scene and the traffic flow scene; and an electric drive system rack of the electric vehicle is built, the electric drive system rack is controlled to carry out a simulated driving test on the target path in the actual traffic scene, data of the simulated driving test are collected, and a test result of the electric vehicle is analyzed according to the data of the simulated driving test. Therefore, the problems that the testing result is unstable, potential safety hazards exist, the testing repeatability is poor due to the fact that real vehicle road testing is adopted in the related technology, but the efficiency of the electric drive system under the actual traffic scene is difficult to reflect truly through the standard cycle working condition in rack testing are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle testing and analysis, and in particular to an electric vehicle bench testing method and system based on actual traffic scenarios. Background Art

[0002] In the wave of the new four modernizations of electric vehicles, as the refinement of vehicle energy management continues to increase, the efficiency test of the electric drive system has become an important direction of energy management. However, the current efficiency test methods of electric drive systems mainly rely on standard cycle conditions or custom conditions as input conditions for electric vehicle testing, such as CLTC-P (China Light Vehicle Driving Cycle) and WLTC (World Light Vehicle Test Cycle).

[0003] In the related art, the use of standard cycle conditions testing has the problem of the driver's intention to follow the speed-time curve, which makes it difficult to truly reflect the efficiency of the electric drive system under actual traffic scenarios; the use of actual vehicle road testing methods has the problem of difficulty in fixing the layout of current and voltage sensors and difficulty in arranging wheel-side speed and torque sensors. In addition, there are safety hazards due to the layout of vehicle sensors, and the actual traffic scenarios change instantly, and the test repeatability is poor. Summary of the invention

[0004] The present invention provides an electric vehicle bench test method and system based on actual traffic scenarios, so as to solve the problems that the related technology adopts real vehicle road testing, resulting in unstable test results and safety hazards, and the test repeatability is poor, but it is difficult to truly reflect the efficiency of the electric drive system in actual traffic scenarios using standard cycle conditions in bench testing.

[0005] The first aspect of the present invention provides an electric vehicle bench testing method based on actual traffic scenarios, comprising the following steps: obtaining map information and driving scenario data of a target path; constructing a road scene of the target path according to the map information, constructing a traffic flow scene of the target path according to the map information and the driving scenario data, and generating an actual traffic scene of the target path according to the road scene and the traffic flow scene; building an electric drive system bench for an electric vehicle, controlling the electric drive system bench to perform a simulated driving test on the target path under the actual traffic scenario, collecting data from the simulated driving test, and analyzing the test results of the electric vehicle based on the data from the simulated driving test.

[0006] Optionally, constructing the road scene of the target path according to the map information includes: acquiring key road data of the target path in the map information; and constructing the road scene of the target path according to the key road data using a step-by-step incremental method.

[0007] Optionally, constructing a road scene of the target path according to the key road data using a segment-by-segment incremental method includes: identifying the segment distance, road type, weather conditions, angle of entry into the road and traffic light position of each segment road in the key road data; querying a preset table according to the road type to determine the corresponding number of lanes, road width and speed limit signs; determining the friction coefficient of each segment road according to the road type and weather conditions; and constructing a road scene according to the number of lanes, the road width, the speed limit sign, the friction coefficient, the segment distance, the angle of entry into the road and the traffic light position.

[0008] Optionally, generating the actual traffic scene of the target path based on the road scene and the traffic flow scene includes: inputting the road scene and the traffic flow scene into a target neural network model, and the target neural network model outputs the traffic flow scene of the target path.

[0009] Optionally, the structure of the target neural network model includes an input layer, an output layer and a hidden layer, wherein the number of input layer nodes is the number of vehicle parameters, the number of output layer nodes is the number of fault indicators, and the number of hidden layer nodes is the target number of nodes, wherein the target number of nodes is obtained using the minimum training number method.

[0010] Optionally, the training method of the target neural network model includes: obtaining map information and key factors of traffic flow scenes to generate a training data set; wherein, the first excitation function is used as the excitation function from the input layer to the hidden layer, and the second excitation function is used as the excitation function from the hidden layer to the output layer; and the neural network model is trained using the training data set until the neural network model converges to the training target.

[0011] Optionally, before the acquisition of key factors of map information and traffic flow scenarios to generate a training data set, it includes: identifying historical map information and driving scene data of a target path; constructing an initial matrix based on the historical map information and driving scene data, and performing dimensionless processing on the initial matrix to generate a target matrix; calculating the difference between any two elements of the target matrix to generate a difference matrix, and identifying the maximum and minimum values ​​of the difference matrix; calculating the correlation coefficient between each factor and traffic volume based on the maximum and minimum values, and calculating the grey correlation between each factor and traffic volume based on the correlation coefficient; selecting factors with a grey correlation greater than a preset threshold as key factors of map information and traffic flow scenarios.

[0012] Optionally, after the target neural network model outputs the traffic flow of the target path, it includes: identifying the traffic volume of adjacent sections in the target path; if the traffic volume difference is greater than a preset threshold, initiating a cut-in and cut-out mechanism.

[0013] Optionally, the construction of an electric drive system bench for an electric vehicle includes: charging the vehicle to be tested to a target power level; removing the tires of the vehicle to be tested; mounting the vehicle to be tested on an axle-coupled dynamometer via a flange, mounting the vehicle's steering tie rod to the vehicle to be tested, and initializing the axle-coupled dynamometer; and arranging sensors and data acquisition equipment to synchronously collect vehicle and dynamometer data.

[0014] The second aspect of the present invention provides a bench test system for electric vehicles based on actual traffic scenarios. The system is applied to the bench test method for electric vehicles based on actual traffic scenarios described in the above embodiments. The system includes: a data acquisition module for synchronously acquiring current and voltage signals, vehicle CAN signals and shaft-coupled dynamometer signals of the vehicle to be tested; a network module for acquiring map information and driving scene data of the target path; an actual traffic scene construction module communicatively connected to the network module, for constructing an actual traffic scene according to the map information and driving scene data of the target path; a vehicle simulation model communicatively connected to the vehicle to be tested; and a shaft-coupled dynamometer for vehicle efficiency testing to simulate the target of the vehicle to be tested. data, and feeds back the target data to the shaft-coupled dynamometer control module; the shaft-coupled dynamometer control module is used to control the shaft-coupled dynamometer; the ring screen is used to display the actual traffic scene to the vehicle to be tested; the environmental chamber is used to simulate the actual ambient temperature to build test conditions; the host computer is used to run the network module to obtain the map information and driving scene data of the target path, input the map information of the target path and the driving scene data into the actual traffic scene construction module to construct the actual traffic scene, and display the actual traffic scene on the vehicle to be tested; run the shaft-coupled dynamometer control module to control the operation of the shaft-coupled dynamometer, input the data collected by the shaft-coupled dynamometer into the simulation model, and run the simulation model to simulate and obtain the test results.

[0015] Therefore, the present invention has at least the following beneficial effects: The embodiments of the present invention can construct a road scene of a target path according to map information, construct a traffic flow scene of the target path according to map information and driving scene data, generate an actual traffic scene of the target path according to the road scene and the traffic flow scene, build an electric drive system test bench for an electric vehicle, control the electric drive system test bench to perform a simulated driving test on the target path under an actual traffic scene, collect data from the simulated driving test, analyze the test results of the electric vehicle according to the data from the simulated driving test, construct an actual traffic scene by integrating online map information, and implement a test of the vehicle's electric drive efficiency on a shaft-coupled dynamometer, which can not only improve the accuracy of the test, but also ensure the safety, stability, consistency and repeatability of the test process, thereby reducing the overall cost.

[0016] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A flowchart of a bench test method for an electric vehicle based on an actual traffic scenario provided according to an embodiment of the present invention; Figure 2 A schematic diagram of building a road scene according to an embodiment of the present invention; Figure 3 A schematic diagram of an actual traffic scene automatic construction method provided according to an embodiment of the present invention; Figure 4 A schematic diagram of a bench test method for the efficiency of an electric vehicle electric drive system based on an actual traffic scenario according to an embodiment of the present invention; Figure 5 This is an example diagram of an electric vehicle bench test system based on actual traffic scenarios provided according to an embodiment of the present invention.

[0018] Explanation of the reference numerals: bench test system 10 for electric vehicles based on actual traffic scenarios, data acquisition module 100, networking module 200, actual traffic scenario construction module 300, vehicle simulation model 400, shaft-coupled dynamometer 500, shaft-coupled dynamometer control module 600, ring curtain 700, environmental chamber 800 and host computer 900. DETAILED DESCRIPTION

[0019] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0020] The following describes an electric vehicle bench testing method and system based on actual traffic scenarios according to an embodiment of the present invention with reference to the accompanying drawings.

[0021] Specifically, Figure 1 A schematic flow chart of a method for testing an electric vehicle bench based on actual traffic scenarios provided in an embodiment of the present invention.

[0022] like Figure 1 As shown, the electric vehicle bench test method based on actual traffic scenarios includes the following steps: In step S101, map information and driving scene data of the target path are obtained.

[0023] Among them, the target path can be selected according to needs without specific limitation.

[0024] It is understandable that the embodiments of the present invention can obtain map information and driving scene data of the target path to facilitate the subsequent construction of the actual traffic scene of the target path.

[0025] Specifically, the network connection module is used to obtain online map information and obtain key information of the test path through the online map API interface, including maneuver turning points, road section distance, road section time, segment distance, angle of entry into the road for each segment, segment time, segment road type, segment congestion level, number of traffic lights, traffic light location, weather, temperature and other information; the angle of entry into the road is divided into 12 levels, including 0-11, "0" represents 345 degrees to 15 degrees, and so on, 11" represents 315 degrees to 345 degrees, where the angle is the difference between the angle of entry into the road and the positive Clockwise angle from north; specifically, 0-11 represent the following: 0-[345°-15°]; 1-[15°-45°]; 2-[45°-75°]; 3-[75°-105°]; 4-[105°-135°]; 5-[135°-165°]; 6-[165°-195°]; 7-[195°-225°]; 8-[225°-255°]; 9-[255°-285°]; 10-[285°-315°]; 11-[315°-345°]. The maneuvering turning points include eight benchmark directions, 1-8, including 1-straight, 2-right front turn, 3-right turn, 4-right rear turn, 5-U-turn, 6-left rear turn, 7-left turn, and 8-left front turn; the segmented road types include 10 types, 0-highway, 1-city highway, 2-national highway, 3-provincial highway, 4-county highway, 5-township and village road, 6-other roads, 7-level 9 road, 8-airline (ferry), and 9-pedestrian road. The segmented congestion levels include four types, 1-smooth, 2-slow, 3-congested, and 4-severely congested.

[0026] In step S102, a road scene of the target path is constructed according to the map information, a traffic flow scene of the target path is constructed according to the map information and the driving scene data, and an actual traffic scene of the target path is generated according to the road scene and the traffic flow scene.

[0027] It can be understood that the embodiments of the present invention can construct a road scene of the target path based on map information, construct a traffic flow scene of the target path based on map information and driving scene data, and generate an actual traffic scene of the target path based on the road scene and the traffic flow scene, so as to simulate the vehicle driving on the actual road as realistically as possible in the subsequent testing process, thereby improving the accuracy and reliability of the efficiency test of the electric drive system of the electric vehicle.

[0028] Specifically, the actual traffic scenario construction module is used to construct actual traffic scenarios, including road scenario construction and traffic flow construction. It automatically constructs road scenarios based on the road segment distance, the angle of entry into the road for each segment, the maneuvering turning point, segment distance, segment road type, and traffic light position in the online map working condition information. It corrects the rolling resistance coefficient of the road in real time based on information such as road type, weather, and temperature, making the constructed road scene more real-time. It is constructed based on historical traffic information and the maneuvering turning point, segment congestion level, traffic light position, weather, and other information provided by the online map.

[0029] In an embodiment of the present invention, constructing a road scene of a target path according to map information includes: acquiring key road data of the target path in the map information; and constructing a road scene of the target path according to the key road data using a step-by-step incremental method.

[0030] It can be understood that the embodiments of the present invention can construct the road scene of the target path by using the segment-by-segment incremental method according to the key road data of the target path in the map information. The segment-by-segment incremental method can accurately simulate the specific characteristics of each section of the road, such as road type, number of lanes, width, speed limit sign, friction coefficient, etc., to ensure that the physical properties of each road section can be accurately reflected; based on the real-time information provided by the online map (such as weather, temperature, traffic light location, etc.), each road section can be dynamically corrected, so that the constructed road scene is closer to the actual situation, thereby improving the accuracy of subsequent tests.

[0031] In an embodiment of the present invention, a road scene of a target path is constructed according to key road data using a step-by-step incremental method, including: identifying the segment distance, road type, weather conditions, angle of entry into the road, and traffic light position of each segment road in the key road data; querying a preset table according to the road type to determine the corresponding number of lanes, road width, and speed limit signs; determining the friction coefficient of each segment road according to the road type and weather conditions; and constructing a road scene according to the number of lanes, road width, speed limit sign, friction coefficient, segment distance, angle of entry into the road, and traffic light position.

[0032] Among them, the preset table can be set according to actual conditions, as shown in Table 1 below, without specific limitation.

[0033] It can be understood that the embodiments of the present invention can identify the segmented distance, road type, weather conditions, angle of entry into the road and traffic light position of each segmented road in the key road data; determine the corresponding number of lanes, road width and speed limit signs by querying the preset table according to the road type; determine the friction coefficient of each segmented road according to the road type and weather conditions; construct a road scene according to the number of lanes, road width, speed limit signs, friction coefficient, segmented distance, angle of entry into the road and traffic light position, so as to accurately simulate the specific characteristics of each section of the road, make the constructed road scene closer to the actual situation, and improve the accuracy of subsequent tests.

[0034] Specifically, if Figure 2 As shown in the figure, the road scene is constructed by a segment-by-segment incremental method, using the online map to feedback the segment distance, segment road type, angle of entry into the road for each segment, and the position of the traffic light. In the scene construction module, for each segment road, the road type is defined, the number of lanes, road width, speed limit sign, and friction coefficient are set. The number of lanes, road width, and speed limit sign are set using the actual collected driving scene database and statistically obtained. For different types of roads, the table lookup index is extracted and constructed as shown in Table 1 below. By entering the road type, the number of lanes, road width, and speed limit sign can be obtained.

[0035] Table 1 Road type correspondence table

[0036] For example, if the road type corresponding to the target path of the present invention is a highway, the corresponding number of lanes is LN1, the road width is LW1, and the speed limit mark is Vlimt1. The final result can be obtained by querying Table 1 according to specific needs.

[0037] In addition, the friction coefficients of different roads of the present invention are obtained through the following Table 2, and the road type and weather are input to obtain the friction coefficients of each segmented road.

[0038] Table 2 Friction coefficient correspondence table

[0039] For example, if the road type corresponding to the target path of the present invention is a highway, and today's weather condition is sunny, then the friction coefficient of the corresponding road is Lf11.

[0040] Then, the angle of each section entering the road and the length of the section are increased step by step, and finally the position of the traffic lights is arranged to finally form a planned road scene.

[0041] In an embodiment of the present invention, an actual traffic scene of a target path is generated based on a road scene and a traffic flow scene, including: inputting the road scene and the traffic flow scene into a target neural network model, and the target neural network model outputs the traffic flow scene of the target path.

[0042] It can be understood that the embodiments of the present invention can input road scenes and traffic flow scenes into the target neural network model, and the target neural network model outputs the traffic flow scene of the target path. The target neural network can quickly adapt to changes in road conditions and traffic conditions, thereby timely adjusting the traffic flow prediction results to ensure that the test scene is always close to reality, so that the efficiency test of the electric vehicle's electric drive system can be carried out under conditions close to reality, while ensuring that the test results are highly representative and repeatable.

[0043] Specifically, if Figure 3 As shown, the present invention can use historical data such as road type, segment distance, segment congestion, segment time, weather, motor vehicle traffic volume, non-motor vehicle traffic volume, pedestrian traffic volume, motor vehicle speed, etc. to train a neural network model.

[0044] The model input data may include maneuver turning points, road section distance, road section time, segment distance, angle of entry into the road for each segment, segment time, segment road type, segment traffic congestion level, number of traffic lights, traffic light location, weather and temperature, etc.; the model outputs traffic flow scenarios such as motor vehicle traffic volume, non-motor vehicle traffic volume, pedestrian traffic volume, and motor vehicle speed.

[0045] In an embodiment of the present invention, the structure of the target neural network model includes an input layer, an output layer and a hidden layer, wherein the number of input layer nodes is the number of vehicle parameters, the number of output layer nodes is the number of fault indicators, and the number of hidden layer nodes is the target number of nodes, wherein the target number of nodes is obtained using the minimum training number method.

[0046] Specifically, the number of input layer nodes of the neural network model is set to N, where N is the number of vehicle parameters; the number of output nodes is set to P, where P is the number of fault indicators; the number of hidden layers is set to N1, where N1 is 1 or 2. If the number of layers is too large, overfitting will occur, affecting the model's running speed and accuracy. The number of nodes in each layer is N2, and the number of N2 is obtained using the minimum training number method: Specifically, we first determine a basic range through empirical formulas, using the first empirical formula: N2= (N+P) 1 / 2 +b; where b is a natural number between [1,10], we get an empirical value n21 for the number of nodes; then according to the second empirical formula: N2=2N+1; we get an empirical value n22 for the number of nodes, and thus we conclude that N2 is a natural number between [n21,n22].

[0047] Then, the golden section method is used to select the value n23 of the number of nodes, and the number of training times when the neural network model with node numbers n21, n22, and n23 is stable is calculated. The settlement quantities corresponding to the two smallest times are selected, and then the golden section method is used to continue to take values ​​between the corresponding nodes, and the process is repeated until a unique value is selected, which is the number of hidden layer nodes. Selecting the appropriate number of nodes in this way can quickly realize result prediction and improve efficiency.

[0048] In an embodiment of the present invention, a method for training a target neural network model includes: obtaining map information and key factors of a traffic flow scene to generate a training data set; wherein a first excitation function is used as an excitation function from an input layer to a hidden layer, and a second excitation function is used as an excitation function from a hidden layer to an output layer; and the neural network model is trained using the training data set until the neural network model converges to a training target.

[0049] Among them, the first excitation function is the tangent S-shaped TansIg function, and the second excitation function is the PureIn function, which can be selected according to actual needs without specific limitation.

[0050] It can be understood that the embodiments of the present invention can obtain map information and key factors of traffic flow scenarios to generate a training data set; wherein the first excitation function is used as the excitation function from the input layer to the hidden layer, and the second excitation function is used as the excitation function from the hidden layer to the output layer; the neural network model is trained using the training data set until the neural network model converges to the training target to improve the accuracy of the model prediction.

[0051] Specifically, the calculation numbers of the neural network model are configured as follows: use MATLAB to establish a three-layer or four-layer BP neural network model and train it, set the training target to 0.01 (that is, the error between the predicted value and the actual value); select the tangent S-shaped TansIg function as the excitation function from the input layer to the hidden layer, select the PurelIn function as the excitation function from the hidden layer to the output layer, set the number of training times to a larger value (generally set to 10,000 times, and the model can converge without reaching the target number of training times), and the learning rate is 0.01. A smaller learning rate can ensure the stability of the system; after training, the network model successfully converges to the training target. At this time, the back propagation neural network model training is completed, and the number of training times is recorded.

[0052] Based on the above neural network model, by inputting real-time online map information, the key quantities of traffic flow information can be obtained, including the traffic volume of motor vehicles, the traffic volume of non-motor vehicles, the traffic volume of pedestrians, and the speed of motor vehicles. In this way, a traffic flow scenario can be constructed on the road scenario.

[0053] The traffic flow scenario obtains online map information from the map at regular intervals, such as 5 minutes, and then uses the above training to complete the neural network model to reconstruct the traffic flow to ensure that it can be accurately reproduced during the test.

[0054] In an embodiment of the present invention, before obtaining the key factors of map information and traffic flow scenes to generate a training data set, it includes: identifying the historical map information and driving scene data of the target path; constructing an initial matrix based on the historical map information and driving scene data, and performing dimensionless processing on the initial matrix to generate a target matrix; calculating the difference between any two elements of the target matrix to generate a difference matrix, and identifying the maximum and minimum values ​​of the difference matrix; calculating the correlation coefficient between each factor and the traffic volume based on the maximum and minimum values, and calculating the gray correlation between each factor and the traffic volume based on the correlation coefficient; selecting factors with a gray correlation greater than a preset threshold as key factors of map information and traffic flow scenes.

[0055] It can be understood that the embodiments of the present invention can identify the historical map information and driving scene data of the target path; construct an initial matrix based on the historical map information and driving scene data, and perform dimensionless processing on the initial matrix to generate a target matrix; calculate the difference between any two elements of the target matrix to generate a difference matrix, and identify the maximum and minimum values ​​of the difference matrix; calculate the correlation coefficient between each factor and the traffic volume based on the maximum and minimum values, and calculate the gray correlation between each factor and the traffic volume based on the correlation coefficient; select factors with a gray correlation greater than a preset threshold as key factors of map information and traffic flow scenes, and use factors with a higher degree of correlation with traffic flow to form training samples to train the neural network model. Compared with conventional training methods, the accuracy of model prediction can be greatly improved, and the calculation efficiency and accuracy can be improved.

[0056] Specifically, the grey correlation algorithm is used to calculate the correlation between each online map information and the traffic volume and the driving speed, and to construct an initial matrix based on the historical online map information and the collected driving scene database; the initial matrix is ​​dimensionlessly processed to obtain the first matrix; the difference matrix is ​​calculated based on the first matrix; the maximum and minimum values ​​of the difference matrix are determined; the correlation coefficient is determined based on the maximum and minimum values; the grey correlation degree is determined based on the correlation coefficient; and the factor with a grey correlation degree greater than a set threshold is selected as the first characteristic factor.

[0057] Specifically, first construct the initial matrix X: (1) In the formula, the first to Nth columns represent the characteristic parameter values ​​of each road section, such as segment distance, segment time consumption, segment congestion level, weather, temperature, segment road type, etc. Columns k to k+3 represent the values ​​of traffic flow scene indicators (i.e., column k is column N+1, column k+1 is column N+2, column k+2 is column N+3, and column k+3 is column N+4), which are the traffic volume of motor vehicles, the traffic volume of non-motor vehicles, the traffic volume of pedestrians, and the speed of motor vehicles, respectively. m represents the number of data rows, that is, the total number of segments collected; Then the initial matrix is ​​dimensionless processed to obtain the first matrix Y. Different weather is represented by different numbers, such as 1 for sunny days and 2 for rainy days. To avoid negative temperature values, the temperature unit uses the thermodynamic temperature scale K.

[0058] (2) In the formula, is the data in the standardized matrix Y, To initialize the data in matrix X, Initialize the j-th column data in the matrix X; i=1,2,3…m, j=1,2,3…N+4.

[0059] Then calculate the difference matrix δY: (3) In the formula, is any element of the difference matrix δY, i=1,2,3…m, j=1,2,3…N+4.

[0060] Calculate the maximum Max and minimum Min of the difference matrix:

[0061] (4) Compute the correlation coefficient matrix: (5) In the formula, It represents the correlation coefficient between the jth map factor and the N+1th factor (in this calculation, the N+1th factor represents the traffic volume of motor vehicles) in the i-th fragment data, ρ is the resolution coefficient, ρ∈(0,1), generally taken as 0.5, i=1,2,3……m, j=1,2,3……N.

[0062] Calculate the grey relational degree: (6) In the formula, Represents the jth map factor and the N+1th factor, i=1,2,3...m, j=1,2,3...N. It can be obtained that the correlation between each factor and the N+1th factor (in this calculation, the N+1th factor represents the traffic volume of motor vehicles) is It represents the correlation coefficient between the jth map factor and the N+1th factor (in this calculation, the N+1th factor represents the traffic volume of motor vehicles) in the i-th segment data.

[0063] Similarly, the grey correlation between the jth map factor and the N+2th, N+3th, and N+4th traffic flow factors can be calculated; the factor with a grey correlation greater than V is selected as the first characteristic factor, and the value of V is generally 0.8.

[0064] If there are N1 factors greater than V, the dimension of the training matrix composed of the first characteristic factor is N1. In this way, factors with low correlation with traffic flow can be eliminated, improving calculation efficiency and accuracy.

[0065] In an embodiment of the present invention, after the target neural network model outputs the traffic flow of the target path, it includes: identifying the traffic volume of adjacent sections in the target path; if the traffic volume difference is greater than a preset threshold, starting the cut-in and cut-out mechanism.

[0066] Among them, the preset threshold can be set according to actual conditions and is not specifically limited.

[0067] It can be understood that the embodiment of the present invention identifies the traffic volume of adjacent sections in the target path; if the traffic volume difference is greater than a preset threshold, the cut-in and cut-out mechanism is activated, thereby ensuring the smoothness of the entire traffic flow and improving the accuracy of the actual traffic scene construction.

[0068] Specifically, the cut-in and cut-out mechanism is to cyclically calculate the difference in motor vehicle traffic volume, non-motor vehicle traffic volume, and pedestrian traffic volume in adjacent segments. If the previous segment is more than the next segment, the intersection is designed to cut out the excess traffic volume. If the previous segment is less than the next segment, the intersection is designed to cut in the excess traffic volume, thereby ensuring the smoothness of the entire traffic flow.

[0069] In step S103, an electric drive system test bench of an electric vehicle is built, the electric drive system test bench is controlled to perform a simulated driving test on a target path under an actual traffic scenario, data of the simulated driving test is collected, and test results of the electric vehicle are analyzed based on the data of the simulated driving test.

[0070] It can be understood that the embodiments of the present invention can build an electric drive system test bench for an electric vehicle, control the electric drive system test bench to perform a simulated driving test on a target path under an actual traffic scenario, collect data from the simulated driving test, analyze the test results of the electric vehicle based on the data from the simulated driving test, and construct an actual traffic scenario by integrating online map information. The vehicle's electric drive efficiency can be tested on a shaft-coupled dynamometer, which can not only improve the accuracy of the test, but also ensure the safety, stability, consistency and repeatability of the test process, thereby reducing the overall cost.

[0071] It should be noted that the test results of the present invention may include the driving efficiency and feedback efficiency during the actual driving process, without specific limitation.

[0072] In an embodiment of the present invention, building an electric drive system test bench for an electric vehicle includes: charging the vehicle to be tested to a target power level; removing the tires of the vehicle to be tested; installing the vehicle to be tested on an axle-coupled dynamometer through a flange, and installing the vehicle's steering tie rod to the vehicle to be tested, initializing the axle-coupled dynamometer; and arranging sensors and data acquisition equipment to synchronously collect vehicle and dynamometer data.

[0073] It can be understood that the embodiments of the present invention can charge the vehicle to be tested to a target power level; remove the tires of the vehicle to be tested; install the vehicle to be tested on an axle-coupled dynamometer through a flange, and install the vehicle's steering tie rod to the vehicle to be tested, initialize the axle-coupled dynamometer, arrange sensors and data acquisition equipment to synchronously collect vehicle and dynamometer data, so as to prepare for the preliminary test, reduce various safety hazards that may be encountered in actual vehicle road testing, and ensure the safety, stability, consistency and repeatability of the test process.

[0074] According to the bench test method for electric vehicles based on actual traffic scenarios proposed in an embodiment of the present invention, a road scene of a target path is constructed according to map information, a traffic flow scene of the target path is constructed according to map information and driving scene data, an actual traffic scene of the target path is generated according to the road scene and the traffic flow scene, an electric drive system bench of the electric vehicle is built, the electric drive system bench is controlled to perform a simulated driving test on the target path under the actual traffic scenario, data of the simulated driving test is collected, and the test results of the electric vehicle are analyzed based on the data of the simulated driving test. The actual traffic scene is constructed by integrating online map information, and the vehicle electric drive efficiency is tested on a shaft-coupled dynamometer. This can not only improve the accuracy of the test, but also ensure the safety, stability, consistency and repeatability of the test process, thereby reducing the overall cost.

[0075] The following will be combined Figure 4 The bench test method of the electric vehicle based on the actual traffic scenario of the present invention is described in detail as follows: Step 1: Scene construction.

[0076] Enter the starting point, end point, and passing points in the online map in the network module to obtain the driving information of the planned route, and then enter the information in the online map into the actual scene construction module to automatically construct the actual traffic scene. During the driving process, automatically obtain the online map information at regular intervals and update the traffic flow scene, so that the constructed scene has strong real-time performance and can be stored in the scene library. Build a simulation model and enter the vehicle parameters to make the simulation model consistent with the parameters of the test vehicle.

[0077] Step 2: Vehicle preparation.

[0078] Charge the vehicle to an appropriate level of power, then remove the tires, install the vehicle on the shaft-coupled dynamometer through the flange, and install the vehicle's steering tie rod on the vehicle. Then enter the wheel diameter, sliding resistance coefficient, and test mass on the shaft-coupled dynamometer control interface. Then perform sliding resistance matching so that the dynamometer can simulate the vehicle's driving resistance at any speed. Arrange current, voltage and other sensors, data acquisition equipment, and synchronously collect vehicle quantity and dynamometer data.

[0079] Step 3: Scene joint debugging.

[0080] Start the host computer, run the simulation model and traffic scene, and start the ring screen, environmental chamber, shaft coupling dynamometer, and data acquisition system at the same time, so that the driver can see the actual traffic scene on the ring screen and drive normally on the route according to the scene, and the vehicle driving is coordinated with the scene.

[0081] Step 4: Test analysis.

[0082] Carry out efficiency tests in actual traffic scenarios, record vehicle information and axle-coupled dynamometer information during the test, and calculate the driving efficiency and feedback efficiency during the actual driving process.

[0083] In summary, the present invention integrates online map information to construct actual traffic scenarios, develops a test device, and implements the test of vehicle electric drive efficiency on an axle-coupled dynamometer, thereby solving the problem of difficulty in energy flow testing in actual traffic scenarios.

[0084] Next, a bench test system for electric vehicles based on actual traffic scenarios proposed according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0085] Figure 5 It is a block diagram of a bench test system for electric vehicles based on actual traffic scenarios according to an embodiment of the present invention.

[0086] like Figure 5As shown, the bench test system 10 for electric vehicles based on actual traffic scenarios includes: a data acquisition module 100, a networking module 200, an actual traffic scenario construction module 300, a vehicle simulation model 400, an axle-coupled dynamometer 500, an axle-coupled dynamometer control module 600, a ring screen 700, an environmental chamber 800 and a host computer 900.

[0087] Among them, the data acquisition module 100 is used to synchronously collect the current and voltage signals, vehicle CAN signals and shaft-coupled dynamometer signals of the vehicle to be tested; the network module 200 is used to obtain the map information and driving scene data of the target path; the actual traffic scene construction module 300, which is connected to the network module 200 in communication, is used to construct the actual traffic scene according to the map information and driving scene data of the target path; the vehicle simulation model 400 is connected to the vehicle to be tested in communication; the shaft-coupled dynamometer 500 is used for vehicle efficiency testing, simulating the target data of the vehicle to be tested, and feeding back the target data to the shaft-coupled dynamometer control module; the shaft-coupled dynamometer control module Block 600 is used to control the shaft-coupled dynamometer; the ring screen 700 is used to display the actual traffic scene to the vehicle to be tested; the environmental chamber 800 is used to simulate the actual ambient temperature to establish the test conditions; the host computer 900 is used to run the network module to obtain the map information and driving scene data of the target path, input the map information and driving scene data of the target path into the actual traffic scene construction module to construct the actual traffic scene, and display the actual traffic scene on the vehicle to be tested; the shaft-coupled dynamometer control module is used to control the shaft-coupled dynamometer, input the data collected by the shaft-coupled dynamometer into the simulation model, and run the simulation model to simulate and obtain the test results.

[0088] Specifically, this system includes a host computer system, which covers a network connection module, an actual traffic scenario construction module, a vehicle simulation model, an axle-coupled dynamometer control module, an axle-coupled dynamometer ring curtain, an environmental chamber, a current sensor, a voltage sensor, and a data acquisition module.

[0089] The host computer system is used to collect and record data from the data acquisition module, run the shaft coupling dynamometer control module to control the operation of the shaft coupling dynamometer, and input the information recorded by the shaft coupling dynamometer into the simulation model, run the simulation model and actual traffic scenarios, run the networking module, and obtain online map information.

[0090] The network connection module is used to obtain online map information and obtain key information of the test path through the online map API (Application Programming Interface) interface.

[0091] The vehicle simulation model is used to integrate the actual vehicle information and run in the actual traffic scenario. The vehicle model includes chassis, electric drive system, steering system, tire system, suspension system, braking system, driver model, thermal management system, aerodynamic model, etc. Among them, the electric drive system, steering system, braking system, and driver model increase the interface for introducing external information, and introduce the wheel speed, steering angle, accelerator pedal, brake pedal and other information fed back by the whole vehicle. Other parts use the existing base model and no additional construction is required.

[0092] The shaft-coupled dynamometer control module is used to control the shaft-coupled dynamometer, input vehicle parameters and test bench parameters, and can collect, record and forward shaft-coupled dynamometer data.

[0093] The shaft-coupled dynamometer is used for vehicle efficiency testing, simulating the driving resistance of the whole vehicle, testing the vehicle's wheel speed, torque, steering angle and other vehicle information, and feeding the above vehicle information into the shaft-coupled dynamometer control module, and then from the control module to the vehicle simulation model, so that the longitudinal and lateral movements of the vehicle are reflected in the actual traffic scene.

[0094] The circular screen is used to display actual traffic scenes, and the driver actually operates the longitudinal and lateral movements of the vehicle based on the information in the scene.

[0095] Environmental chambers are used to simulate actual ambient temperatures.

[0096] Current sensors and voltage sensors are used to test the current and voltage of the vehicle's high-voltage components, such as the electric drive system, power battery, compressor, PTC (Positive Temperature Coefficient) heater, DCDC, etc. The data acquisition module is used to synchronously collect current and voltage signals, vehicle CAN signals, and shaft-coupled dynamometer signals. The collected data is used to analyze the vehicle's energy efficiency in actual traffic scenarios.

[0097] According to the bench test system for electric vehicles based on actual traffic scenarios proposed in an embodiment of the present invention, a road scene of a target path is constructed according to map information, a traffic flow scene of the target path is constructed according to map information and driving scene data, an actual traffic scene of the target path is generated according to the road scene and the traffic flow scene, an electric drive system bench for the electric vehicle is built, the electric drive system bench is controlled to perform a simulated driving test on the target path under the actual traffic scenario, data of the simulated driving test is collected, and the test results of the electric vehicle are analyzed according to the data of the simulated driving test. The actual traffic scene is constructed by integrating online map information, and the vehicle electric drive efficiency is tested on a shaft-coupled dynamometer. This can not only improve the accuracy of the test, but also ensure the safety, stability, consistency and repeatability of the test process, thereby reducing the overall cost.

[0098] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. 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 invention. 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 may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0099] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0100] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0101] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0102] A person skilled in the art may understand that all or part of the steps in the above-mentioned embodiment method may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

Claims

1. A bench test method for electric vehicles based on actual traffic scenarios, characterized in that: The following steps are involved: Obtain map information and driving scenario data of the target path; Constructing a road scene of the target path according to the map information, constructing a traffic flow scene of the target path according to the map information and the driving scene data, and generating an actual traffic scene of the target path according to the road scene and the traffic flow scene; Build an electric drive system test bench for an electric vehicle, control the electric drive system test bench to perform a simulated driving test on the target path under the actual traffic scenario, collect data from the simulated driving test, and analyze the test results of the electric vehicle based on the data from the simulated driving test.

2. The electric vehicle bench test method based on actual traffic scenarios according to claim 1 is characterized in that: The step of constructing the road scene of the target path according to the map information includes: Acquire key road data of the target path in the map information; A road scene of a target path is constructed according to the key road data using a step-by-step incremental method.

3. The electric vehicle bench test method based on actual traffic scenarios according to claim 2 is characterized in that: The step of constructing a road scene of a target path according to the key road data by using a step-by-step incremental method includes: Identify key road data including segment distance, road type, weather conditions, angle of entry to the road, and traffic light locations for each segment of the road; According to the road type, a preset table is searched to determine the corresponding number of lanes, road width and speed limit sign; Determining the friction coefficient of each road segment according to the road type and weather conditions; A road scene is constructed according to the number of lanes, the road width, the speed limit sign, the friction coefficient, the segment distance, the angle of entering the road and the traffic light position.

4. The electric vehicle bench test method based on actual traffic scenarios according to claim 1 is characterized in that: The generating the actual traffic scene of the target path according to the road scene and the traffic flow scene comprises: The road scene and the traffic flow scene are input into a target neural network model, and the target neural network model outputs a traffic flow scene of a target path.

5. The electric vehicle bench test method based on actual traffic scenarios according to claim 4 is characterized in that: The structure of the target neural network model includes an input layer, an output layer and a hidden layer, wherein the number of input layer nodes is the number of vehicle parameters, the number of output layer nodes is the number of fault indicators, and the number of hidden layer nodes is the target number of nodes, wherein the target number of nodes is obtained by using the minimum training number method.

6. The electric vehicle bench test method based on actual traffic scenarios according to claim 5 is characterized in that: The training method of the target neural network model includes: The key factors of the map information and the traffic flow scene are obtained to generate a training data set; wherein the first excitation function is used as the excitation function from the input layer to the hidden layer, and the second excitation function is used as the excitation function from the hidden layer to the output layer; The neural network model is trained using the training data set until the neural network model converges to a training target.

7. The electric vehicle bench test method based on actual traffic scenarios according to claim 6 is characterized in that: Before obtaining the map information and the key factors of the traffic flow scene to generate a training data set, it includes: Identify historical map information and driving scenario data of the target path; constructing an initial matrix according to the historical map information and the driving scene data, and performing dimensionless processing on the initial matrix to generate a target matrix; Calculating the difference between any two elements of the target matrix to generate a difference matrix, and identifying the maximum and minimum values ​​of the difference matrix; Calculate the correlation coefficient between each factor and the traffic volume according to the maximum value and the minimum value, and calculate the grey correlation degree between each factor and the traffic volume according to the correlation coefficient; Factors with grey correlation greater than the preset threshold are selected as key factors of map information and traffic flow scenes.

8. The electric vehicle bench test method based on actual traffic scenarios according to claim 4 is characterized in that: After the target neural network model outputs the traffic flow of the target path, the method further comprises: Identify the traffic volume of adjacent road segments in the target path; If the traffic volume difference is greater than a preset threshold, the cut-in and cut-out mechanism is activated.

9. The electric vehicle bench test method based on actual traffic scenarios according to claim 1 is characterized in that: The electric drive system bench for building an electric vehicle includes: Charge the vehicle to be tested to the target power; Removing the tires of the vehicle to be tested; The vehicle to be tested is mounted on the shaft coupling dynamometer via a flange, and the steering tie rod of the vehicle is mounted on the vehicle to be tested; Initialize the settings of the shaft coupling dynamometer; Arrange sensors and data acquisition equipment to synchronously collect vehicle and dynamometer data.

10. An electric vehicle bench test system based on actual traffic scenarios, characterized in that: The system is applied to implement the electric vehicle bench test method based on actual traffic scenarios as described in any one of claims 1 to 9, wherein the system comprises: Data acquisition module, used to synchronously acquire the current and voltage signals of the vehicle to be tested, the vehicle CAN signals and the shaft coupling dynamometer signals; Network connection module, used to obtain map information and driving scene data of the target path; An actual traffic scene construction module communicatively connected to the network connection module, used to construct an actual traffic scene according to the map information of the target path and the driving scene data; A vehicle simulation model communicatively connected to the vehicle to be tested; A shaft-coupled dynamometer is used for vehicle efficiency testing, simulating target data of the vehicle to be tested, and feeding back the target data to a shaft-coupled dynamometer control module; A shaft-coupled dynamometer control module, used for controlling the shaft-coupled dynamometer; Ring screen, used to display the actual traffic scene to the vehicle under test; Environmental chamber, used to simulate actual ambient temperature to establish test conditions; The host computer is used to run the network connection module to obtain the map information and driving scene data of the target path, input the map information of the target path and the driving scene data into the actual traffic scene construction module to construct the actual traffic scene, and display the actual traffic scene on the vehicle to be tested; the shaft coupling dynamometer control module controls the shaft coupling dynamometer, inputs the data collected by the shaft coupling dynamometer into the simulation model, and runs the simulation model to simulate and obtain the test results.

Images