Detection method and system of traveling data recorder and storage medium

Abstract

The invention relates to the field of data testing, in particular to a detection method and system of a traveling data recorder and a storage medium. The method comprises the following steps: constructing various driving environments; the method comprises the following steps: simulating a driving state in various driving environments to detect recorded data and an audio and video recording function of a traveling data recorder to obtain the recorded data of the traveling data recorder; automatically analyzing the recorded data of the traveling data recorder to obtain recorded detection data; performing communication function detection; detecting a driving behavior identification function and a driving safety prompting function of the traveling data recorder; detecting the safety early warning response speed and accuracy of the recorder; and based on a big data technology, collecting and analyzing various detection data to obtain a final detection result. According to the invention, the performance of the traveling data recorder can be evaluated comprehensively. And secondly, by constructing various driving environments, the performance of the traveling data recorder can be reflected more truly by simulating driving behaviors, and the obtained detection result is more suitable for the actual situation, so that the detection accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the field of data testing, and more specifically, to a detection method, system and storage medium of a driving recorder. Background Art

[0002] The driving recorder is a widely used device in real life. It accurately records the driving conditions of the car and is an important basis for accident judgment, driving behavior management and other scenarios. In recent years, with the advancement of science and technology and the trend of intelligence, the driving recorder no longer has only a single recording function, but also adds intelligent functions such as navigation, safety warning, audio and video recording, and data analysis. With the development of science and technology, the functions of driving recorders are increasing, such as vehicle status monitoring, navigation, voice recognition, audio and video recording, vehicle safety warning, etc. Therefore, the detection of driving recorders is particularly important to ensure their quality and accuracy to meet daily use needs. However, because the driving recorder involves many types of data, including video data, voice data, car driving parameters, etc., in the process of designing and manufacturing driving recorders, there may be problems in the data accuracy, reliability, stability and other aspects of the driving recorder in actual use. Therefore, the detection and evaluation of driving recorders is crucial. Traditional detection methods often detect a single parameter and cannot comprehensively evaluate the performance of the driving recorder. Furthermore, due to the limitations of the detection environment, it is impossible to simulate the complex and changeable driving environment in the real world.

[0003] The prior art discloses a functional testing device and method for a vehicle driving recorder, wherein the device includes an industrial computer, an electrical control cabinet, and a test fixture. The industrial computer is installed with a host computer software, and the electrical control cabinet is used to receive the test instructions sent by the host computer software, perform full-function testing on the driving recorder according to the test instructions, and receive feedback information from the driving recorder and send it to the industrial computer. The industrial computer receives the feedback information and determines the test result through the host computer software. However, this method only performs the test through the test instructions sent by the host computer software, which cannot well reflect the function of the driving recorder in the actual driving situation, resulting in the accuracy of the test being not high enough. Summary of the invention

[0004] The purpose of the present invention is to disclose a detection method, system and storage medium of a driving recorder with better detection effect.

[0005] In order to achieve the above object, the present invention provides a detection method of a driving recorder, comprising:

[0006] S1: Construct multiple driving environments;

[0007] S2: simulating driving conditions in various driving environments to detect the recorded data and audio and video recording functions of the driving recorder to obtain the recorded data of the driving recorder; automatically analyzing the recorded data of the driving recorder to obtain recorded detection data;

[0008] S3: Perform communication function detection in various driving environments to obtain dynamic detection data of wireless communication capability;

[0009] S4: Simulate various driving behaviors in various driving environments, and detect the driving behavior recognition and driving safety prompt functions of the driving recorder to obtain behavior-based driving detection data;

[0010] S5: Testing the safety warning reaction speed and accuracy of the recorder when simulating an accident in various driving environments to obtain intelligent safety warning detection data;

[0011] S6: Based on big data technology, collect and analyze record detection data, wireless communication capability dynamic detection data, driving detection data and intelligent safety warning detection data to obtain the final detection results.

[0012] Furthermore, the multiple driving environments include: at least one driving environment of rainy and snowy weather, driving at night, and urban congestion environment.

[0013] Furthermore, the construction of various driving environments includes:

[0014] Define parameterized driving environment characteristics, including: road type, weather conditions, lighting conditions, vehicle status and other traffic factors;

[0015] Generate a simulation environment based on the specific conditions to be tested:

[0016] When simulating the situation of urban roads, the road is divided into multiple sections, and the length, driving speed, and number of vehicles ahead of each section are generated by the given parameter generation model;

[0017] The driving environment is described by the parameter vector P

[0018] P=[p1,p2,...,pn]

[0019] Among them, pi represents the i-th environmental parameter;

[0020] According to the parameter vector P, the corresponding simulation environment is generated; specifically, each environment parameter is generated by a generating function fi:

[0021] pi=fi(P')

[0022] Among them, P' is a set of input parameters, which are random variables.

[0023] Furthermore, the automatic analysis of the records of the driving recorder includes:

[0024] Using graph neural networks (GNNs) to analyze driving recorder data:

[0025] Build a graph model, consider vehicles and roads as nodes, and the distance and direction between vehicles and roads as edges, so as to construct a traffic network graph;

[0026] Feature extraction: For each node, a feature vector is defined to express the node status, including the speed and acceleration of the vehicle, the length, width and speed limit of the road. The graph network will automatically propagate and integrate information, so that each node contains not only its own information, but also the information of adjacent nodes. Feature extraction is expressed as a function H, where H(v) represents the feature vector of node v.

[0027] Graph network training: training is performed based on real traffic data, including traffic flow and traffic accidents, to obtain a trained graph model;

[0028] Based on the trained graph model, the traffic data and conditions in the future are predicted, and the predicted traffic data and conditions in the future are compared with the traffic data and conditions of the simulated driving state to obtain the recorded detection data.

[0029] Furthermore, communication function detection is performed in various driving environments to obtain dynamic detection data of wireless communication capability including:

[0030] Define the measurement parameter set Q = {moving speed V, communication distance D, network bandwidth B, network delay L}, where V, D, B, L represent moving speed, communication distance, network bandwidth and network delay respectively;

[0031] Assume that the data transmission stability of the USB, Bluetooth, and wireless public network of the driving recorder is U, B, and W respectively, and is measured by the functions f(Q)=U, g(Q)=B, and h(Q=W;

[0032] In the simulated mobile environment, set a specific mobile speed V1, communication distance D1, network bandwidth B1, and network delay L1, that is, set the parameter set Q1 = {V1, D1, B1, L1}, and then measure the data transmission stability U1, B1, W1 at this time according to f(Q1), g(Q1), and h(Q1);

[0033] Then, change the mobile environment conditions, set a new parameter set Q2 = {V2, D1, B1, L1}, V2> V1, and measure U2, B2, and W2 at this time;

[0034] According to the above steps, multiple groups of different parameter sets are obtained respectively. By measuring the data transmission stability under each group of conditions, a series of data transmission stability and efficiency values ​​are obtained: (U1, B1, W1), (U2, B2, W2), ..., (Un, Bn, Wn); the series of data transmission stability and efficiency values ​​are used as dynamic detection data of wireless communication capabilities.

[0035] Furthermore, various driving behaviors are simulated in various driving environments, and the driving behavior recognition and driving safety prompt functions of the driving recorder are tested to obtain behavior-based driving test data:

[0036] Define a driving behavior description set A = {a1, a2, ..., an}, where a1, a2, ..., an are parameters related to driving behavior such as speed, steering angle, braking force, and relative position of the vehicle; and define a driving behavior recognition result set R = {r1, r2, ..., rn}, where r1, r2, ..., rn represent the recognized driving behavior types, including normal driving, sudden acceleration, and sudden braking;

[0037] The driving behavior recognition function of the driving recorder is defined as:

[0038] R = f(A),

[0039] Where f is the internal algorithm of the driving recorder, which inputs the driving behavior description set A and outputs the identified driving behavior type set R;

[0040] Similarly, the driving safety reminder function of the driving recorder is defined as a function g, the input is the identified driving behavior type set R, and the output is the safety reminder set S, that is,

[0041] S = g(R);

[0042] In simulating various driving behaviors, a series of driving behavior description sets A are first generated in the driving environment, including normal driving situations and various abnormal driving behaviors, and then input into the driving recorder, and the driving behavior recognition result set R is obtained through f(A), and the safety prompt set S is further obtained through g(R);

[0043] For each driving behavior description set A, there is an expected driving behavior recognition result set R' and an expected safety prompt set S'. The actual R and S are compared with the expected R' and S' to obtain the driving behavior recognition accuracy and driving safety prompt accuracy of the driving recorder; the driving behavior recognition accuracy and driving safety prompt accuracy of the driving recorder are used as behavior-based driving detection data.

[0044] Furthermore, the safety warning reaction speed and accuracy of the recorder are tested in various driving environments when simulating accidents, and intelligent safety warning detection data is obtained:

[0045] Define a set of simulated driving environment parameters E = {e1, e2, ..., ei}, where ei represents the i-th driving environment parameter, including: vehicle speed, vehicle distance, and road condition;

[0046] Define the reaction speed of intelligent safety warning as V, and the reaction accuracy as A. These two parameters are calculated by functions f(E) and g(E): V = f(E), A = g(E);

[0047] During the test, the parameter set E of the simulation environment is first set as a specific accident environment. If the distance between vehicles is too close, e1 = vehicle distance, then it is set as too close distance. If the speed is too fast, e2 = vehicle speed, then it is set as too fast speed.

[0048] Calculate the reaction speed V and reaction accuracy A, that is, V1 = f(E), A1 = g(E);

[0049] Then change the parameters of the simulation environment and obtain a series of V and A according to the above steps;

[0050] This series of reaction speeds and reaction accuracies are summarized and recorded as (V1, A1), (V2, A2), ..., (Vn, An); and the values ​​of V and A are statistically analyzed, and the results are used as intelligent safety early warning detection data.

[0051] Furthermore, based on big data technology, the final test results obtained by collecting and analyzing the recorded test data, wireless communication capability dynamic test data, driving test data and intelligent safety warning test data include:

[0052] The data collected during all tests are defined as data set D, which includes driving environment parameters, driving recorder parameters, and various performance test results, namely

[0053] D={E,A,R,S,U,B,W,V1,A1,V2,A2,...,Vn,An};

[0054] Wherein, E = {e1, e2, ..., en} represents the parameters of various driving environments; A = {a1, a2, ..., an} represents the parameters of driving actions; R = {r1, r2, ..., rn} represents the type of driving behavior identified; S = {s1, s2, ..., sn} represents the safety prompts of the driving recorder; U, B, W represent the data transmission stability and efficiency of USB, Bluetooth, and wireless public networks respectively; V1, A1, V2, A2, ..., Vn, An represent the safety warning response speed and accuracy of the intelligent warning system in various driving environments respectively;

[0055] Use big data technology to preprocess data, remove outliers and invalid data, and perform data cleaning:

[0056] The outliers Z and invalid data I are defined by the following formulas;

[0057] Z = x|x∈D, and the value of x is outside its maximum and minimum range,

[0058] I=x|x∈D, and the value of x is missing or unreasonable,

[0059] The excluded data set D' is defined as:

[0060] D' = DZI;

[0061] Get the cleaned data set D',

[0062] D'={E',A',R',S',U',B',W',V1',A1',V2',A2',...,Vn',An'};

[0063] For the data set D', a variety of statistical analysis and machine learning methods are used for in-depth analysis or descriptive statistical analysis is used to analyze it, including calculating the mean μ, median θ and standard deviation σ of each parameter. The formula is as follows:

[0064]

[0065] Where N is the number of samples, xi is the i-th sample value;

[0066] Correlation analysis is used to describe the correlation between various parameters, and Pearson correlation coefficient r is used for correlation analysis. The calculation method is as follows:

[0067]

[0068] Among them, xi and yi represent the x and y values ​​of the i-th sample, μx, μy are the average values ​​of x and y, is the standard deviation of x and y;

[0069] For key performance indicators such as recognition accuracy and warning accuracy, the machine learning model SVM is used for prediction and optimization; the driving behavior recognition accuracy P is provided, and the prediction function p(X) = a*X+b is defined, where X is the parameter set, and a and b are the coefficients to be solved; the parameters of the prediction function are optimized by minimizing the error between the predicted value and the actual value. The error function E(a,b) is defined as:

[0070] E(a,b)=Σ[p(Xi)-Pi] 2

[0071] Among them, Pi is the actual driving behavior recognition accuracy, and p(Xi) is the predicted value;

[0072] Cluster analysis is used to perform data mining to find the underlying patterns of data. A commonly used indicator for evaluating clustering results is the silhouette coefficient S, which is calculated as follows:

[0073] S=(ba) / max(a,b)

[0074] Among them, a represents the average distance between the sample and samples of the same category, and b represents the minimum average distance between the sample and samples of other categories.

[0075] In addition, the present invention also provides a driving recorder detection system, comprising:

[0076] Environment construction module: build various driving environments;

[0077] Recording detection module: simulates driving status in various driving environments to detect the recording data and audio and video recording functions of the driving recorder, and obtains the recording data of the driving recorder; automatically analyzes the recording data of the driving recorder to obtain the recording detection data;

[0078] Wireless communication capability dynamic detection module: performs communication function detection in various driving environments to obtain dynamic detection data of wireless communication capability;

[0079] Driving detection module: simulates various driving behaviors in various driving environments, detects the driving behavior recognition and driving safety prompt functions of the driving recorder, and obtains behavior-based driving detection data;

[0080] Intelligent safety warning detection module: Test the safety warning response speed and accuracy of the recorder when simulating an accident in a variety of driving environments to obtain intelligent safety warning detection data;

[0081] Final detection module: Based on big data technology, it collects and analyzes recorded detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data to obtain the final detection results.

[0082] In addition, the present invention also provides a driving recorder detection storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned driving recorder detection method is implemented.

[0083] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0084] The present invention conducts a comprehensive test on the various functions of the driving recorder by constructing various driving environments, and obtains recording detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data, which can comprehensively evaluate the performance of the driving recorder. Secondly, by constructing various driving environments, simulating driving behaviors can more realistically reflect the performance of the driving recorder, and the obtained detection results are more in line with the actual situation, thereby improving the detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 This is a flow chart of a detection method of a driving recorder according to Embodiment 1;

[0086] Figure 2 This is the final detection flow chart described in Example 2;

[0087] Figure 3 A block diagram of a driving recorder detection system according to Embodiment 3; DETAILED DESCRIPTION

[0088] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0089] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0090] Embodiment 1:

[0091] This embodiment provides Figure 1 A detection method of a driving recorder shown includes:

[0092] Build multiple driving environments;

[0093] Simulate driving conditions in various driving environments to detect the recorded data and audio and video recording functions of the driving recorder to obtain the recorded data of the driving recorder; automatically analyze the recorded data of the driving recorder to obtain the recorded detection data;

[0094] Conduct communication function tests in various driving environments to obtain dynamic test data of wireless communication capabilities;

[0095] Simulate various driving behaviors in various driving environments, test the driving behavior recognition and driving safety prompt functions of the driving recorder to obtain behavior-based driving test data;

[0096] Test the safety warning response speed and accuracy of the recorder when simulating accidents in various driving environments to obtain intelligent safety warning detection data;

[0097] Based on big data technology, the final test results are obtained by collecting and analyzing recorded test data, dynamic test data on wireless communication capabilities, driving test data, and intelligent safety warning test data.

[0098] This embodiment conducts a comprehensive test on the various functions of the driving recorder by constructing various driving environments, and obtains recording detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data, which can comprehensively evaluate the performance of the driving recorder. Secondly, by constructing various driving environments and simulating driving behaviors, the performance of the driving recorder can be more realistically reflected, and the detection results obtained are more in line with the actual situation, thereby improving the detection effect.

[0099] Embodiment 2:

[0100] This embodiment further discloses on the basis of the first embodiment:

[0101] The multiple driving environments include: at least one driving environment in rainy and snowy weather, driving at night, and a congested city environment.

[0102] The construction of various driving environments includes:

[0103] Define parameterized driving environment characteristics, including: road type, weather conditions, lighting conditions, vehicle status and other traffic factors;

[0104] Generate a simulation environment based on the specific conditions to be tested:

[0105] When simulating the situation of urban roads, the road is divided into multiple sections, and the length, driving speed, and number of vehicles ahead of each section are generated by the given parameter generation model;

[0106] The driving environment is described by the parameter vector P

[0107] P=[p1,p2,...,pn]

[0108] Among them, pi represents the i-th environmental parameter;

[0109] According to the parameter vector P, the corresponding simulation environment is generated; specifically, each environment parameter is generated by a generating function fi:

[0110] pi=fi(P')

[0111] Among them, P' is a set of input parameters, which are random variables.

[0112] The automatic analysis of the records of the driving recorder includes:

[0113] Using graph neural networks (GNNs) to analyze driving recorder data:

[0114] Build a graph model, consider vehicles and roads as nodes, and the distance and direction between vehicles and roads as edges, so as to construct a traffic network graph;

[0115] Feature extraction: For each node, a feature vector is defined to express the node status, including the speed and acceleration of the vehicle, the length, width and speed limit of the road. The graph network will automatically propagate and integrate information, so that each node contains not only its own information, but also the information of adjacent nodes. Feature extraction is expressed as a function H, where H(v) represents the feature vector of node v.

[0116] Graph network training: training is performed based on real traffic data, including traffic flow and traffic accidents, to obtain a trained graph model;

[0117] Based on the trained graph model, the traffic data and conditions in the future are predicted, and the predicted traffic data and conditions in the future are compared with the traffic data and conditions of the simulated driving state to obtain the recorded detection data.

[0118] The communication function is tested in various driving environments, and the dynamic detection data of wireless communication capability is obtained, including:

[0119] Define the measurement parameter set Q = {moving speed V, communication distance D, network bandwidth B, network delay L}, where V, D, B, L represent moving speed, communication distance, network bandwidth and network delay respectively;

[0120] Assume that the data transmission stability of the USB, Bluetooth, and wireless public network of the driving recorder is U, B, and W respectively, and is measured by the functions f(Q)=U, g(Q)=B, and h(Q=W;

[0121] In the simulated mobile environment, set a specific mobile speed V1, communication distance D1, network bandwidth B1, and network delay L1, that is, set the parameter set Q1 = {V1, D1, B1, L1}, and then measure the data transmission stability U1, B1, W1 at this time according to f(Q1), g(Q1), and h(Q1);

[0122] Then, change the mobile environment conditions, set a new parameter set Q2 = {V2, D1, B1, L1}, V2> V1, and measure U2, B2, and W2 at this time;

[0123] According to the above steps, multiple groups of different parameter sets are obtained respectively. By measuring the data transmission stability under each group of conditions, a series of data transmission stability and efficiency values ​​are obtained: (U1, B1, W1), (U2, B2, W2), ..., (Un, Bn, Wn); the series of data transmission stability and efficiency values ​​are used as dynamic detection data of wireless communication capabilities.

[0124] Simulate various driving behaviors in various driving environments, test the driving behavior recognition and driving safety reminder functions of the driving recorder to obtain behavior-based driving test data:

[0125] Define a driving behavior description set A = {a1, a2, ..., an}, where a1, a2, ..., an are parameters related to driving behavior such as speed, steering angle, braking force, and relative position of the vehicle; and define a driving behavior recognition result set R = {r1, r2, ..., rn}, where r1, r2, ..., rn represent the recognized driving behavior types, including normal driving, sudden acceleration, and sudden braking;

[0126] The driving behavior recognition function of the driving recorder is defined as:

[0127] R = f(A),

[0128] Where f is the internal algorithm of the driving recorder, which inputs the driving behavior description set A and outputs the identified driving behavior type set R;

[0129] Similarly, the driving safety reminder function of the driving recorder is defined as a function g, the input is the identified driving behavior type set R, and the output is the safety reminder set S, that is,

[0130] S = g(R);

[0131] In simulating various driving behaviors, a series of driving behavior description sets A are first generated in the driving environment, including normal driving situations and various abnormal driving behaviors, and then input into the driving recorder, and the driving behavior recognition result set R is obtained through f(A), and the safety prompt set S is further obtained through g(R);

[0132] For each driving behavior description set A, there is an expected driving behavior recognition result set R' and an expected safety prompt set S'. The actual R and S are compared with the expected R' and S' to obtain the driving behavior recognition accuracy and driving safety prompt accuracy of the driving recorder; the driving behavior recognition accuracy and driving safety prompt accuracy of the driving recorder are used as behavior-based driving detection data.

[0133] The safety warning reaction speed and accuracy of the recorder are tested in various driving environments when simulating accidents, and intelligent safety warning detection data is obtained:

[0134] Define a set of simulated driving environment parameters E = {e1, e2, ..., ei}, where ei represents the i-th driving environment parameter, including: vehicle speed, vehicle distance, and road condition;

[0135] Define the reaction speed of intelligent safety warning as V, and the reaction accuracy as A. These two parameters are calculated by functions f(E) and g(E): V = f(E), A = g(E);

[0136] During the test, the parameter set E of the simulation environment is first set as a specific accident environment. If the distance between vehicles is too close, e1 = vehicle distance, then it is set as too close distance. If the speed is too fast, e2 = vehicle speed, then it is set as too fast speed.

[0137] Calculate the reaction speed V and reaction accuracy A, that is, V1 = f(E), A1 = g(E);

[0138] Then change the parameters of the simulation environment and obtain a series of V and A according to the above steps;

[0139] This series of reaction speeds and reaction accuracies are summarized and recorded as (V1, A1), (V2, A2), ..., (Vn, An); and the values ​​of V and A are statistically analyzed, and the results are used as intelligent safety early warning detection data.

[0140] Based on big data technology, the final test results are collected and analyzed, including record test data, wireless communication capability dynamic test data, driving test data, and intelligent safety warning test data. Figure 2 Shown include:

[0141] The data collected during all tests are defined as data set D, which includes driving environment parameters, driving recorder parameters, and various performance test results, namely

[0142] D={E,A,R,S,U,B,W,V1,A1,V2,A2,...,Vn,An};

[0143] Wherein, E = {e1, e2, ..., en} represents the parameters of various driving environments; A = {a1, a2, ..., an} represents the parameters of driving actions; R = {r1, r2, ..., rn} represents the type of driving behavior identified; S = {s1, s2, ..., sn} represents the safety prompts of the driving recorder; U, B, W represent the data transmission stability and efficiency of USB, Bluetooth, and wireless public networks respectively; V1, A1, V2, A2, ..., Vn, An represent the safety warning response speed and accuracy of the intelligent warning system in various driving environments respectively;

[0144] Use big data technology to preprocess data, remove outliers and invalid data, and perform data cleaning:

[0145] The outliers Z and invalid data I are defined by the following formulas;

[0146] Z = x|x∈D, and the value of x is outside its maximum and minimum range,

[0147] I=x|x∈D, and the value of x is missing or unreasonable,

[0148] The excluded data set D' is defined as:

[0149] D' = DZI;

[0150] Get the cleaned data set D',

[0151] D'={E',A',R',S',U',B',W',V1',A1',V2',A2',...,Vn',An'};

[0152] For the data set D', a variety of statistical analysis and machine learning methods are used for in-depth analysis or descriptive statistical analysis is used to analyze it, including calculating the mean μ, median θ and standard deviation σ of each parameter. The formula is as follows:

[0153]

[0154] Where N is the number of samples, xi is the i-th sample value;

[0155] Correlation analysis is used to describe the correlation between various parameters, and Pearson correlation coefficient r is used for correlation analysis. The calculation method is as follows:

[0156]

[0157] Among them, xi and yi represent the x and y values ​​of the i-th sample, μx, μy are the average values ​​of x and y, is the standard deviation of x and y;

[0158] For key performance indicators such as recognition accuracy and warning accuracy, the machine learning model SVM is used for prediction and optimization; the driving behavior recognition accuracy P is provided, and the prediction function p(X) = a*X+b is defined, where X is the parameter set, and a and b are the coefficients to be solved; the parameters of the prediction function are optimized by minimizing the error between the predicted value and the actual value. The error function E(a,b) is defined as:

[0159] E(a,b)=Σ[p(Xi)-Pi] 2

[0160] Among them, Pi is the actual driving behavior recognition accuracy, and p(Xi) is the predicted value;

[0161] Cluster analysis is used to perform data mining to find the underlying patterns of data. A commonly used indicator for evaluating clustering results is the silhouette coefficient S, which is calculated as follows:

[0162] S=(ba) / max(a,b)

[0163] Among them, a represents the average distance between the sample and samples of the same category, and b represents the minimum average distance between the sample and samples of other categories.

[0164] This embodiment conducts a comprehensive test on the various functions of the driving recorder by constructing various driving environments, and obtains recording detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data, which can comprehensively evaluate the performance of the driving recorder. Secondly, by constructing various driving environments and simulating driving behaviors, the performance of the driving recorder can be more realistically reflected, and the detection results obtained are more in line with the actual situation, thereby improving the detection effect.

[0165] Embodiment three:

[0166] This embodiment provides Figure 3 A driving recorder detection system is shown, comprising:

[0167] Environment construction module: build various driving environments;

[0168] Recording detection module: simulates driving status in various driving environments to detect the recording data and audio and video recording functions of the driving recorder, and obtains the recording data of the driving recorder; automatically analyzes the recording data of the driving recorder to obtain the recording detection data;

[0169] Wireless communication capability dynamic detection module: performs communication function detection in various driving environments to obtain dynamic detection data of wireless communication capability;

[0170] Driving detection module: simulates various driving behaviors in various driving environments, detects the driving behavior recognition and driving safety prompt functions of the driving recorder, and obtains behavior-based driving detection data;

[0171] Intelligent safety warning detection module: Test the safety warning response speed and accuracy of the recorder when simulating an accident in a variety of driving environments to obtain intelligent safety warning detection data;

[0172] Final detection module: Based on big data technology, it collects and analyzes recorded detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data to obtain the final detection results.

[0173] This embodiment conducts a comprehensive test on the various functions of the driving recorder by constructing various driving environments, and obtains recording detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data, which can comprehensively evaluate the performance of the driving recorder. Secondly, by constructing various driving environments and simulating driving behaviors, the performance of the driving recorder can be more realistically reflected, and the detection results obtained are more in line with the actual situation, thereby improving the detection effect.

[0174] Embodiment 4:

[0175] This embodiment provides a driving recorder detection storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned driving recorder detection method is implemented.

[0176] This embodiment conducts a comprehensive test on the various functions of the driving recorder by constructing various driving environments, and obtains recording detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data, which can comprehensively evaluate the performance of the driving recorder. Secondly, by constructing various driving environments and simulating driving behaviors, the performance of the driving recorder can be more realistically reflected, and the detection results obtained are more in line with the actual situation, thereby improving the detection effect.

[0177] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A detection method for a driving recorder, It is characterized in that include: Build multiple driving environments; Simulate driving conditions in various driving environments to detect the recorded data and audio and video recording functions of the driving recorder to obtain the recorded data of the driving recorder; automatically analyze the recorded data of the driving recorder to obtain the recorded detection data; Conduct communication function tests in various driving environments to obtain dynamic test data of wireless communication capabilities; Simulate various driving behaviors in various driving environments, test the driving behavior recognition and driving safety prompt functions of the driving recorder to obtain behavior-based driving test data; Test the safety warning response speed and accuracy of the recorder when simulating accidents in various driving environments to obtain intelligent safety warning detection data; Based on big data technology, the final test results are obtained by collecting and analyzing recorded test data, dynamic test data on wireless communication capabilities, driving test data, and intelligent safety warning test data.

2. The detection method for a driving recorder according to claim 1, It is characterized in that The multiple driving environments include: at least one driving environment in rainy and snowy weather, driving at night, and a congested city environment.

3. The detection method for a driving recorder according to claim 1, It is characterized in that The construction of various driving environments includes: Define parameterized driving environment characteristics, including: road type, weather conditions, lighting conditions, vehicle status and other traffic factors; Generate a simulation environment based on the specific conditions to be tested: When simulating the situation of urban roads, the road is divided into multiple sections, and the length, driving speed, and number of vehicles ahead of each section are generated by the given parameter generation model; The driving environment is described by the parameter vector P P=[p1,p2,...,pn] Among them, pi represents the i-th environmental parameter; According to the parameter vector P, the corresponding simulation environment is generated; specifically, each environment parameter is generated by a generating function fi: pi=fi(P') Among them, P' is a set of input parameters, which are random variables.

4. The detection method for a driving recorder according to claim 3, It is characterized in that The automatic analysis of the records of the driving recorder includes: Using graph neural networks (GNNs) to analyze driving recorder data: Build a graph model, consider vehicles and roads as nodes, and the distance and direction between vehicles and roads as edges, so as to construct a traffic network graph; Feature extraction: For each node, a feature vector is defined to express the node status, including the speed and acceleration of the vehicle, the length, width and speed limit of the road. The graph network will automatically propagate and integrate information, so that each node contains not only its own information, but also the information of adjacent nodes. Feature extraction is expressed as a function H, where H(v) represents the feature vector of node v. Graph network training: training is performed based on real traffic data, including traffic flow and traffic accidents, to obtain a trained graph model; Based on the trained graph model, the traffic data and conditions in the future are predicted, and the predicted traffic data and conditions in the future are compared with the traffic data and conditions of the simulated driving state to obtain the recorded detection data.

5. The detection method for a driving recorder according to claim 1, It is characterized in that Communication function tests are performed in various driving environments to obtain dynamic detection data of wireless communication capabilities, including: Define the measurement parameter set Q = {moving speed V, communication distance D, network bandwidth B, network delay L}, where V, D, B, L represent moving speed, communication distance, network bandwidth and network delay respectively; Assume that the data transmission stability of the USB, Bluetooth, and wireless public network of the driving recorder is U, B, and W respectively, and is measured by the functions f(Q)=U, g(Q)=B, and h(Q=W; In the simulated mobile environment, set a specific mobile speed V1, communication distance D1, network bandwidth B1, and network delay L1, that is, set the parameter set Q1 = {V1, D1, B1, L1}, and then measure the data transmission stability U1, B1, W1 at this time according to f(Q1), g(Q1), and h(Q1); Then, change the mobile environment conditions, set a new parameter set Q2 = {V2, D1, B1, L1}, V2> V1, and measure U2, B2, and W2 at this time; According to the above steps, multiple groups of different parameter sets are obtained respectively. By measuring the data transmission stability under each group of conditions, a series of data transmission stability and efficiency values ​​are obtained: (U1, B1, W1), (U2, B2, W2), ..., (Un, Bn, Wn); the series of data transmission stability and efficiency values ​​are used as dynamic detection data of wireless communication capabilities.

6. The detection method for a driving recorder according to claim 1, It is characterized in that Simulate various driving behaviors in various driving environments, test the driving behavior recognition and driving safety reminder functions of the driving recorder to obtain behavior-based driving test data: Define a driving behavior description set A = {a1, a2, ..., an}, where a1, a2, ..., an are parameters related to driving behavior such as speed, steering angle, braking force, and relative position of the vehicle; and define a driving behavior recognition result set R = {r1, r2, ..., rn}, where r1, r2, ..., rn represent the recognized driving behavior types, including normal driving, sudden acceleration, and sudden braking; The driving behavior recognition function of the driving recorder is defined as: R = f(A), Where f is the internal algorithm of the driving recorder, which inputs the driving behavior description set A and outputs the identified driving behavior type set R; Similarly, the driving safety reminder function of the driving recorder is defined as a function g, the input is the identified driving behavior type set R, and the output is the safety reminder set S, that is, S = g(R); In simulating various driving behaviors, a series of driving behavior description sets A are first generated in the driving environment, including normal driving situations and various abnormal driving behaviors, and then input into the driving recorder, and the driving behavior recognition result set R is obtained through f(A), and the safety prompt set S is further obtained through g(R); For each driving behavior description set A, there is an expected driving behavior recognition result set R' and an expected safety prompt set S'. The actual R and S are compared with the expected R' and S' to obtain the driving behavior recognition accuracy and driving safety prompt accuracy of the driving recorder; the driving behavior recognition accuracy and driving safety prompt accuracy of the driving recorder are used as behavior-based driving detection data.

7. The detection method for a driving recorder according to claim 1, It is characterized in that The safety warning reaction speed and accuracy of the recorder are tested in various driving environments when simulating accidents, and intelligent safety warning detection data is obtained: Define a set of simulated driving environment parameters E = {e1, e2, ..., ei}, where ei represents the i-th driving environment parameter, including: vehicle speed, vehicle distance, and road condition; Define the reaction speed of intelligent safety warning as V, and the reaction accuracy as A. These two parameters are calculated by functions f(E) and g(E): V = f(E), A = g(E); During the test, the parameter set E of the simulation environment is first set as a specific accident environment. If the distance between vehicles is too close, e1 = vehicle distance, then it is set as too close distance. If the speed is too fast, e2 = vehicle speed, then it is set as too fast speed. Calculate the reaction speed V and reaction accuracy A, that is, V1 = f(E), A1 = g(E); Then change the parameters of the simulation environment and obtain a series of V and A according to the above steps; This series of reaction speeds and reaction accuracies are summarized and recorded as (V1, A1), (V2, A2), ..., (Vn, An); and the values ​​of V and A are statistically analyzed, and the results are used as intelligent safety early warning detection data.

8. The detection method for a driving recorder according to claim 1, It is characterized in that Based on big data technology, the final test results obtained by collecting and analyzing the recorded test data, wireless communication capability dynamic test data, driving test data and intelligent safety warning test data include: The data collected during all tests are defined as data set D, which includes driving environment parameters, driving recorder parameters, and various performance test results, namely D={E,A,R,S,U,B,W,V1,A1,V2,A2,...,Vn,An}; Wherein, E = {e1, e2, ..., en} represents the parameters of various driving environments; A = {a1, a2, ..., an} represents the parameters of driving actions; R = {r1, r2, ..., rn} represents the type of driving behavior identified; S = {s1, s2, ..., sn} represents the safety prompts of the driving recorder; U, B, W represent the data transmission stability and efficiency of USB, Bluetooth, and wireless public networks respectively; V1, A1, V2, A2, ..., Vn, An represent the safety warning response speed and accuracy of the intelligent warning system in various driving environments respectively; Use big data technology to preprocess data, remove outliers and invalid data, and perform data cleaning: The outliers Z and invalid data I are defined by the following formulas; Z = x|x∈D, and the value of x is outside its maximum and minimum range, I=x|x∈D, and the value of x is missing or unreasonable, The excluded data set D' is defined as: D' = DZI; Get the cleaned data set D', D'={E',A',R',S',U',B',W',V1',A1',V2',A2',...,Vn',An'}; For the data set D', a variety of statistical analysis and machine learning methods are used for in-depth analysis or descriptive statistical analysis is used to analyze it, including calculating the mean μ, median θ and standard deviation σ of each parameter. The formula is as follows: (if N is an odd number); or θ=(xi+xi+1) / 2 (if N is an even number); Where N is the number of samples, xi is the i-th sample value; Correlation analysis is used to describe the correlation between various parameters, and Pearson correlation coefficient r is used for correlation analysis. The calculation method is as follows: Among them, xi and yi represent the x and y values ​​of the i-th sample, μx, μy are the average values ​​of x and y, is the standard deviation of x and y; For key performance indicators such as recognition accuracy and warning accuracy, the machine learning model SVM is used for prediction and optimization; the driving behavior recognition accuracy P is provided, and the prediction function p(X) = a*X+b is defined, where X is the parameter set, and a and b are the coefficients to be solved; the parameters of the prediction function are optimized by minimizing the error between the predicted value and the actual value. The error function E(a,b) is defined as: E(a,b)=Σ[p(Xi)-Pi] 2 Among them, Pi is the actual driving behavior recognition accuracy, and p(Xi) is the predicted value; Cluster analysis is used to perform data mining to find the underlying patterns of data. A commonly used indicator for evaluating clustering results is the silhouette coefficient S, which is calculated as follows: S=(ba) / max(a,b) Among them, a represents the average distance between the sample and samples of the same category, and b represents the minimum average distance between the sample and samples of other categories.

9. A driving recorder detection system, It is characterized in that include: Environment construction module: build various driving environments; Recording detection module: simulates driving status in various driving environments to detect the recording data and audio and video recording functions of the driving recorder, and obtains the recording data of the driving recorder; automatically analyzes the recording data of the driving recorder to obtain the recording detection data; Wireless communication capability dynamic detection module: performs communication function detection in various driving environments to obtain dynamic detection data of wireless communication capability; Driving detection module: simulates various driving behaviors in various driving environments, detects the driving behavior recognition and driving safety prompt functions of the driving recorder, and obtains behavior-based driving detection data; Intelligent safety warning detection module: Test the safety warning response speed and accuracy of the recorder when simulating an accident in a variety of driving environments to obtain intelligent safety warning detection data; Final detection module: Based on big data technology, it collects and analyzes recorded detection data, wireless communication capability dynamic detection data, driving detection data, and intelligent safety warning detection data to obtain the final detection results.

10. A driving recorder detection storage medium, It is characterized in that A computer program is stored thereon, and when the computer program is executed by a processor, a detection method for a driving recorder as claimed in any one of claims 1 to 8 is implemented.

Images