Real vehicle following model calibration method, device and equipment based on multi-dimensional grid thought and storage medium

Abstract

The invention discloses a real vehicle following model calibration method, device and equipment based on a multi-dimensional grid thought and a storage medium, and relates to the technical field of vehicle motion models. Generalizing the calibration quantity of the theoretical driving type car-following model to be calibrated into a multi-dimensional space coordinate, and constructing a grid value of the calibration quantity under each dimension; determining a multi-dimensional grid calibration algorithm according to the nested cycles with the same number of layers in the multi-dimensional space; combining the grid values of the subscript quantity of each dimension as an input value of the theoretical driving type car-following model to be calibrated, and obtaining a plurality of output values of the theoretical driving type car-following model to be calibrated for representing performance indexes through a multi-dimensional grid calibration algorithm; and according to the output values, the true values in the actual car-following data and the deviation test function, the to-be-calibrated theoretical driving type car-following model is calibrated, the calibrated car-following model is obtained, and thus the efficiency and accuracy of model calibration are improved.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle engineering and vehicle motion model, and in particular to a real vehicle following model calibration method, device, equipment and storage medium based on a multi-dimensional grid concept. Background Art

[0002] The vehicle following model is used to mathematically describe the reaction of the rear vehicle to the motion state adjustment of the front vehicle in the same lane without overtaking. Currently, autonomous driving systems have received widespread attention. Modeling the following behavior can quantify the interaction between vehicles in the longitudinal space, which is of great significance for the development and optimization of vehicle following functions. It has become a research hotspot in this field.

[0003] The car-following model can be divided into theory-driven and data-driven according to the form of expression. However, the data-driven model mainly uses the emerging machine learning technology in recent years to train, learn, iterate, and evolve a large amount of driving data, and finally derives a set of laws that can express the car-following behavior. Most of these models have more stringent requirements on the amount of data, and are usually black box models. The interpretability of the car-following behavior is weak, and it is difficult to optimize after the model training is completed. The research object of this application is a theory-driven car-following model. This type of car-following model can express different output effects by changing the calibration quantity in the model. The formed theoretical model can describe the car-following phenomenon and has strong practical significance.

[0004] Therefore, in view of the defects of data-driven models and future development trends, how to improve the efficiency and accuracy of theory-driven model calibration is a problem that needs to be solved urgently.

[0005] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] The main purpose of this application is to provide a real vehicle following model calibration method, device, equipment and storage medium based on the multi-dimensional grid concept, aiming to solve the technical problem of how to improve the efficiency and accuracy of theory-driven model calibration.

[0007] To achieve the above objectives, the present application proposes a real vehicle following model calibration method based on a multi-dimensional grid concept, the method comprising:

[0008] Obtaining the calibration value of the theoretical driven car-following model to be calibrated;

[0009] Generalizing the calibration quantity into multi-dimensional space coordinates, and constructing a grid value of the calibration quantity in each dimension;

[0010] Determine the multi-dimensional grid calibration algorithm based on the nested loops with the same number of multi-dimensional space layers;

[0011] Obtain actual car-following data collected from real vehicles;

[0012] Combining the grid values ​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated, and obtaining multiple output values ​​of the theoretical driven car-following model to be calibrated that represent performance indicators through the multi-dimensional grid calibration algorithm;

[0013] The theoretical driven car-following model to be calibrated is calibrated according to each of the output values, the true value in the actual car-following data, and a deviation check function to obtain a calibrated car-following model.

[0014] In one embodiment, generalizing the calibration quantity into a multi-dimensional space coordinate and constructing a grid value of the calibration quantity in each dimension includes:

[0015] Determine the corresponding value range of each calibration quantity according to the physical definition;

[0016] Taking the value range as the initial value interval of the corresponding calibration amount;

[0017] Acquiring the calibration accuracy of the theoretical driven car-following model to be calibrated, and determining the step amount corresponding to each calibration amount according to the calibration accuracy;

[0018] Circularly calibrate the calibration amount according to the initial value interval of the calibration amount and the step amount to obtain a target value interval of each calibration amount;

[0019] According to the target value interval of each calibration quantity, a grid value of the calibration quantity under each dimension is constructed.

[0020] In one embodiment, the method of determining the multidimensional grid calibration algorithm based on the nested loops with the same number of multidimensional space layers includes:

[0021] Determine the number of multidimensional space layers according to the grid value of each dimensional subscript quantity;

[0022] A multidimensional grid calibration algorithm is constructed according to the number of multidimensional space layers and the target value intervals of each calibration quantity.

[0023] In one embodiment, the obtaining of actual car-following data collected from a real vehicle includes:

[0024] Collect the motion state data of the vehicle and the preceding vehicle;

[0025] The motion state data of the vehicle and the motion state data of the preceding vehicle are subjected to denoising to obtain actual car-following data.

[0026] In one embodiment, the denoising process is performed on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain the actual following data, including:

[0027] Perform frequency synchronization processing on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain motion state data after frequency synchronization;

[0028] Get the following status;

[0029] Based on the car-following state, deleting the non-car-following data segment in the motion state data after the same frequency, to obtain the car-following data segment;

[0030] The car-following data segments are classified and processed to obtain actual car-following data.

[0031] In one embodiment, the grid values ​​calibrated in each dimension are combined as input values ​​of the theoretical driven car-following model to be calibrated, and multiple output values ​​of the theoretical driven car-following model to be calibrated that characterize performance indicators are obtained through the multi-dimensional grid calibration algorithm, including:

[0032] Based on the actual car-following data, obtaining existing parameter values ​​of the theoretical driven car-following model;

[0033] Combining the grid values ​​of the calibration quantity under each dimension as input values ​​of the theoretical driven car-following model to be calibrated;

[0034] According to each of the input values ​​and the existing parameter values, a plurality of output values ​​of the theoretical driven car-following model to be calibrated that characterize performance indicators are obtained through the multi-dimensional grid calibration algorithm.

[0035] In one embodiment, calibrating the theoretical driven car-following model to be calibrated according to each of the output values, the true value in the actual car-following data, and the deviation check function to obtain the calibrated car-following model includes:

[0036] Obtaining the deviation check function;

[0037] Based on the deviation verification function, the judgment criteria for obtaining the model calibration results are determined;

[0038] Calculating the deviation verification result of each of the output values ​​and the true value sequence in the actual car-following data by using the deviation verification function;

[0039] If the deviation check result meets the judgment standard, the spatial point represented by the corresponding multidimensional grid value of the corresponding output value is the model calibration result.

[0040] In addition, to achieve the above purpose, the present application also proposes a real vehicle following model calibration device based on the multi-dimensional grid concept, and the real vehicle following model calibration device based on the multi-dimensional grid concept includes:

[0041] A calibration module is used to obtain the calibration quantity of the theoretical driven car-following model to be calibrated, generalize the calibration quantity into multi-dimensional space coordinates, construct grid values ​​of the calibration quantity under each dimension, and determine the multi-dimensional grid calibration algorithm according to nested loops with the same number of multi-dimensional space layers;

[0042] A collection module, used to obtain actual car-following data collected from a real vehicle;

[0043] a calculation module, used for combining the grid values ​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated, and obtaining a plurality of output values ​​of the theoretical driven car-following model to be calibrated representing performance indicators through the multi-dimensional grid calibration algorithm;

[0044] The output module is used to calibrate the theoretical driven car-following model to be calibrated according to each of the output values, the true value in the actual car-following data and the deviation check function to obtain a calibrated car-following model.

[0045] In addition, to achieve the above-mentioned purpose, the present application also proposes a real vehicle following model calibration device based on the multi-dimensional grid concept, the device comprising: a memory, a processor and a computer program stored on the memory and executable on the processor, the computer program being configured to implement the steps of the real vehicle following model calibration method based on the multi-dimensional grid concept as described above.

[0046] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the real vehicle following model calibration method based on the multi-dimensional grid concept as described above are implemented.

[0047] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the real vehicle following model calibration method based on the multi-dimensional grid concept as described above.

[0048] One or more technical solutions proposed in this application have at least the following technical effects:

[0049] The performance of the car-following model when each grid value combination in each dimension is calculated through the constructed calibration algorithm. The calibration process is hierarchical and progressive, which can ensure that any point in the multi-dimensional space is traversed, and the vehicle status information is collected and processed to obtain a true value sequence. The model calibration result is determined based on the deviation function model output value and the true value sequence, thereby solving the technical problem of how to improve the efficiency and accuracy of theory-driven model calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0052] Figure 1 A flow chart of a first embodiment of a method for calibrating a real vehicle following model based on a multi-dimensional grid concept of the present application;

[0053] Figure 2 Vehicles and equipment that can be used as references when collecting car-following data provided in Example 1 of the real-car car-following model calibration method based on the multi-dimensional grid concept of this application;

[0054] Figure 3 When the speed provided in the first embodiment of the method for calibrating a real car following model based on the multi-dimensional grid concept of this application is used as the performance indicator, the comparison between the prediction result of the car following behavior of the car following model calibrated by the method and the actual value;

[0055] Figure 4 When the acceleration provided in the first embodiment of the method for calibrating a real car following model based on the multi-dimensional grid concept of this application is used as the performance indicator, the comparison between the prediction result of the following behavior of the following model calibrated by the method and the actual value;

[0056] Figure 5 A flow chart of a second embodiment of a method for calibrating a real vehicle following model based on a multi-dimensional grid concept of the present application;

[0057] Figure 6 A brief flowchart of a real vehicle following model calibration method based on a multi-dimensional grid concept in this application;

[0058] Figure 7 This is a schematic diagram of the module structure of a real vehicle following model calibration device based on the multi-dimensional grid concept in an embodiment of the present application;

[0059] Figure 8Schematic diagram of the device structure of the hardware operating environment involved in the real vehicle following model calibration method based on the multi-dimensional grid concept in the embodiment of the present application.

[0060] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0061] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0062] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0063] The main solution of the embodiment of the present application is: obtaining the calibration quantity of the theoretical driven car-following model to be calibrated; generalizing the calibration quantity into multi-dimensional space coordinates, and constructing grid values ​​of the calibration quantity under each dimension; determining the multi-dimensional grid calibration algorithm according to the nested loops with the same number of multi-dimensional space layers; obtaining the actual car-following data collected by the real vehicle; combining the grid values ​​of the calibration quantity under each dimension as the input value of the theoretical driven car-following model to be calibrated, and obtaining multiple output values ​​of the theoretical driven car-following model to be calibrated that characterize the performance indicators through the multi-dimensional grid calibration algorithm; calibrating the theoretical driven car-following model to be calibrated according to each of the output values, the true value in the actual car-following data, and the deviation check function to obtain a calibrated car-following model.

[0064] Since most of the existing data-driven technologies have more stringent requirements on the amount of data and are usually black-box models, the interpretability of following behavior is weak and it is difficult to optimize the model after training. Therefore, we choose to adopt a theory-driven model, which focuses on the assumptions about the driver's reactions. Since the model comes from the observation of driving behavior, the expression of the model can be more in line with the characteristics of human drivers, which is consistent with the anthropomorphic driving in autonomous driving technology.

[0065] The present application provides a solution. The present application calculates the performance of the following model when each grid value combination in each dimension through a constructed calibration algorithm. The calibration process is hierarchical and progressive, which can ensure that any point in the multi-dimensional space is traversed, and the vehicle status information is collected and processed to obtain a true value sequence. The model calibration result is determined based on the deviation function model output value and the true value sequence, thereby solving the technical problem of how to improve the efficiency and accuracy of theory-driven model calibration.

[0066] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, etc. The following takes the vehicle-mounted computer as an example to illustrate this embodiment and the following embodiments.

[0067] Based on this, the embodiment of the present application provides a real vehicle following model calibration method based on the multi-dimensional grid concept, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the real vehicle following model calibration method based on the multi-dimensional grid concept of this application.

[0068] In this embodiment, the real vehicle following model calibration method based on the multi-dimensional grid concept includes steps S10 to S60:

[0069] Step S10, obtaining a calibration value of a theoretical driving car-following model to be calibrated;

[0070] It should be noted that the theory-driven car-following model refers to a car-following model with clear physical properties that is established based on the car-following theory using traditional mathematical and physical modeling methods. It considers the physical state of the vehicle and the driver's behavioral decisions during the car-following process, and converts them into physical quantities and introduces them into the model. There are usually multiple calibration quantities in the model. The calibration quantity can be used to calibrate and verify the car-following model to ensure that the model can accurately reflect the actual traffic flow situation. This includes adjusting the model parameters to match the actual observed traffic behavior.

[0071] Furthermore, taking the generalized theory-driven car-following model as an example, for the model V(X,Y,Z)=lX+mY+nZ, the model calibration quantities are l, m, n,

[0072] The method is explained in detail using the classic IDM (Intelligent Driver Model, IDM) intelligent driver following model. The model is expressed as follows:

[0073]

[0074] The quantity to be calibrated is the comfortable deceleration Maximum acceleration / deceleration Expected speed Expected headway Blocking distance at rest Acceleration exponent β.

[0075] It can be understood that, since a calibration algorithm needs to be constructed, step S10 is performed to determine the object of each layer cycle of the calibration algorithm, thereby improving the accuracy of the calibration algorithm.

[0076] Step S20, generalizing the calibration quantity into multi-dimensional space coordinates, and constructing grid values ​​of the calibration quantity in each dimension;

[0077] It should be noted that the multidimensional space coordinates are obtained by generalizing the model calibration using the multidimensional grid concept. For example, the multidimensional space coordinates (x 1 ,x 2 ,x 3 ,……x n ), the multidimensional grid idea is to divide the multidimensional space into a finite number of cells or grids so that calculations, analysis and processing can be performed on these grids. The grid value usually refers to the value at a specific location in the multidimensional grid space. This value can represent a variety of different physical quantities or data points.

[0078] Furthermore, taking the generalized theory-driven car-following model as an example, for the model V(X,Y,Z)=lX+mY+nZ, the model calibration quantities are l, m, n, and the multi-dimensional grid idea is used to generalize the model calibration quantities to three-dimensional space coordinates (l,m,n). According to the physical definition, it is determined that l, m, n have a corresponding value range. Any coordinate (l′,m′,n′) within the value range can represent the model calibration result, and the corresponding car-following model can be expressed as V(X,Y,Z)=l′X+m′Y+n′Z.

[0079] Based on this, a three-layer nested loop is constructed. For the calibration quantity l∈[a, b], when the model calibration accuracy is set to i, the step amount of the calibration quantity l in the coordinate system is also . At this time, the value range of the calibration quantity l is [a, a+i, a+2i,…, a+ni, b]. In the first layer of the loop, the calibration quantity l traverses all values ​​in this interval. The same is true for the calibration quantities m and n, and the construction of the calibration model is completed.

[0080] It is understandable that, since a calibration algorithm needs to be constructed, step S20 is performed to obtain grid values ​​of the calibration quantity, thereby providing support for constructing the calibration algorithm.

[0081] In a feasible implementation, step S10 may include steps A11 to A15:

[0082] Step A11, determining the corresponding value range of each calibration amount according to the physical definition;

[0083] It should be noted that the corresponding value range of the calibration quantity refers to the value range of the coordinates on the grid of each dimension defined according to the physical meaning of the calibration quantity. n ∈[a,b].

[0084] Step A12, using the value range as the initial value interval of the corresponding calibration amount;

[0085] It should be noted that the initial value interval refers to the value range of the coordinates on the grid of each dimension defined according to the physical meaning of the calibration quantity x n ∈[a,b].

[0086] Exemplarily, the method is further explained using the classic IDM (Intelligent Driver Model, IDM) intelligent driver following model.

[0087] There are 6 calibration quantities in the IDM model, based on which the calibration results can be expressed using 6-dimensional spatial coordinates. The range of the spatial grid in each dimension is consistent with the range of the calibration quantity, for example The physical value range of is 0.15-4, and the corresponding initial value interval is also 0.15-4.

[0088] Step A13, obtaining the calibration accuracy of the theoretical driven car-following model to be calibrated, and determining the step amount corresponding to each calibration amount according to the calibration accuracy.

[0089] It should be noted that calibration accuracy usually refers to the accuracy level that an instrument or system can achieve during the calibration process. The step size is the linear displacement of the target value range of the indicator quantification. The step size of the calibration quantity in each dimension is equal to the calibration accuracy.

[0090] Step A14, cyclically calibrating the calibration quantity according to the initial value interval of the calibration quantity and the step amount, to obtain a target value interval of each calibration quantity.

[0091] For example, the IDM model is used to further illustrate that, assuming the model calibration accuracy is 0.05, then for The target value range can be expressed as [0.15, 0.20, 0.25, …, 3.95, 4.0].

[0092] Step A15, constructing grid values ​​of the calibration quantities in each dimension according to the target value intervals of each calibration quantity.

[0093] In this implementation, the range of the calibration quantity is determined according to the physical meaning of the calibration quantity, and then further refined to obtain the grid values ​​of the calibration quantity in each dimension, ensuring that the calibration algorithm can verify the values ​​on the grid of each dimension as much as possible and without omission, thereby improving the accuracy of the calibration algorithm.

[0094] The above is only a feasible implementation of step S20 provided in this embodiment, and this embodiment does not specifically limit the specific implementation of step S20.

[0095] Step S30, determining a multidimensional grid calibration algorithm according to nested loops with the same number of multidimensional space layers;

[0096] It should be noted that the number of multidimensional space layers is equal to the number of calibration quantities. The multidimensional grid calibration algorithm needs to ensure that the values ​​on the grid of each dimension are verified as much as possible and without omission. Based on the multidimensional grid idea, nested loops with the same number of multidimensional space layers are constructed. The step distance of the calibration quantity in each loop is equal to the calibration accuracy. This algorithm is the multidimensional grid calibration algorithm.

[0097] For example, the IDM model is used to further illustrate that The target value set interval is [0.15, 0.20, 0.25, …, 3.95, 4.0]. In the constructed nested loop algorithm, the corresponding dimension is Recursively follow the values ​​in the set.

[0098] It can be understood that determining the number of multidimensional space layers according to the grid value of the calibration quantity under each dimension improves the accuracy of the calibration algorithm according to the number of multidimensional space layers and the target value range of each calibration quantity.

[0099] Step S40, obtaining actual car-following data collected from a real vehicle.

[0100] It should be noted that the actual car-following data is obtained by collecting information such as vehicle distance, vehicle speed, throttle / pedal opening, steering wheel steering angle, etc. when following the vehicle in a real road environment. The collected data is then processed by synchronization, cleaning, slicing, and classification. The processed data is the actual car-following data, which can be used as the target set for model calibration.

[0101] For example, the data collection equipment can be realized by using a test vehicle with perception capabilities and related data collection equipment. The vehicles and equipment that can be used as references in the data collection process are as follows: Figure 2 As shown, for example, ADAS (Advanced Driver Assistance System), vehicle-mounted forward radar, vehicle-mounted forward camera, CAN data acquisition equipment, and CANoe / CANalyzer analysis software.

[0102] It can be understood that since it is necessary to calculate the mean square deviation between the model output value and the actual vehicle speed sequence, S40 is performed to reduce the error existing in the actual car-following data and improve the data accuracy.

[0103] In a feasible implementation, step S40 may include steps B11 to B15:

[0104] Step B11, collecting the motion state data of the vehicle and the motion state data of the preceding vehicle;

[0105] It should be noted that the motion state data generally includes the vehicle's position, speed, acceleration, attitude (including roll angle, pitch angle and yaw angle), yaw angular velocity and vehicle sideslip angle, etc.

[0106] Step B12, performing frequency synchronization processing on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain frequency synchronized motion state data;

[0107] Step B13, obtaining the following car status.

[0108] It should be noted that the following state usually refers to the state in which the rear vehicle follows the front vehicle in the traffic flow, in which the rear vehicle adjusts its own acceleration to maintain a safe distance from the front vehicle. In autonomous driving and vehicle control systems, understanding and simulating this following behavior is crucial to improving the stability, safety and efficiency of traffic flow.

[0109] Step B14, based on the car-following state, deleting the non-car-following data segments in the motion state data after the synchronization, to obtain the car-following data segments.

[0110] Step B15, classifying the car-following data segments to obtain actual car-following data.

[0111] In this implementation manner, actual car-following data is obtained by performing denoising processing on the motion state data of the vehicle and the motion state data of the preceding vehicle, thereby improving the accuracy of the actual car-following data.

[0112] Step S50, combining the grid values ​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated, and obtaining multiple output values ​​of the theoretical driven car-following model to be calibrated that characterize performance indicators through the multi-dimensional grid calibration algorithm.

[0113] It should be noted that performance indicators refer to a series of quantitative standards used to measure the performance and efficiency of a system, equipment, service or process.

[0114] Furthermore, taking the generalized theory-driven car-following model as an example, for the model V(X,Y,Z)=lX+mY+nZ, the model output vehicle speed V(X,Y,Z) is used as the performance indicator, and the constructed calibration algorithm is used to calculate when the multi-dimensional grid coordinates are (l′,m′,n′), the model outputs multiple vehicle speeds V(X,Y,Z) within the driving time t.

[0115] It can be understood that through the multidimensional network calibration algorithm, multiple output values ​​of the model are calculated, which provides data support for the subsequent verification of model fit and improves the efficiency of model calibration.

[0116] Step S60, calibrating the theoretical driven car-following model to be calibrated according to each of the output values, the true value in the actual car-following data, and a deviation check function to obtain a calibrated car-following model.

[0117] It should be noted that the true value refers to the numerical sequence of the same attribute as the output value in the actual car-following data collected. The deviation test function refers to the function or method used in statistics and machine learning to evaluate and correct the model prediction deviation. The bias refers to the difference between the model prediction value and the true value, which reflects the degree of fit of the model.

[0118] Furthermore, taking the generalized theory-driven car-following model as an example, for the model V(X,Y,Z)=lX+mY+nZ, the model output speed V(X,Y,Z) is used as the performance indicator, and the mean square error R 2 As a deviation check function, the constructed calibration algorithm is used to calculate the difference between the model output speed V(X,Y,Z) sequence and the actual speed V in the target set within the driving time t when the multidimensional grid coordinates are (l′,m′,n′). R The mean square deviation R of the series 2 , when satisfied , this spatial point is the model calibration result.

[0119] For example, the IDM model is used to further illustrate that the output of the model is the rear vehicle acceleration a n Therefore, acceleration is used as a performance indicator to evaluate the calibration effect of the model (a n Back-calculated into other physical quantities as performance indicators), when the calibration grid coordinates are ( β(x)), the IDM model can output a set of acceleration predictions a according to the known motion states of the front and rear vehicles. n (x), compare the mean square deviation R between the output acceleration value and the real acceleration value in the target data set 2 After jumping out of each layer of loop, find satisfaction The spatial point of comf , a max , S jam , β is the calibration result. Figure 3 and Figure 4 When speed and acceleration are used as performance indicators, the following behavior prediction results of the car-following model calibrated using this method are compared with the actual values.

[0120] It can be understood that by calculating the deviation between each output value of the model and the true value sequence, the degree of fit between different calibration value combinations and the model is determined, thereby improving the accuracy of model calibration.

[0121] This embodiment provides a real vehicle following model calibration method based on the multi-dimensional grid concept. The performance of the following model when each grid value combination in each dimension is calculated through a constructed calibration algorithm. The calibration process is hierarchical and progressive, which can ensure that any point in the multi-dimensional space is traversed, and the vehicle status information is collected and processed to obtain a true value sequence. The model calibration result is determined based on the deviation function model output value and the true value sequence, thereby solving the technical problem of how to improve the efficiency and accuracy of theory-driven model calibration.

[0122] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 5 Before step S60, the real vehicle following model calibration method based on the multi-dimensional grid concept further includes steps S61 to S64:

[0123] Step S61, obtaining the deviation check function;

[0124] It should be noted that the deviation check function refers to a function or algorithm used to verify and correct the deviation between the measured or estimated value and the true value. Taking the generalized theory-driven car-following model as an example, the mean square deviation R 2 as a deviation test function.

[0125] Step S62, determining a judgment criterion for obtaining a model calibration result based on the deviation verification function;

[0126] Step S63, calculating the deviation verification result between each of the output values ​​and the real value sequence in the actual car-following data by using a deviation verification function;

[0127] Step S64: if the deviation check result meets the judgment standard, the spatial point represented by the corresponding multidimensional grid value of the corresponding output value is the model calibration result.

[0128] For example, taking the generalized theory-driven car-following model as an example, the mean square error R 2 As a deviation check function, the judgment standard of the model calibration result is to calculate the difference between the model output speed V(X,Y,Z) sequence and the actual speed V in the target set within the driving time t when the multidimensional grid coordinates are (l′,m′,n′). R The mean square deviation R of the series 2 , when satisfied , this spatial point is the model calibration result.

[0129] For example, in order to help understand the implementation process of the real vehicle following model calibration method based on the multi-dimensional grid concept obtained by combining the present embodiment with the above-mentioned embodiment 1 and embodiment 2, please refer to Figure 6 , Figure 6A brief flow chart of the real vehicle following model calibration method based on the multi-dimensional grid concept is provided. Specifically:

[0130] The first step is to select the car-following model to be calibrated and extract the model calibration quantity. In order to make the output of the car-following model match the vehicle motion state in the real car-following state, the multi-dimensional grid idea is used to generalize these calibration quantities into multi-dimensional space coordinates (x 1 ,x 2 ,x 3 ,……x n ), and then determine the value range according to the physical definition of the calibration quantity, set the accuracy of the model calibration quantity, and obtain that the coordinate values ​​of each dimension satisfy x n ∈[a, b], based on which a grid of values ​​for the quantified values ​​under each dimension can be constructed. The calibration algorithm needs to ensure that the values ​​on the grid of values ​​of each dimension are verified as much as possible without omission. Based on this idea, a nested loop with the same number of layers in the multidimensional space is constructed. The step distance of the quantified value in each loop is equal to the calibration accuracy, and a multidimensional grid calibration algorithm can be obtained.

[0131] The second step is to collect car-following data from the actual vehicle, process the collected car-following data, and obtain the calibration target set. This is to set up cameras, radars and other sensing devices on the test vehicle to support the collection of information such as vehicle distance, vehicle speed, throttle / pedal opening, steering wheel steering angle, etc. The vehicle is operated in a real road environment to follow the vehicle in front, and the collected valid car-following data is used as the target set for calibrating the car-following model. The target set is the actual vehicle following data mentioned above.

[0132] Finally, the calibration algorithm is run with the output of the car-following model as the performance indicator and the mean square deviation as the deviation verification function. When the corresponding spatial coordinates meet the calibration results, the model calibration is completed. That is, the output of the theoretical driven car-following model is used as the performance indicator, and the mean square deviation R 2 As a deviation check function, a multidimensional grid calibration algorithm is used to calculate the performance of the car-following model when each grid value combination in each dimension is used. The calibration process is hierarchical and progressive, which can ensure that any point in the multidimensional space is traversed. When any point in the space is brought into the car-following model, the mean square deviation R between the model output value and the true value sequence of the performance index can be obtained. 2 ,satisfy The spatial point is the model calibration result.

[0133] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the real vehicle following model calibration method based on the multi-dimensional grid concept of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0134] This application also provides a real vehicle following model calibration device based on the multi-dimensional grid concept, please refer to Figure 7The real vehicle following model calibration device based on the multi-dimensional grid concept includes:

[0135] The calibration module 10 is used to obtain the calibration quantity of the theoretical driven car-following model to be calibrated, generalize the calibration quantity into multi-dimensional space coordinates, construct grid values ​​of the calibration quantity under each dimension, and determine the multi-dimensional grid calibration algorithm according to the nested loops with the same number of multi-dimensional space layers;

[0136] A collection module 20, used to obtain actual car-following data collected from a real vehicle;

[0137] A calculation module 30 is used to combine the grid values ​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated, and obtain multiple output values ​​of the theoretical driven car-following model to be calibrated that represent performance indicators through the multi-dimensional grid calibration algorithm;

[0138] The output module 40 is used to calibrate the theoretical driven car-following model to be calibrated according to the output values, the real values ​​in the actual car-following data and the deviation check function to obtain a calibrated car-following model.

[0139] The real vehicle following model calibration device based on the multi-dimensional grid concept provided by the present application adopts the real vehicle following model calibration method based on the multi-dimensional grid concept in the above-mentioned embodiment, which can solve the technical problem of real vehicle following model calibration based on the multi-dimensional grid concept. Compared with the prior art, the beneficial effects of the real vehicle following model calibration device based on the multi-dimensional grid concept provided by the present application are the same as the beneficial effects of the real vehicle following model calibration method based on the multi-dimensional grid concept provided by the above-mentioned embodiment, and the other technical features in the real vehicle following model calibration device based on the multi-dimensional grid concept are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0140] In one embodiment, the calibration module 10 is further used to determine the corresponding value range of each calibration quantity according to the physical definition; use the value range as the initial value interval of the corresponding calibration quantity; obtain the calibration accuracy of the theoretical driven car-following model to be calibrated, and determine the step amount corresponding to each calibration quantity according to the calibration accuracy; cyclically calibrate the calibration quantity according to the initial value interval of the calibration quantity and the step amount to obtain the target value interval of each calibration quantity; and construct the grid value of the calibration quantity under each dimension according to the target value interval of each calibration quantity.

[0141] In one embodiment, the calibration module 10 is further used to determine the number of multidimensional space layers according to the grid value of the calibration quantity under each dimension; and to construct a multidimensional grid calibration algorithm according to the number of multidimensional space layers and the target value range of each calibration quantity.

[0142] In one embodiment, the acquisition module 20 is further used to acquire the motion state data of the vehicle and the motion state data of the preceding vehicle; and to perform denoising on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain actual car-following data.

[0143] In one embodiment, the acquisition module 20 is also used to perform isochronous processing on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain the isochronous motion state data; obtain the following state; based on the following state, delete the non-following data segments in the isochronous motion state data to obtain the following data segments; classify the following data segments to obtain the actual following data.

[0144] In one embodiment, the calculation module 30 is also used to obtain the existing parameter values ​​of the theoretical driven car-following model based on the actual car-following data; combine the grid values ​​​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated; and obtain multiple output values ​​of the theoretical driven car-following model to be calibrated that characterize performance indicators through the multi-dimensional grid calibration algorithm based on each of the input values ​​and the existing parameter values.

[0145] In one embodiment, the output module 40 is also used to obtain the deviation verification function; based on the deviation verification function, determine the judgment standard for obtaining the model calibration result; through the deviation verification function, calculate the deviation verification result of each output value and the true value sequence in the actual following data; if the deviation verification result meets the judgment standard, the spatial point represented by the corresponding multidimensional grid value of the corresponding output value is the model calibration result.

[0146] The present application provides a real vehicle following model calibration device based on the multi-dimensional grid concept, and the real vehicle following model calibration device based on the multi-dimensional grid concept includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor so that at least one processor can execute the real vehicle following model calibration method based on the multi-dimensional grid concept in the above-mentioned embodiment one.

[0147] Reference below Figure 8, which shows a schematic diagram of the structure of a real car following model calibration device based on the multi-dimensional grid concept suitable for implementing the embodiment of the present application. The real car following model calibration device based on the multi-dimensional grid concept in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The real vehicle following model calibration device based on the multi-dimensional grid concept shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0148] like Figure 8 As shown, the real vehicle following model calibration device based on the multi-dimensional grid concept may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 to the random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the real vehicle following model calibration device based on the multi-dimensional grid concept are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other through a bus 1005. The input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the real vehicle following model calibration device based on the multi-dimensional grid concept to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a real vehicle following model calibration device based on the multi-dimensional grid concept with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0149] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0150] The real vehicle following model calibration device based on the multi-dimensional grid concept provided by the present application adopts the real vehicle following model calibration method based on the multi-dimensional grid concept in the above-mentioned embodiment, which can solve the technical problem of real vehicle following model calibration based on the multi-dimensional grid concept. Compared with the prior art, the beneficial effect of the real vehicle following model calibration device based on the multi-dimensional grid concept provided by the present application is the same as the beneficial effect of the real vehicle following model calibration method based on the multi-dimensional grid concept provided by the above-mentioned embodiment, and the other technical features in the real vehicle following model calibration device based on the multi-dimensional grid concept are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0151] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0152] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0153] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the real vehicle following model calibration method based on the multi-dimensional grid concept in the above-mentioned embodiment.

[0154] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, 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 this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0155] The computer-readable storage medium may be included in the real vehicle car-following model calibration device based on the multi-dimensional grid concept; or may exist independently without being assembled into the real vehicle car-following model calibration device based on the multi-dimensional grid concept.

[0156] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the real vehicle following model calibration device based on the multi-dimensional grid concept, the real vehicle following model calibration device based on the multi-dimensional grid concept enables the real vehicle following model calibration device based on the multi-dimensional grid concept to: obtain the calibration quantity of the theoretical driven following model to be calibrated; generalize the calibration quantity into multi-dimensional space coordinates, and construct grid values ​​of the calibration quantity under each dimension; determine the multi-dimensional grid calibration algorithm according to the nested loops with the same number of multi-dimensional space layers; obtain the actual following data collected by the real vehicle; combine the grid values ​​of the calibration quantity under each dimension as the input value of the theoretical driven following model to be calibrated, and obtain multiple output values ​​of the theoretical driven following model to be calibrated that represent performance indicators through the multi-dimensional grid calibration algorithm; calibrate the theoretical driven following model to be calibrated according to each of the output values, the true value in the actual following data and the deviation check function to obtain a calibrated following model.

[0157] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, 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 separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of 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., via the Internet using an Internet service provider).

[0158] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0159] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0160] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned real vehicle following model calibration method based on the multi-dimensional grid concept, and can solve the technical problems of real vehicle following model calibration based on the multi-dimensional grid concept. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the real vehicle following model calibration method based on the multi-dimensional grid concept provided in the above-mentioned embodiment, and will not be elaborated here.

[0161] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the real vehicle following model calibration method based on the multi-dimensional grid concept as described above.

[0162] The computer program product provided by the present application can solve the technical problem of real vehicle following model calibration based on the multi-dimensional grid concept. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the real vehicle following model calibration method based on the multi-dimensional grid concept provided by the above embodiment, which will not be repeated here.

[0163] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A real vehicle following model calibration method based on multi-dimensional grid concept, characterized in that: The method includes: Obtaining the calibration value of the theoretical driven car-following model to be calibrated; Generalizing the calibration quantity into multi-dimensional space coordinates, and constructing a grid value of the calibration quantity in each dimension; Determine the multi-dimensional grid calibration algorithm based on the nested loops with the same number of multi-dimensional space layers; Obtain actual car-following data collected from real vehicles; Combining the grid values ​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated, and obtaining multiple output values ​​of the theoretical driven car-following model to be calibrated that represent performance indicators through the multi-dimensional grid calibration algorithm; The theoretical driven car-following model to be calibrated is calibrated according to each of the output values, the true value in the actual car-following data, and a deviation check function to obtain a calibrated car-following model.

2. The method according to claim 1, characterized in that The step of generalizing the calibration quantity into a multi-dimensional space coordinate and constructing a grid value of the calibration quantity in each dimension includes: Determine the corresponding value range of each calibration quantity according to the physical definition; Taking the value range as the initial value interval of the corresponding calibration amount; Acquiring the calibration accuracy of the theoretical driven car-following model to be calibrated, and determining the step amount corresponding to each calibration amount according to the calibration accuracy; Circularly calibrate the calibration amount according to the initial value interval of the calibration amount and the step amount to obtain a target value interval of each calibration amount; According to the target value interval of each calibration quantity, a grid value of the calibration quantity under each dimension is constructed.

3. The method according to claim 1, characterized in that Determining the multidimensional grid calibration algorithm according to the nested loops with the same number of multidimensional space layers includes: Determine the number of multidimensional space layers according to the grid value of each dimensional subscript quantity; A multidimensional grid calibration algorithm is constructed according to the number of multidimensional space layers and the target value intervals of each calibration quantity.

4. The method according to claim 1, characterized in that The obtaining of actual car-following data collected from a real vehicle includes: Collect the motion state data of the vehicle and the preceding vehicle; The motion state data of the vehicle and the motion state data of the preceding vehicle are subjected to denoising to obtain actual car-following data.

5. The method according to claim 4, characterized in that The denoising process is performed on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain actual car-following data, including: Perform frequency synchronization processing on the motion state data of the vehicle and the motion state data of the preceding vehicle to obtain motion state data after frequency synchronization; Get the following status; Based on the car-following state, deleting the non-car-following data segment in the motion state data after the same frequency, to obtain the car-following data segment; The car-following data segments are classified and processed to obtain actual car-following data.

6. The method according to claim 1, characterized in that The grid values ​​calibrated in each dimension are combined as input values ​​of the theoretical driven car-following model to be calibrated, and multiple output values ​​of the theoretical driven car-following model to be calibrated representing performance indicators are obtained through the multi-dimensional grid calibration algorithm, including: Based on the actual car-following data, obtaining existing parameter values ​​of the theoretical driven car-following model; Combining the grid values ​​of the calibration quantity under each dimension as input values ​​of the theoretical driven car-following model to be calibrated; According to each of the input values ​​and the existing parameter values, a plurality of output values ​​of the theoretical driven car-following model to be calibrated that characterize performance indicators are obtained through the multi-dimensional grid calibration algorithm.

7. The method according to claim 1, characterized in that The step of calibrating the theoretical driven car-following model to be calibrated according to each of the output values, the true value in the actual car-following data, and the deviation check function to obtain a calibrated car-following model includes: Obtaining the deviation check function; Based on the deviation verification function, the judgment criteria for obtaining the model calibration results are determined; Calculating the deviation verification result of each of the output values ​​and the true value sequence in the actual car-following data by using the deviation verification function; If the deviation check result meets the judgment standard, the spatial point represented by the corresponding multidimensional grid value of the corresponding output value is the model calibration result.

8. A real vehicle following model calibration device based on multi-dimensional grid concept, characterized in that: The device comprises: A calibration module is used to obtain the calibration quantity of the theoretical driven car-following model to be calibrated, generalize the calibration quantity into multi-dimensional space coordinates, construct grid values ​​of the calibration quantity under each dimension, and determine the multi-dimensional grid calibration algorithm according to nested loops with the same number of multi-dimensional space layers; A collection module, used to obtain actual car-following data collected from a real vehicle; a calculation module, used for combining the grid values ​​calibrated in each dimension as input values ​​of the theoretical driven car-following model to be calibrated, and obtaining a plurality of output values ​​of the theoretical driven car-following model to be calibrated representing performance indicators through the multi-dimensional grid calibration algorithm; The output module is used to calibrate the theoretical driven car-following model to be calibrated according to each of the output values, the true value in the actual car-following data and the deviation check function to obtain a calibrated car-following model.

9. A real vehicle following model calibration device based on multi-dimensional grid concept, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the real vehicle following model calibration method based on the multi-dimensional grid concept as described in any one of claims 1 to 7, and the steps of the method.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the real vehicle following model calibration method based on the multi-dimensional grid concept as described in any one of claims 1 to 7 are implemented.

Images