Method and system for evaluating comprehensive performance of vehicle-road collaborative navigation of whole vehicle

Abstract

The invention discloses a comprehensive performance evaluation method and system for vehicle-road collaborative navigation of a whole vehicle, relates to the technical field of aided driving, and solves the technical problems that complexity of different road areas is difficult to consider, reasonable variation of formed navigation performance is poor, and potential faults existing in a navigation system are difficult to accurately analyze. The actual condition of the road is visually obtained through the road top view image data, weather factors are conveniently considered in combination with weather data, and the influence of visibility on the road environment is reduced. By comprehensively analyzing the factors, the complexity of the road environment can be evaluated more accurately, and a reliable basis is provided for subsequent grid data grading. According to the method, the complexity of the road environment is divided into different grid data grades, the grid data grades can clearly display the complexity of the road environment in different areas, vehicle operation data, road state data, vehicle-road cooperative communication data and weather data are analyzed, and the problem of the vehicle in the vehicle-road cooperative navigation process can be conveniently analyzed in real time.

Description

Technical Field

[0001] The present invention belongs to the field of assisted driving, and is specifically a comprehensive performance evaluation method and system for vehicle-road collaborative navigation. A method for positioning and labeling intervertebral discs based on digital twins. Background Art

[0002] With the rapid development of intelligent transportation systems, vehicle-road cooperative technology has become a key means to improve road traffic safety, efficiency and intelligence. As an important part of it, vehicle-road cooperative navigation provides drivers with more accurate and real-time navigation services by realizing information interaction and sharing between vehicles and road infrastructure. However, with the continuous advancement of technology and the expansion of application scenarios, the performance requirements for vehicle-road cooperative navigation of the whole vehicle are becoming higher and higher. Therefore, it is of great significance to establish a scientific and comprehensive comprehensive performance evaluation system for evaluating and optimizing the performance of vehicle-road cooperative navigation of the whole vehicle. The comprehensive performance evaluation system of vehicle-road cooperative navigation of the whole vehicle analyzes the problems existing in the navigation of the whole vehicle during the vehicle-road cooperative navigation process by integrating multiple technical means and evaluation indicators.

[0003] Most of the existing performance evaluation schemes for vehicle-road cooperative navigation perform a unified analysis of vehicle navigation data, which makes it difficult to take into account the complexity of different road areas, resulting in a reasonable deterioration in navigation performance and accurately analyzing problems existing in the navigation system. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a comprehensive performance evaluation method and system for vehicle-road collaborative navigation of the whole vehicle, which is used to solve the technical problems that it is difficult to take into account the complexity of different road areas, resulting in reasonable deterioration of navigation performance, and it is difficult to accurately analyze potential faults in the navigation system.

[0005] To solve the above problems, the first aspect of the present invention provides a comprehensive performance evaluation method and system for vehicle-road collaborative navigation, comprising the following steps: Real-time acquisition of vehicle operation data, road status data, vehicle-road cooperative communication data and weather data for vehicle-road cooperative navigation; The detection area is divided into grids and numbered, and several grid detection nodes are set in the grid to collect road overhead monitoring image data, weather data and vehicle operation data at different locations on the road; Analyze the complexity of the road environment through road overhead image data and weather data, and assign grid data with the same range of road environment complexity to the same grid data classification; According to the grid data that the vehicle has passed through and the classification of each grid, the vehicle data in the detection area is approximately associated; Establish data evaluation models for vehicle-road cooperative navigation at different grid data levels, and analyze vehicle navigation problems, road data delay problems, and communication delay problems of vehicle-road cooperative navigation in the grid; Based on the vehicle operation data, the vehicle decision similarity of the approximately associated vehicle data is evaluated, and the abnormal behavior of the vehicles in the grid data classification with the highest road environment complexity is analyzed. Based on the vehicle decision similarity of the approximately associated vehicle data and the vehicle abnormal behavior analysis results, the performance defect coefficient of vehicle-road collaborative navigation is evaluated.

[0006] Optionally, in an example of the above aspect, the vehicle operation data, road status data, vehicle-road cooperative communication data and weather data include; Vehicle operation data includes: vehicle real-time driving route, speed, route optimization times, NOA function on and off time; Road status data includes: road type, road lights, road obstacles and road congestion status; Vehicle-road cooperative communication data includes: communication delay and packet loss rate; Weather data includes: real-time weather data and visibility.

[0007] Optionally, in an example of the above aspect, collecting road bird's-eye view monitoring image data, environmental monitoring data and vehicle operation data at different road locations includes the following steps: The grid detection nodes collect road overhead monitoring image data at different locations on the road, and establish a connection with the authorized weather service center to obtain the latest weather data in real time; The grid detection node sends vehicle driving data transmission requests to surrounding vehicles to obtain vehicle operation data transmitted in real time by the authorized user's vehicle.

[0008] Optionally, in an example of the above aspect, analyzing the road environment complexity through road bird's-eye view image data and weather data, and assigning grid data belonging to the same range of road environment complexity to the same grid data classification includes the following steps: Obtain road bird's-eye view monitoring video data, and extract monitoring image data once every 100 frames as road bird's-eye view monitoring image data; By preprocessing the image, including denoising and contrast enhancement, and extracting the road overhead monitoring image through image recognition algorithm, the average speed of vehicles, vehicle density, pedestrian flow and road obstacle area range of each detection time period of the same video frame are obtained; The complexity of the road environment is analyzed by the average speed of vehicles, vehicle density, pedestrian flow, and area of ​​road obstacles, combined with the latest weather data obtained in real time, using the following formula: Wherein, F is the road environment complexity analysis value, N is the number of preset total detection time periods, Vmax, Omax, Lmax and Amax are the maximum detection values ​​of the average vehicle speed, vehicle density, pedestrian flow and road width in the total detection time period, Vi, Oi, Li and Ai are the average vehicle speed, vehicle density, pedestrian flow and obstacle-occupied road width detected in the i-th detection time period, λ is the attenuation coefficient, Di is the road visibility in the weather data in the i-th detection time period, Db is the preset standard value of road visibility on a sunny day; attenuation coefficient λ=0.05; Grid data with road environment complexity analysis values ​​belonging to the same range are allocated to the same grid data level.

[0009] Optionally, in an example of the above aspect, according to the grid data through which the vehicle passes and the level of each grid, approximately correlating the vehicle data of the detection area includes the following steps: Analyze the vehicle's current driving route data through vehicle-road collaborative navigation to obtain the grids crossed by the vehicle and their corresponding number data; Based on the grid data that the vehicle passes through and the level of each grid, the similarity between vehicles with the same number of grids crossed is calculated using the following formula: Where S is the similarity between vehicles, Ch is the average number of highest-level grids crossed by vehicles with equal number of grids crossed, C0 is the number of different grids crossed by the two vehicles, Ct is the number of the same grids crossed by the vehicles, C is the sum of all grids crossed by the two vehicles, and Fa-Fb is the difference in the road environment complexity analysis values ​​of the grids crossed by the two vehicles; Approximately associate the vehicle data within the detection area whose similarity between vehicles is greater than a threshold.

[0010] Optionally, in an example of the above aspect, data evaluation models for vehicle-road cooperative navigation of different grid data levels are established respectively, and vehicle navigation problems, road data delay problems, and communication delay problems of vehicle-road cooperative navigation of vehicles in the grid are analyzed, including the following steps: Obtain road overhead monitoring image data, weather data and vehicle operation data collected by grid detection nodes at different road locations, as well as vehicle-road collaborative navigation data collected within the grid; Obtain historical data collected by grid detection nodes at different grid data levels, including good user feedback, navigation problem and road problem feedback, and corresponding historical data of vehicle-road cooperative navigation, and add labels of good navigation status, vehicle navigation problem, road data delay problem and communication delay problem to the corresponding data. At the same time, add communication delay problem labels to the corresponding historical data when the communication delay and packet loss rate exceed the preset threshold; By adding labeled historical data of different grid data levels, different machine learning models are trained respectively, and a data evaluation model for vehicle-road collaborative navigation of different grid data levels is established to analyze whether the vehicle's vehicle-road collaborative navigation is in good navigation condition, or whether there are vehicle navigation problems, road data delay problems and communication delay problems.

[0011] Optionally, in an example of the above aspect, evaluating the vehicle decision similarity of approximately associated vehicle data according to the vehicle operation data comprises the following steps: Obtain the real-time driving route, speed and route optimization times of the vehicle in the vehicle operation data; Count the real-time driving route lengths of approximately associated vehicles, the average speed of vehicles in the driving routes, and the number of route optimizations; The ratio of the difference between the real-time driving route lengths of approximately associated vehicles and their sum, the ratio of the difference between the average speeds of vehicles in the driving routes and their sum, and the ratio of the difference between the number of route optimizations and their sum are calculated respectively, and the absolute value of each ratio is taken and weightedly added to obtain the vehicle decision similarity.

[0012] Optionally, in an example of the above aspect, analyzing abnormal behaviors of vehicles in the grid data hierarchy with the highest road environment complexity includes the following steps: The vehicle operation data corresponding to the vehicles in the grid data classification with the highest road environment complexity are analyzed, and abnormal vehicle behavior is defined as the vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimizations exceeding the preset threshold; The vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimization times exceeding the preset threshold are assigned values ​​respectively, with the assignment range being between 0.1 and 1. The sum of the vehicle abnormal behavior assignments is calculated as the abnormal behavior analysis value.

[0013] Optionally, in an example of the above aspect, evaluating the performance defect coefficient of the vehicle-road cooperative navigation according to the vehicle decision similarity of the approximately associated vehicle data and the vehicle abnormal behavior analysis result includes the following steps: The vehicle decision similarity values ​​and the vehicle abnormal behavior analysis values ​​of the approximately associated vehicle data are both normalized, and the normalized vehicle decision similarity values ​​and the vehicle abnormal behavior analysis values ​​are weightedly added, and the obtained value is used as the performance defect coefficient of the vehicle-road cooperative navigation; If the performance defect coefficient of the vehicle-road cooperative navigation exceeds a preset threshold, then the vehicle-road cooperative navigation of the corresponding vehicle has a performance defect, otherwise there is no performance defect.

[0014] According to another aspect of the present disclosure, a comprehensive performance evaluation system for vehicle-road collaborative navigation of a whole vehicle is provided. The system adopts a comprehensive performance evaluation method for vehicle-road collaborative navigation of a whole vehicle as described above to implement comprehensive performance evaluation of vehicle-road collaborative navigation of a whole vehicle.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention uses road overhead image data to intuitively obtain the actual conditions of the road, and combines weather data to facilitate further consideration of weather factors and the impact of reduced visibility on the road environment. By comprehensively analyzing these factors, the complexity of the road environment can be more accurately evaluated, providing a reliable basis for subsequent grid data classification. The road environment complexity is divided into different grid data classifications, and the grid data classification can clearly display the complexity of the road environment in different areas, which is convenient for the subsequent simplification of vehicle data processing. When a vehicle passes through a certain grid, the vehicle data can be quickly associated with the corresponding road environment complexity based on the classification information of the grid. This approximate association method avoids complex calculation processes and improves the efficiency of data processing.

[0016] The present invention uses a data evaluation model to analyze vehicle operation data, road status data, vehicle-road cooperative communication data and weather data, so as to facilitate real-time analysis of vehicle navigation problems, road data delay problems and communication delay problems during vehicle-road cooperative navigation. By dividing the road environment into different grids and assigning data classification to each grid, the vehicle navigation problems within each grid can be analyzed in more detail. This refined management method helps to quickly and accurately identify specific problems such as navigation deviation, untimely road data updates or communication delays, so as to take targeted solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention 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, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0019] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] See also Figure 1 The first embodiment of the present invention provides a comprehensive performance evaluation method and system for vehicle-road collaborative navigation, including the following steps: Real-time acquisition of vehicle operation data, road status data, vehicle-road cooperative communication data and weather data for vehicle-road cooperative navigation; The detection area is divided into grids and numbered, and several grid detection nodes are set in the grid to collect road overhead monitoring image data, weather data and vehicle operation data at different locations on the road; Analyze the complexity of the road environment through road overhead image data and weather data, and assign grid data with the same range of road environment complexity to the same grid data classification; According to the grid data that the vehicle has passed through and the classification of each grid, the vehicle data in the detection area is approximately associated; Establish data evaluation models for vehicle-road cooperative navigation at different grid data levels, and analyze vehicle navigation problems, road data delay problems, and communication delay problems of vehicle-road cooperative navigation in the grid; Based on the vehicle operation data, the vehicle decision similarity of the approximately associated vehicle data is evaluated, and the abnormal behavior of the vehicles in the grid data classification with the highest road environment complexity is analyzed. Based on the vehicle decision similarity of the approximately associated vehicle data and the vehicle abnormal behavior analysis results, the performance defect coefficient of vehicle-road collaborative navigation is evaluated.

[0021] In this embodiment, the detection area is divided into grids and numbered, and the grid division standard is set: Based on the average braking distance on urban expressways, the grid is divided into 200m×200m; based on the intersection spacing on ordinary urban roads, the grid is divided into 100m×100m; on expressways, the grid is divided into 300m×300m; in complex interchange areas, the grid is divided into 50m×50m; and the grids are numbered, such as: Urban Expressway-0001.

[0022] Several grid detection nodes are set in the grid, road overhead monitoring cameras are set, and each grid detection node is connected to the authorized weather service center to collect road overhead monitoring image data, weather data and vehicle operation data at different locations on the road; The road environment complexity is analyzed through road overhead image data and weather data, and grid data belonging to the same range of road environment complexity are assigned to the same grid data level; vehicle data in the detection area are approximately associated according to the grid data passed by the vehicle and the level of each grid; The use of road overhead image data makes it easy to intuitively obtain the actual conditions of the road, such as traffic volume, road conditions and other information. Combined with weather data, it is convenient to further consider weather factors and the impact of reduced visibility on the road environment. By comprehensively analyzing these factors, the complexity of the road environment can be more accurately assessed, providing a reliable basis for subsequent grid data classification.

[0023] The complexity of the road environment is divided into different grid data levels. The grid data levels can clearly show the complexity of the road environment in different areas, which is convenient for the subsequent simplification of the vehicle data processing process.

[0024] When a vehicle passes through a grid, the vehicle data can be quickly associated with the corresponding road environment complexity based on the grid's classification information. This approximate association method avoids complex calculations and improves data processing efficiency.

[0025] Establish data evaluation models for vehicle-road cooperative navigation at different grid data levels, and analyze vehicle navigation problems, road data delay problems, and communication delay problems of vehicle-road cooperative navigation in the grid; Through the data evaluation model, vehicle operation data, road status data, vehicle-road cooperative communication data and weather data are analyzed, which facilitates real-time analysis of vehicle navigation problems, road data delay problems and communication delay problems during the vehicle-road cooperative navigation process.

[0026] By dividing the road environment into different grids and assigning data levels to each grid, the vehicle navigation problems within each grid can be analyzed in more detail. This refined management method helps to quickly and accurately identify specific problems such as navigation deviations, untimely road data updates, or communication delays, so that targeted solutions can be taken. The data evaluation model can be based on grid data classification. By continuously monitoring and analyzing vehicle navigation problems, road data delays, and communication delays within the grid, the data evaluation model can promptly detect and warn of potential system failures or performance degradation trends in the vehicle-road cooperative navigation of vehicles in different hierarchical grids.

[0027] Based on the vehicle operation data, the vehicle decision similarity of the approximately associated vehicle data is evaluated, and the abnormal behavior of the vehicles in the grid data classification with the highest road environment complexity is analyzed. Based on the vehicle decision similarity of the approximately associated vehicle data and the vehicle abnormal behavior analysis results, the performance defect coefficient of vehicle-road collaborative navigation is evaluated.

[0028] By evaluating the vehicle decision similarity of approximately related vehicle data based on vehicle operation data, the decision consistency of vehicles during navigation can be quantified more accurately. At the same time, in-depth analysis of abnormal vehicle behavior in the grid data classification with the highest road environment complexity can accurately identify potential problems of vehicle navigation systems in highly complex environments. Compared with a single indicator or simple data statistics, this comprehensive evaluation method can significantly improve the accuracy of identifying defects in vehicle-road collaborative navigation performance.

[0029] By continuously monitoring and analyzing vehicle decision similarities and abnormal vehicle behaviors, potential system failures or performance degradation trends can be discovered and warned in a timely manner. This helps to take preventive or repair measures in advance. For grid areas or vehicles that frequently have decision deviations or abnormal behaviors, inspections can be prioritized to ensure that navigation services and communication quality in key areas are prioritized.

[0030] In one embodiment of the present invention, the vehicle operation data, road status data, vehicle-road cooperative communication data and weather data include: Vehicle operation data includes: vehicle real-time driving route, speed, route optimization times, NOA function on and off time; Road status data includes: road type, road lights, road obstacles and road congestion status; Vehicle-road cooperative communication data includes: communication delay and packet loss rate; Weather data includes: real-time weather data and visibility.

[0031] In one embodiment of the present invention, collecting road overhead monitoring image data, environmental monitoring data and vehicle operation data at different positions of the road includes the following steps: The grid detection nodes collect road overhead monitoring image data at different locations on the road, and establish a connection with the authorized weather service center to obtain the latest weather data in real time; The grid detection node sends vehicle driving data transmission requests to surrounding vehicles to obtain vehicle operation data transmitted in real time by the authorized user's vehicle.

[0032] In one embodiment of the present invention, the complexity of the road environment is analyzed by using road overhead image data and weather data, and grid data belonging to the same range of road environment complexity are assigned to the same grid data classification, including the following steps: Obtain road bird's-eye view monitoring video data, and extract monitoring image data once every 100 frames as road bird's-eye view monitoring image data; By preprocessing the image, including denoising and contrast enhancement, and extracting the road overhead monitoring image through image recognition algorithm, the average speed of vehicles, vehicle density, pedestrian flow and road obstacle area range of each detection time period of the same video frame are obtained; The complexity of the road environment is analyzed by the average speed of vehicles, vehicle density, pedestrian flow, and area of ​​road obstacles, combined with the latest weather data obtained in real time, using the following formula: Wherein, F is the road environment complexity analysis value, N is the number of preset total detection time periods, Vmax, Omax, Lmax and Amax are the maximum detection values ​​of the average vehicle speed, vehicle density, pedestrian flow and road width in the total detection time period, Vi, Oi, Li and Ai are the average vehicle speed, vehicle density, pedestrian flow and obstacle-occupied road width detected in the i-th detection time period, λ is the attenuation coefficient, Di is the road visibility in the weather data in the i-th detection time period, Db is the preset standard value of road visibility on a sunny day; attenuation coefficient λ=0.05; Grid data with road environment complexity analysis values ​​belonging to the same range are allocated to the same grid data level.

[0033] In this embodiment, the grid data is classified into five levels according to the road environment complexity analysis value: Level 1: CCI < 0.6; Level 2: 0.6≤CCI<1.1; Level 3: 1.1≤CCI<2.2; Level 4: 2.2≤CCI<2.9; Level 5: CCI ≥ 2.9.

[0034] In one embodiment of the present invention, according to the grid data through which the vehicle passes and the level of each grid, the vehicle data of the detection area is approximately associated, including the following steps: Analyze the vehicle's current driving route data through vehicle-road collaborative navigation to obtain the grids crossed by the vehicle and their corresponding number data; Based on the grid data that the vehicle passes through and the level of each grid, the similarity between vehicles with the same number of grids crossed is calculated using the following formula: Where S is the similarity between vehicles, Ch is the average number of highest-level grids crossed by vehicles with equal number of grids crossed, C0 is the number of different grids crossed by the two vehicles, Ct is the number of the same grids crossed by the vehicles, C is the sum of all grids crossed by the two vehicles, and Fa-Fb is the difference in the road environment complexity analysis values ​​of the grids crossed by the two vehicles; Approximately associate the vehicle data within the detection area whose similarity between vehicles is greater than a threshold.

[0035] In this embodiment, the driving route data of a large number of vehicles with close starting and ending positions are collected, and the similarities between the corresponding vehicles are counted, and the mean of the statistical similarities is calculated as the threshold. Based on this, in this embodiment, if the similarity S between vehicles with an equal number of grids crossed by the vehicles is greater than the threshold value 0.32, the two vehicles are judged to have similar driving data, and the data of the corresponding vehicles are approximately associated.

[0036] In one embodiment of the present invention, data evaluation models for vehicle-road cooperative navigation of different grid data levels are established respectively, and vehicle navigation problems, road data delay problems, and communication delay problems of vehicle-road cooperative navigation of vehicles in the grid are analyzed, including the following steps: Obtain road overhead monitoring image data, weather data and vehicle operation data collected by grid detection nodes at different road locations, as well as vehicle-road collaborative navigation data collected within the grid; Obtain historical data collected by grid detection nodes at different grid data levels, including good user feedback, navigation problem and road problem feedback, and corresponding historical data of vehicle-road cooperative navigation, and add labels of good navigation status, vehicle navigation problem, road data delay problem and communication delay problem to the corresponding data. At the same time, add communication delay problem labels to the corresponding historical data when the communication delay and packet loss rate exceed the preset threshold; By adding labeled historical data of different grid data levels, different machine learning models are trained respectively, and a data evaluation model for vehicle-road collaborative navigation of different grid data levels is established to analyze whether the vehicle's vehicle-road collaborative navigation is in good navigation condition, or whether there are vehicle navigation problems, road data delay problems and communication delay problems.

[0037] In one embodiment of the present invention, evaluating the vehicle decision similarity of approximately associated vehicle data based on vehicle operation data includes the following steps: Obtain the real-time driving route, speed and route optimization times of the vehicle in the vehicle operation data; Count the real-time driving route lengths of approximately associated vehicles, the average speed of vehicles in the driving routes, and the number of route optimizations; The ratio of the difference between the real-time driving route lengths of approximately associated vehicles and their sum, the ratio of the difference between the average speeds of vehicles in the driving routes and their sum, and the ratio of the difference between the number of route optimizations and their sum are calculated respectively, and the absolute value of each ratio is taken and weightedly added to obtain the vehicle decision similarity.

[0038] In this embodiment, the absolute values ​​of the ratio of the difference between the lengths of the real-time driving routes of the vehicles and their sum, and the ratio of the difference between the average speeds of the vehicles in the driving routes and their sum are weighted and added, and the weights are 0.4, 0.3 and 0.3 respectively.

[0039] In one embodiment of the present invention, analyzing abnormal behaviors of vehicles in the grid data hierarchy with the highest road environment complexity includes the following steps: The vehicle operation data corresponding to the vehicles in the grid data classification with the highest road environment complexity are analyzed, and abnormal vehicle behavior is defined as the vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimizations exceeding the preset threshold; The vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimization times exceeding the preset threshold are assigned values ​​respectively, with the assignment range being between 0.1 and 1. The sum of the vehicle abnormal behavior assignments is calculated as the abnormal behavior analysis value.

[0040] In this embodiment, the vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimization times exceeding the preset threshold are assigned values ​​of: 0.5, 0.3, 0.6, and 0.5, respectively.

[0041] In one embodiment of the present invention, the performance defect coefficient of vehicle-road cooperative navigation is evaluated based on the vehicle decision similarity of approximately associated vehicle data and the vehicle abnormal behavior analysis result, including the following steps: The vehicle decision similarity values ​​and the vehicle abnormal behavior analysis values ​​of the approximately associated vehicle data are both normalized, and the normalized vehicle decision similarity values ​​and the vehicle abnormal behavior analysis values ​​are weightedly added, and the obtained value is used as the performance defect coefficient of the vehicle-road cooperative navigation; If the performance defect coefficient of the vehicle-road cooperative navigation exceeds a preset threshold, then the vehicle-road cooperative navigation of the corresponding vehicle has a performance defect, otherwise there is no performance defect.

[0042] In this embodiment, the normalized vehicle decision similarity value and the vehicle abnormal behavior analysis value are weighted and added, and the weights are 0.4 and 0.6 respectively; By counting the performance defect coefficient values ​​of the vehicle-road collaborative navigation of a large number of vehicles and the approximately associated vehicle data when there are navigation failures, the minimum value is used as the preset threshold value of the performance defect coefficient of the vehicle-road collaborative navigation; in this embodiment, if the performance defect coefficient of the vehicle-road collaborative navigation exceeds 0.42, the vehicle-road collaborative navigation of the corresponding vehicle has a performance defect, otherwise there is no performance defect.

[0043] According to the above-mentioned embodiments of the present application, a comprehensive performance evaluation system for vehicle-road collaborative navigation of the whole vehicle is also provided. The system adopts a comprehensive performance evaluation method for vehicle-road collaborative navigation of the whole vehicle as mentioned above to realize the comprehensive performance evaluation of vehicle-road collaborative navigation of the whole vehicle.

[0044] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A comprehensive performance evaluation method for vehicle-road collaborative navigation, characterized in that: The following steps are involved: Real-time acquisition of vehicle operation data, road status data, vehicle-road cooperative communication data and weather data for vehicle-road cooperative navigation; The detection area is divided into grids and numbered, and several grid detection nodes are set in the grid to collect road overhead monitoring image data, weather data and vehicle operation data at different locations on the road; Analyze the complexity of the road environment through road overhead image data and weather data, and assign grid data with the same range of road environment complexity to the same grid data classification; According to the grid data that the vehicle has passed through and the classification of each grid, the vehicle data in the detection area is approximately associated; Establish data evaluation models for vehicle-road cooperative navigation at different grid data levels, and analyze vehicle navigation problems, road data delay problems, and communication delay problems of vehicle-road cooperative navigation in the grid; Based on the vehicle operation data, the vehicle decision similarity of the approximately associated vehicle data is evaluated, and the abnormal behavior of the vehicles in the grid data classification with the highest road environment complexity is analyzed. Based on the vehicle decision similarity of the approximately associated vehicle data and the vehicle abnormal behavior analysis results, the performance defect coefficient of vehicle-road collaborative navigation is evaluated.

2. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: Vehicle operation data, road status data, vehicle-road cooperative communication data and weather data, including; Vehicle operation data includes: vehicle real-time driving route, speed, route optimization times, NOA function on and off time; Road status data includes: road type, road lights, road obstacles and road congestion status; Vehicle-road cooperative communication data includes: communication delay and packet loss rate; Weather data includes: real-time weather data and visibility.

3. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: Collecting road overhead monitoring image data, environmental monitoring data and vehicle operation data at different road locations includes the following steps: The grid detection nodes collect road overhead monitoring image data at different locations on the road, and establish a connection with the authorized weather service center to obtain the latest weather data in real time; The grid detection node sends vehicle driving data transmission requests to surrounding vehicles to obtain vehicle operation data transmitted in real time by the authorized user's vehicle.

4. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: The road environment complexity is analyzed by using road overhead image data and weather data, and grid data belonging to the same range of road environment complexity are assigned to the same grid data classification, including the following steps: Obtain road bird's-eye view monitoring video data, and extract monitoring image data once every 100 frames as road bird's-eye view monitoring image data; By preprocessing the image, including denoising and contrast enhancement, and extracting the road overhead monitoring image through image recognition algorithm, the average speed of vehicles, vehicle density, pedestrian flow and road obstacle area range of each detection time period of the same video frame are obtained; The complexity of the road environment is analyzed by the average speed of vehicles, vehicle density, pedestrian flow, and area of ​​road obstacles, combined with the latest weather data obtained in real time, using the following formula: Wherein, F is the road environment complexity analysis value, N is the number of preset total detection time periods, Vmax, Omax, Lmax and Amax are the maximum detection values ​​of the average vehicle speed, vehicle density, pedestrian flow and road width in the total detection time period, Vi, Oi, Li and Ai are the average vehicle speed, vehicle density, pedestrian flow and obstacle-occupied road width detected in the i-th detection time period, λ is the attenuation coefficient, Di is the road visibility in the weather data in the i-th detection time period, and Db is the preset standard value of road visibility on a sunny day; Grid data with road environment complexity analysis values ​​belonging to the same range are allocated to the same grid data level.

5. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 4 is characterized in that: According to the grid data that the vehicle has passed through and the classification of each grid, the vehicle data in the detection area is approximately associated, including the following steps: Analyze the vehicle's current driving route data through vehicle-road collaborative navigation to obtain the grids crossed by the vehicle and their corresponding number data; Based on the grid data that the vehicle passes through and the level of each grid, the similarity between vehicles with the same number of grids crossed is calculated using the following formula: Where S is the similarity between vehicles, Ch is the average number of highest-level grids crossed by vehicles with equal number of grids crossed, C0 is the number of different grids crossed by the two vehicles, Ct is the number of the same grids crossed by the vehicles, C is the sum of all grids crossed by the two vehicles, and Fa-Fb is the difference in the road environment complexity analysis values ​​of the grids crossed by the two vehicles; Approximately associate the vehicle data within the detection area whose similarity between vehicles is greater than a threshold.

6. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: The data evaluation models of vehicle-road cooperative navigation with different grid data levels are established respectively, and the vehicle navigation problem, road data delay problem and communication delay problem of vehicle-road cooperative navigation of vehicles in the grid are analyzed, including the following steps: Obtain road overhead monitoring image data, weather data and vehicle operation data collected by grid detection nodes at different road locations, as well as vehicle-road collaborative navigation data collected within the grid; Obtain historical data collected by grid detection nodes at different grid data levels, including good user feedback, navigation problem and road problem feedback, and corresponding historical data of vehicle-road cooperative navigation, and add labels of good navigation status, vehicle navigation problem, road data delay problem and communication delay problem to the corresponding data. At the same time, add communication delay problem labels to the corresponding historical data when the communication delay and packet loss rate exceed the preset threshold; By adding labeled historical data of different grid data levels, different machine learning models are trained respectively, and a data evaluation model for vehicle-road collaborative navigation of different grid data levels is established to analyze whether the vehicle's vehicle-road collaborative navigation is in good navigation condition, or whether there are vehicle navigation problems, road data delay problems and communication delay problems.

7. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: Based on the vehicle operation data, the vehicle decision similarity of approximately associated vehicle data is evaluated, including the following steps: Obtain the real-time driving route, speed and route optimization times of the vehicle in the vehicle operation data; Count the real-time driving route lengths of approximately associated vehicles, the average speed of vehicles in the driving routes, and the number of route optimizations; The ratio of the difference between the real-time driving route lengths of approximately associated vehicles and their sum, the ratio of the difference between the average speeds of vehicles in the driving routes and their sum, and the ratio of the difference between the number of route optimizations and their sum are calculated respectively, and the absolute value of each ratio is taken and weightedly added to obtain the vehicle decision similarity.

8. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: The abnormal behavior of vehicles in the grid data level with the highest road environment complexity is analyzed, including the following steps: The vehicle operation data corresponding to the vehicles in the grid data classification with the highest road environment complexity are analyzed, and abnormal vehicle behavior is defined as the vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimizations exceeding the preset threshold; The vehicle stationary time exceeding the preset threshold, the number of vehicle lane changes, the total time consumed by the vehicle route exceeding the preset threshold, and the number of route optimization times exceeding the preset threshold are assigned values ​​respectively, with the assignment range being between 0.1 and 1. The sum of the vehicle abnormal behavior assignments is calculated as the abnormal behavior analysis value.

9. The comprehensive performance evaluation method for vehicle-road collaborative navigation according to claim 1 is characterized in that: Based on the vehicle decision similarity and vehicle abnormal behavior analysis results of the approximately associated vehicle data, the performance defect coefficient of the vehicle-road cooperative navigation is evaluated, including the following steps: The vehicle decision similarity values ​​and the vehicle abnormal behavior analysis values ​​of the approximately associated vehicle data are both normalized, and the normalized vehicle decision similarity values ​​and the vehicle abnormal behavior analysis values ​​are weightedly added, and the obtained value is used as the performance defect coefficient of the vehicle-road cooperative navigation; If the performance defect coefficient of the vehicle-road cooperative navigation exceeds a preset threshold, then the vehicle-road cooperative navigation of the corresponding vehicle has a performance defect, otherwise there is no performance defect.

10. A comprehensive performance evaluation system for vehicle-road collaborative navigation, characterized in that: The system adopts a comprehensive performance evaluation method for vehicle-road collaborative navigation of a whole vehicle as described in any one of claims 1-9 to realize comprehensive performance evaluation of vehicle-road collaborative navigation of the whole vehicle.

Images