Processing method and device for evaluating driving control performance of vehicle

Abstract

The embodiment of the invention relates to a processing method and device for evaluating the driving control performance of a vehicle. The method comprises the steps that a processing model for evaluating the driving control performance of the vehicle according to a collected data recording sequence of the vehicle is constructed and recorded as a corresponding first evaluation model; constructing a corresponding first data set through data acquisition; training the first evaluation model based on the first data set; after the model training is finished, receiving a first acquisition data record sequence of any vehicle; and the first evaluation model evaluates the driving control performance of the current vehicle according to the first acquisition data record sequence to obtain a corresponding first evaluation score. According to the invention, the evaluation efficiency and the evaluation real-time performance can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a processing method and device for evaluating the driving control performance of a vehicle. Background Art

[0002] When evaluating the driving control performance of an automatic / unmanned vehicle, it is necessary to conduct a multi-faceted evaluation of the following vehicle performances: straight-line performance, turning performance, pedestrian passability, lane change performance, meeting performance, overtaking performance, U-turn performance, etc. The current conventional processing method for evaluating these performances is to first collect and save real-time data during vehicle driving, and then analyze and score the collected data manually or semi-manually according to different performance evaluation purposes after the vehicle has finished driving. Obviously, this conventional processing method is difficult to effectively improve its evaluation efficiency due to the existence of manual operation processes. Summary of the invention

[0003] The purpose of the present invention is to provide a processing method, device, electronic device and computer-readable storage medium for evaluating the driving control performance of a vehicle in view of the defects of the prior art. The present invention constructs and trains an end-to-end driving control performance evaluation model; based on the evaluation model, the seven types of performance (straight performance, left turn performance, sidewalk passing performance, lane change performance, meeting performance, overtaking performance, and U-turn performance) of the vehicle can be instantly scored according to a real-time / historical sequence of collected data records of the vehicle, and the seven types of scores are weighted summed to output a corresponding total performance score. The evaluation model of the present invention can not only get rid of the manual operation process, but also perform real-time evaluation during the vehicle driving process, thereby achieving the purpose of improving the evaluation efficiency and real-time evaluation.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present invention provides a processing method for evaluating the driving control performance of a vehicle, the method comprising:

[0005] Constructing a processing model for evaluating the driving control performance of the vehicle according to the vehicle's collected data record sequence as a corresponding first evaluation model; constructing a corresponding first data set through data collection; and training the first evaluation model based on the first data set;

[0006] After the model training is completed, a first collected data record sequence of any vehicle is received; and the first evaluation model evaluates the driving control performance of the current vehicle according to the first collected data record sequence to obtain a corresponding first evaluation score.

[0007] Preferably, the model input end of the first evaluation model is used to receive the collected data record sequence, and the model output end is used to output the corresponding evaluation score; the first evaluation model is used to evaluate the driving control performance of the vehicle according to the collected data record sequence input by the model and output the corresponding evaluation score;

[0008] The collected data record sequence includes multiple collected data records; the collected data record includes a timestamp t i , coordinate c i , vehicle speed i , acceleration a i , longitudinal acceleration ay i , lateral acceleration ax i 、Steering wheel angle ω i and driving distance s i , 1≤index i≤N, N is the total number of records in the acquisition data record sequence; the time difference between every two adjacent timestamps △t=t i+1 -t i fixed and divisible by 0.3 seconds; the coordinate c i , the vehicle speed v i , the acceleration a i For the vehicle at the time stamp t i The real-time coordinates, real-time vehicle speed and real-time acceleration on the vehicle; the longitudinal acceleration ay i and the lateral acceleration ax i is the acceleration a i The longitudinal and lateral components of the steering wheel angle ω i The vehicle steering wheel is at the time stamp t i The real-time turning angle on the axis; the corresponding driving distance s when the index i is 1 i＝1 is 0, and the corresponding driving distance s when the index i is greater than 1 i>1 For vehicles from timestamp t i＝1 To the current timestamp t i>1 Relative driving distance;

[0009] The first evaluation model includes a feature recognition module, a straight-ahead performance prediction head, a lane-changing performance prediction head, a sidewalk passing performance prediction head, a meeting performance prediction head, a overtaking performance prediction head, a U-turn performance prediction head, a left-turn performance prediction head and a comprehensive scoring module;

[0010] The input end of the feature recognition module is connected to the model input end, and the first, second, third, fourth, fifth, sixth and seventh output ends are respectively connected to the input ends of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head; the first, second, third, fourth, fifth, sixth and seventh input ends of the comprehensive scoring module are respectively connected to the output ends of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head;

[0011] The feature recognition module is used to perform feature item recognition processing according to the collected data record sequence to obtain corresponding thirteen types of feature recognition items; and perform prediction head input vector preparation processing according to the preset correlation relationship between the feature recognition items and the predicted performance items and the thirteen types of feature recognition items to obtain corresponding seven types of prediction head input vectors; and send the seven types of prediction head input vectors to the corresponding straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head respectively; the thirteen types of feature recognition items include average speed, number of sudden accelerations, number of sudden decelerations, speed variation coefficient, acceleration smoothness, longitudinal acceleration smoothness, lateral acceleration smoothness, initial maximum steering wheel angle per 100 meters, ratio of sudden lane changes, maximum speed, relative displacement of the first deceleration point, maximum steering wheel speed and number of positive and negative changes in steering wheel angle; the seven types of prediction head input vectors include first, second, third, fourth, fifth, sixth and seventh prediction head input vectors;

[0012] The straight-ahead performance prediction head, the lane-changing performance prediction head, the sidewalk-passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left-turn performance prediction head are each implemented based on a type of nonlinear regression model; the nonlinear regression model at least includes a decision tree model, a random forest model, a gradient boosting tree model and an MLP model;

[0013] The straight performance prediction head is used to predict the straight performance score of the vehicle according to the first prediction head input vector to obtain the corresponding first prediction score and send it to the comprehensive scoring module; the lane change performance prediction head is used to predict the lane change performance score of the vehicle according to the second prediction head input vector to obtain the corresponding second prediction score and send it to the comprehensive scoring module; the sidewalk passing performance prediction head is used to predict the sidewalk passing performance score of the vehicle according to the third prediction head input vector to obtain the corresponding third prediction score and send it to the comprehensive scoring module; the meeting performance prediction head is used to predict the meeting performance score of the vehicle according to the fourth prediction head input vector to obtain the corresponding fourth prediction score and send it to the comprehensive scoring module; the overtaking performance prediction head is used to predict the overtaking performance score of the vehicle according to the fifth prediction head input vector to obtain the corresponding fifth prediction score and send it to the comprehensive scoring module; the U-turn performance prediction head is used to predict the U-turn performance score of the vehicle according to the sixth prediction head input vector to obtain the corresponding sixth prediction score and send it to the comprehensive scoring module; the left turn performance prediction head is used to predict the left turn performance score of the vehicle according to the seventh prediction head input vector to obtain the corresponding seventh prediction score and send it to the comprehensive scoring module;

[0014] The comprehensive scoring module is used to calculate the weighted sum of the first, second, third, fourth, fifth, sixth and seventh prediction scores to obtain the corresponding evaluation score and output it; Evaluation score = w 1 ×First prediction score+w 2 × Second prediction score + w 3 ×Third prediction score + w 4 ×Fourth prediction score+w 5 × Fifth Prediction Rating + w 6 ×Sixth prediction score+w 7 ×Seventh prediction score, w 1 、w 2 、w 3 、w 4 、w 5 、w 6 、w 7 There are seven preset weighting coefficients, and the sum of the seven weighting coefficients is 1.

[0015] Furthermore, the feature item recognition processing is performed according to the collected data record sequence to obtain corresponding thirteen types of feature recognition items, specifically including:

[0016] The feature recognition module records all the vehicle speeds v in the collected data sequence. i Calculate the mean value to obtain the corresponding average vehicle speed; and record the maximum vehicle speed v in the collected data record sequence. ias the corresponding maximum speed;

[0017] And the data collection sequence exceeds 3m / s 2 The longitudinal acceleration ay i The number of rapid accelerations is counted and the statistical result is used as the corresponding rapid acceleration number;

[0018] The acquisition data record sequence is sequentially divided into a plurality of first subsequences with a division time of 0.3 seconds, and the first and last timestamps t i The time interval is 0.3 seconds; and all the first subsequences are traversed in one round; and in this round of traversal, the first subsequence currently traversed is used as the corresponding current subsequence; and all the longitudinal accelerations ay in the current subsequence are traversed in one round. i Are they all less than -4.5m / s? 2 Identify, if yes, record the current subsequence as the corresponding first abnormal subsequence; and count the total number of the first abnormal subsequences at the end of this round of traversal and use the statistical result as the corresponding number of rapid decelerations;

[0019] And record all the vehicle speeds v in the acquisition data sequence i The corresponding speed standard deviation is calculated by calculating the standard deviation of the speed; and the corresponding speed variation coefficient is calculated based on the speed standard deviation and the average vehicle speed = speed standard deviation / average vehicle speed;

[0020] And record each acceleration a in the acquisition data sequence i , each of the longitudinal accelerations ay i and each of the lateral accelerations ax i The first-order derivative of is calculated to obtain the corresponding first derivative, second derivative and third derivative; and the standard deviation of all the first derivatives obtained is calculated and the calculation result is used as the corresponding acceleration smoothness; and the standard deviation of all the second derivatives obtained is calculated and the calculation result is used as the corresponding longitudinal acceleration smoothness; and the standard deviation of all the third derivatives obtained is calculated and the calculation result is used as the corresponding lateral acceleration smoothness;

[0021] The first driving distance s greater than 100 meters in the acquisition data recording sequence is recorded as i The corresponding collected data record is recorded as the corresponding current end record; and all the steering wheel angles ω between the first collected data and the current end record are recorded. i The maximum value among is extracted as the corresponding initial maximum steering wheel angle of 100 meters;

[0022] And based on all the coordinates c in the acquisition data record sequence i The total number of lane changes during vehicle driving is counted with a preset high-precision road map to obtain a corresponding first total number; the driving trajectory of each lane change is extracted to obtain a corresponding first lane change trajectory; each of the first lane change trajectories is identified as a sudden lane change trajectory, and if so, the current first lane change trajectory is recorded as a corresponding abnormal lane change trajectory; the total number of all the abnormal lane change trajectories obtained is counted to obtain a corresponding second total number; and the corresponding ratio of the number of sudden lane changes is calculated based on the first and second total numbers = (second total number / first total number) × 100%;

[0023] And record the first acceleration a less than 0 in the acquisition data recording sequence i The corresponding collected data record is recorded as the corresponding first deceleration record; and the driving distance s of the first deceleration record is recorded as i as the corresponding relative displacement of the first deceleration point;

[0024] And record each time stamp t in the acquisition data sequence i Calculate the absolute value of the steering wheel speed on the steering wheel, and use the maximum absolute speed as the corresponding maximum steering wheel speed;

[0025] and counting the number of positive and negative switching of the steering wheel angle in the collected data record sequence and using the statistical result as the corresponding number of positive and negative changes of the steering wheel angle;

[0026] The obtained average speed, the number of sudden accelerations, the number of sudden decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the initial maximum steering wheel angle per 100 meters, the ratio of the number of sudden lane changes, the maximum speed, the relative displacement of the first deceleration point, the maximum steering wheel speed and the number of positive and negative changes in the steering wheel angle constitute the corresponding thirteen categories of feature identification items.

[0027] Further, the first prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the initial maximum steering wheel angle per 100 meters;

[0028] The second prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the ratio of the number of rapid lane changes;

[0029] The third prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the maximum speed and the relative displacement of the first deceleration point;

[0030] The fourth prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the maximum steering wheel speed;

[0031] The fifth prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the ratio of the number of rapid lane changes;

[0032] The sixth prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the maximum speed, and the number of positive and negative changes in the steering wheel angle;

[0033] The seventh prediction head input vector consists of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness and the lateral acceleration smoothness.

[0034] Preferably, the first data set includes multiple first data records; the first data record includes a first training data record sequence, a first label score, a second label score, a third label score, a fourth label score, a fifth label score, a sixth label score and a seventh label score; the first training data record sequence includes multiple first training data records, and the first training data record has the same data format as the collected data record; the first, second, third, fourth, fifth, sixth and seventh label scores are respectively the label scores of the corresponding straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head; the first, second, third, fourth, fifth, sixth and seventh label scores are scores obtained by other manual scoring methods or machine scoring methods.

[0035] Preferably, the training of the first evaluation model based on the first data set specifically includes:

[0036] Step 61, based on a preset first segmentation ratio, the first data set is divided into two sub-data sets recorded as a corresponding first training set and a first evaluation set;

[0037] Wherein, both the first training set and the first evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first segmentation ratio;

[0038] Step 62, perform a round of traversal on all the first data records of the first training set; and in this round of traversal, use the first data record currently traversed as the corresponding current test record; and input the first training data record sequence of the current test record into the first evaluation model for processing to obtain the output scores of the seven prediction heads as the corresponding first, second, third, fourth, fifth, sixth and seventh scores; and bring the first score and the first label score of the current test record into the preset first model loss function; and bring the second score and the second label score of the current test record into the preset second model loss function; and bring the third score and the third label score of the current test record into the preset third model loss function; and bring the fourth score and the The fourth label score of the current test record is brought into the preset fourth model loss function; and the fifth score and the fifth label score of the current test record are brought into the preset fifth model loss function; and the sixth score and the sixth label score of the current test record are brought into the preset sixth model loss function; and the seventh score and the seventh label score of the current test record are brought into the preset seventh model loss function; and a round of optimization is performed on the model parameters of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head in a manner that minimizes the first, second, third, fourth, fifth, sixth and seventh model loss functions;

[0039] Wherein, the first, second, third, fourth, fifth, sixth and seventh model loss functions are implemented based on the L1 loss function or the L2 loss function;

[0040] Step 63, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and input the first training data record sequence of the current evaluation record into the first evaluation model for processing to obtain the output scores of the seven prediction heads as the corresponding eighth, ninth, tenth, eleventh, twelfth, thirteenth and fourteenth scores to form a corresponding first prediction vector; and the first, second, third, fourth, fifth, sixth and seventh label scores of the current evaluation record form a corresponding first label vector; and the first prediction vector and the first label vector form a corresponding first prediction-label pair; and at the end of this round of traversal, bring all the obtained first prediction-label pairs into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value;

[0041] Wherein, the first model evaluation function is RMSE function;

[0042] Step 64, identifying whether the first evaluation value satisfies a preset first evaluation value range; if the first evaluation value does not satisfy the first evaluation value range, returning to step 61 to continue training; if the first evaluation value satisfies the first evaluation value range, stopping the mode training and confirming that the model training is completed.

[0043] A second aspect of an embodiment of the present invention provides a device for implementing the processing method for evaluating the driving control performance of a vehicle as described in the first aspect, the device comprising: a model building and training module and a model application module;

[0044] The model building and training module is used to build a processing model for evaluating the driving control performance of the vehicle according to the vehicle's collected data record sequence, which is recorded as the corresponding first evaluation model; and to build a corresponding first data set through data collection; and to train the first evaluation model based on the first data set;

[0045] The model application module is used to receive a first collected data record sequence of any vehicle after the model training is completed; and the first evaluation model evaluates the driving control performance of the current vehicle according to the first collected data record sequence to obtain a corresponding first evaluation score.

[0046] A third aspect of an embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0047] The processor is used to be coupled to the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;

[0048] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0049] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions. When the computer instructions are executed by a computer, the computer executes the instructions of the method described in the first aspect above.

[0050] The embodiment of the present invention provides a processing method, device, electronic device and computer-readable storage medium for evaluating the driving control performance of a vehicle. It can be seen from the above invention content that the embodiment of the present invention constructs and trains an end-to-end driving control performance evaluation model; based on the evaluation model, the seven types of performance (straight performance, left turn performance, sidewalk passing performance, lane change performance, meeting performance, overtaking performance, U-turn performance) of the vehicle can be instantly scored according to a real-time / historical sequence of collected data records of the vehicle, and the seven types of scores are weighted summed to output a corresponding total performance score. The evaluation model of the embodiment of the present invention can not only get rid of the manual operation process, but also perform real-time evaluation during the vehicle driving process, which not only improves the real-time nature of the evaluation, but also improves the evaluation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a processing method for evaluating the driving control performance of a vehicle provided in the first embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the modules of the first evaluation model of the first embodiment of the present invention;

[0053] Figure 3 This is a schematic diagram of the data structure of the acquisition data record sequence of the first embodiment of the present invention;

[0054] Figure 4 A schematic diagram of the association relationship between feature identification items and prediction performance items according to the first embodiment of the present invention;

[0055] Figure 5 A module structure diagram of a processing device for evaluating the driving control performance of a vehicle provided in the second embodiment of the present invention;

[0056] Figure 6 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all 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.

[0058] A processing method for evaluating the driving control performance of a vehicle is provided in a first embodiment of the present invention, such as Figure 1 A schematic diagram of a processing method for evaluating the driving control performance of a vehicle provided in the first embodiment of the present invention is shown, and includes the following main steps:

[0059] Step 1, constructing a processing model for evaluating the driving control performance of the vehicle according to the vehicle's collected data record sequence, recorded as the corresponding first evaluation model; and constructing the corresponding first data set through data collection; and training the first evaluation model based on the first data set;

[0060] Specifically, it includes: step 11, constructing a processing model for evaluating the driving control performance of the vehicle according to the collected data record sequence of the vehicle, which is recorded as the corresponding first evaluation model;

[0061] Here, if Figure 2 As shown in the module schematic diagram of the first evaluation model of the first embodiment of the present invention, the model input end of the first evaluation model of the embodiment of the present invention is used to receive the collection data record sequence, and the model output end is used to output the corresponding evaluation score; the first evaluation model is used to evaluate the driving control performance of the vehicle according to the collection data record sequence input by the model and output the corresponding evaluation score;

[0062] The collected data record sequence input to the model is a segment of real-time or historical collected data of any vehicle, and the collected data record sequence includes multiple collected data records; Figure 3 The data structure diagram of the data collection record sequence of the first embodiment of the present invention is shown in the following figure. The data collection record of the embodiment of the present invention includes a timestamp t i , coordinate c i , vehicle speed i , acceleration a i , longitudinal acceleration ay i , lateral acceleration ax i 、Steering wheel angle ω i and driving distance s i , 1≤index i≤N, N is the total number of records in the data collection record sequence; the time difference between every two adjacent timestamps △t=t i+1 -t i Fixed and divisible by 0.3 seconds; coordinate c i, vehicle speed i , acceleration a i is the vehicle at timestamp t i Real-time coordinates, real-time vehicle speed and real-time acceleration on the vehicle; longitudinal acceleration ay i and the lateral acceleration ax i is the acceleration a i The longitudinal and lateral components of the steering wheel angle ω i The vehicle steering wheel at timestamp t i The real-time turning angle on the ; the corresponding driving distance s when the index i is 1 i＝1 is 0, and the corresponding driving distance s when index i is greater than 1 i>1 For vehicles from timestamp t i＝1 To the current timestamp t i>1 Relative driving distance;

[0063] like Figure 2 As shown, the model components of the first evaluation model of the embodiment of the present invention include: a feature recognition module, a straight performance prediction head, a lane change performance prediction head, a sidewalk passing performance prediction head, a meeting performance prediction head, a overtaking performance prediction head, a U-turn performance prediction head, a left turn performance prediction head and a comprehensive scoring module;

[0064] The connection relationship of each model component is as follows: the input end of the feature recognition module is connected to the model input end, and the first, second, third, fourth, fifth, sixth and seventh output ends are respectively connected to the input ends of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head; the first, second, third, fourth, fifth, sixth and seventh input ends of the comprehensive scoring module are respectively connected to the output ends of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head;

[0065] The component functions of each model component are briefly described as follows:

[0066] 1) Feature recognition module:

[0067] The feature recognition module of the embodiment of the present invention is used to perform feature item recognition processing according to the collected data record sequence to obtain corresponding thirteen types of feature recognition items; and perform prediction head input vector preparation processing according to the preset association relationship between the feature recognition items and the prediction performance items and the thirteen types of feature recognition items to obtain corresponding seven types of prediction head input vectors; and send the seven types of prediction head input vectors to the corresponding straight performance prediction head, lane change performance prediction head, sidewalk passing performance prediction head, meeting performance prediction head, overtaking performance prediction head, U-turn performance prediction head and left turn performance prediction head respectively;

[0068] Among them, thirteen types of feature recognition items include average speed, number of sudden accelerations, number of sudden decelerations, speed variation coefficient, acceleration smoothness, longitudinal acceleration smoothness, lateral acceleration smoothness, initial maximum steering wheel angle per 100 meters, ratio of sudden lane changes, maximum speed, relative displacement of the first deceleration point, maximum steering wheel speed, and number of positive and negative changes in steering wheel angle; seven types of prediction head input vectors include the first, second, third, fourth, fifth, sixth, and seventh prediction head input vectors;

[0069] In the above module processing steps, feature item recognition processing is performed according to the collected data record sequence to obtain corresponding thirteen types of feature recognition items, which is specifically composed of the following steps A1-A11:

[0070] Step A1: the feature recognition module collects all the vehicle speed v in the data record sequence. i Calculate the mean to get the corresponding average speed; and record the maximum speed v in the sequence of collected data i As the corresponding maximum speed;

[0071] Step A2: record the data in the sequence that exceeds 3m / s 2 The longitudinal acceleration ay i The number of is counted and the statistical result is used as the corresponding number of rapid accelerations;

[0072] Step A3, the collected data record sequence is sequentially divided into a plurality of first subsequences with a segmentation time of 0.3 seconds, and the first and last timestamps of each first subsequence are t i The time interval is 0.3 seconds; and all the first subsequences are traversed in one round; and in this round of traversal, the first subsequence currently traversed is used as the corresponding current subsequence; and all the longitudinal accelerations ay in the current subsequence are traversed in one round. i Are they all less than -4.5m / s? 2 Identify, if yes, record the current subsequence as the corresponding first abnormal subsequence; and at the end of this round of traversal, count the total number of the first abnormal subsequences and use the statistical result as the corresponding number of rapid decelerations;

[0073] Step A4, and record all the vehicle speeds v in the collected data sequence i The corresponding speed standard deviation is calculated by the standard deviation of the speed; and the corresponding speed variation coefficient is calculated based on the speed standard deviation and the average vehicle speed = speed standard deviation / average vehicle speed;

[0074] Step A5, and record each acceleration a in the acquisition data sequence i , each longitudinal acceleration ay i And each lateral acceleration ax iThe first-order derivative of is calculated to obtain the corresponding first derivative, second derivative and third derivative; and the standard deviation of all the first derivatives is calculated and the calculation result is used as the corresponding acceleration smoothness; and the standard deviation of all the second derivatives is calculated and the calculation result is used as the corresponding longitudinal acceleration smoothness; and the standard deviation of all the third derivatives is calculated and the calculation result is used as the corresponding lateral acceleration smoothness;

[0075] Step A6, and record the first driving distance s greater than 100 meters in the collected data record sequence i The corresponding collected data record is recorded as the corresponding current end record; and all steering wheel angles ω between the first collected data and the current end record are recorded i The maximum value among is extracted as the corresponding initial maximum steering wheel angle per 100 meters;

[0076] Step A7, and based on the collected data record sequence all coordinates c i The total number of lane changes during vehicle driving is counted with a preset high-precision road map to obtain a corresponding first total number; the driving trajectory of each lane change is extracted to obtain a corresponding first lane change trajectory; and each first lane change trajectory is identified as a sudden lane change trajectory, and if so, the current first lane change trajectory is recorded as a corresponding abnormal lane change trajectory; and the total number of all abnormal lane change trajectories obtained is counted to obtain a corresponding second total number; and the corresponding sudden lane change frequency ratio is calculated based on the first and second total numbers = (second total number / first total number) × 100%;

[0077] Step A8, and record the first acceleration a less than 0 in the collected data sequence i The corresponding collected data record is recorded as the corresponding first deceleration record; and the driving distance s of the first deceleration record is recorded as i As the corresponding relative displacement of the first deceleration point;

[0078] Step A9, and record the timestamps t in the data collection sequence. i The absolute values ​​of the steering wheel speeds are calculated, and the maximum absolute speed is used as the corresponding maximum steering wheel speed;

[0079] Step A10, counting the number of positive and negative switching of the steering wheel angle in the collected data record sequence and using the statistical result as the corresponding number of positive and negative changes of the steering wheel angle;

[0080] Step A11, the obtained average speed, number of sudden accelerations, number of sudden decelerations, speed variation coefficient, acceleration smoothness, longitudinal acceleration smoothness, lateral acceleration smoothness, initial maximum steering wheel angle per 100 meters, ratio of sudden lane changes, maximum speed, relative displacement of the first deceleration point, maximum steering wheel speed, and number of positive and negative changes in the steering wheel angle form corresponding thirteen types of feature recognition items;

[0081] The correlation between the feature identification item and the prediction performance item in the embodiment of the present invention is as follows: Figure 4 As shown in the schematic diagram of the association relationship between the feature identification items and the prediction performance items of the first embodiment of the present invention, this association relationship is reflected in the feature data components of the prediction head input vector: 1) the first prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the initial maximum steering wheel angle of 100 meters; 2) the second prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the ratio of the number of rapid lane changes; 3) the third prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the maximum speed and the relative displacement of the first deceleration point; 4) The fourth prediction head input vector is composed of the average speed, the number of sudden accelerations, the number of sudden decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the maximum steering wheel speed; 5) The fifth prediction head input vector is composed of the average speed, the number of sudden accelerations, the number of sudden decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the ratio of the number of sudden lane changes; 6) The sixth prediction head input vector is composed of the average speed, the number of sudden accelerations, the number of sudden decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the maximum speed and the number of positive and negative changes in the steering wheel angle; 7) The seventh prediction head input vector is composed of the average speed, the number of sudden accelerations, the number of sudden decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness and the lateral acceleration smoothness;

[0082] 2) Seven types of prediction heads:

[0083] like Figure 2 As shown, the seven types of prediction heads of the first evaluation model of the embodiment of the present invention are: straight performance prediction head, lane change performance prediction head, sidewalk passing performance prediction head, meeting performance prediction head, overtaking performance prediction head, U-turn performance prediction head and left turn performance prediction head; these seven types of prediction heads are each implemented based on a type of nonlinear regression model; the nonlinear regression model mentioned here includes at least a decision tree model, a random forest model, a gradient boosting tree model and an MLP model; the functions of these seven types of prediction heads are as follows:

[0084] a. The straight driving performance prediction head is used to predict the straight driving performance score of the vehicle according to the first prediction head input vector to obtain the corresponding first prediction score and send it to the comprehensive scoring module;

[0085] b. The lane change performance prediction head is used to predict the lane change performance score of the vehicle according to the second prediction head input vector to obtain a corresponding second prediction score and send it to the comprehensive scoring module;

[0086] c. The pedestrian passability performance prediction head is used to predict the pedestrian passability performance score of the vehicle according to the third prediction head input vector to obtain the corresponding third prediction score and send it to the comprehensive scoring module;

[0087] d. The meeting performance prediction head is used to predict the meeting performance score of the vehicle according to the fourth prediction head input vector to obtain the corresponding fourth prediction score and send it to the comprehensive scoring module;

[0088] e. The overtaking performance prediction head is used to predict the overtaking performance score of the vehicle according to the fifth prediction head input vector to obtain the corresponding fifth prediction score and send it to the comprehensive scoring module;

[0089] f. The U-turn performance prediction head is used to predict the U-turn performance score of the vehicle according to the sixth prediction head input vector to obtain the corresponding sixth prediction score and send it to the comprehensive scoring module;

[0090] g. The left-turn performance prediction head is used to predict the left-turn performance score of the vehicle according to the seventh prediction head input vector to obtain the corresponding seventh prediction score and send it to the comprehensive scoring module;

[0091] 3) Comprehensive scoring module:

[0092] The comprehensive scoring module of the embodiment of the present invention is used to calculate the weighted sum of the first, second, third, fourth, fifth, sixth and seventh prediction scores to obtain the corresponding evaluation scores and output them;

[0093] Where, evaluation score = w 1 ×First prediction score+w 2 × Second prediction score + w 3 ×Third prediction score + w 4 ×Fourth prediction score+w 5 × Fifth Prediction Rating + w 6 ×Sixth prediction score+w 7 ×Seventh prediction score; w 1 、w 2 、w 3 、w 4 、w 5 、w 6 、w 7There are seven preset weighting coefficients, and the sum of the seven weighting coefficients is 1;

[0094] Step 12, and constructing a corresponding first data set through data collection;

[0095] Among them, the first data set includes multiple first data records; the first data record includes a first training data record sequence, a first label score, a second label score, a third label score, a fourth label score, a fifth label score, a sixth label score and a seventh label score; the first training data record sequence includes multiple first training data records, and the data format of the first training data record is the same as that of the collected data record; the first, second, third, fourth, fifth, sixth and seventh label scores are the label scores of the corresponding straight performance prediction head, lane change performance prediction head, sidewalk passing performance prediction head, meeting performance prediction head, overtaking performance prediction head, U-turn performance prediction head and left turn performance prediction head respectively; the first, second, third, fourth, fifth, sixth and seventh label scores are scores obtained by other manual scoring methods or machine scoring methods;

[0096] Step 13, training the first evaluation model based on the first data set;

[0097] Specifically, it includes: step 131, dividing the first data set into two sub-data based on a preset first division ratio and recording them as a corresponding first training set and a first evaluation set;

[0098] The first segmentation ratio is a preset ratio, such as 8:2; the first training set and the first evaluation set both include a plurality of first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first segmentation ratio;

[0099] Step 132, perform a round of traversal on all the first data records of the first training set; and in this round of traversal, use the currently traversed first data record as the corresponding current test record; and input the first training data record sequence of the current test record into the first evaluation model for processing to obtain the output scores of the seven prediction heads as the corresponding first, second, third, fourth, fifth, sixth and seventh scores; and bring the first score and the first label score of the current test record into the preset first model loss function; and bring the second score and the second label score of the current test record into the preset second model loss function; and bring the third score and the third label score of the current test record into the preset third model loss function; and bring the fourth score into the preset third model loss function. The fourth label score of the current test record is brought into the preset fourth model loss function; the fifth score and the fifth label score of the current test record are brought into the preset fifth model loss function; the sixth score and the sixth label score of the current test record are brought into the preset sixth model loss function; the seventh score and the seventh label score of the current test record are brought into the preset seventh model loss function; and a round of optimization is performed on the model parameters of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head in a way that the first, second, third, fourth, fifth, sixth and seventh model loss functions are minimized;

[0100] Among them, the first, second, third, fourth, fifth, sixth and seventh model loss functions are implemented based on the L1 loss function or the L2 loss function;

[0101] Step 133, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and input the first training data record sequence of the current evaluation record into the first evaluation model for processing to obtain the output scores of the seven prediction heads as the corresponding eighth, ninth, tenth, eleventh, twelfth, thirteenth and fourteenth scores to form a corresponding first prediction vector; and the first, second, third, fourth, fifth, sixth and seventh label scores of the current evaluation record form a corresponding first label vector; and the first prediction vector and the first label vector form a corresponding first prediction-label pair; and at the end of this round of traversal, bring all the obtained first prediction-label pairs into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value;

[0102] Among them, the first model evaluation function is the RMSE function;

[0103] Step 134, identifying whether the first evaluation value satisfies the preset first evaluation value range; if the first evaluation value does not satisfy the first evaluation value range, returning to step 131 to continue training; if the first evaluation value satisfies the first evaluation value range, stopping the mode training and confirming the end of the model training.

[0104] Here, the first evaluation value range is a preset evaluation value range.

[0105] Step 2: After the model training is completed, a first collected data record sequence of any vehicle is received; and a first evaluation model is used to evaluate the driving control performance of the current vehicle according to the first collected data record sequence to obtain a corresponding first evaluation score.

[0106] Here, the first collected data record sequence is a record sequence output by a certain vehicle in real time, which can be regarded as a real-time evaluation of the vehicle's driving control performance based on the first evaluation model; the first collected data record sequence can also be a record sequence collected by a certain vehicle in a certain historical period, which can be regarded as a historical retrospection of the vehicle's driving control performance based on the first evaluation model. After receiving the first collected data record sequence, the first evaluation model will first perform feature item recognition processing according to the first collected data record sequence by the feature recognition module to obtain the corresponding thirteen categories of feature recognition items, and perform prediction head input vector preparation processing according to the correlation between the feature recognition items and the predicted performance items and the thirteen categories of feature recognition items to obtain the corresponding seven categories of prediction head input vectors, which are respectively input into the corresponding straight performance prediction head, lane change performance prediction head, sidewalk passing performance prediction head, meeting performance prediction head, overtaking performance prediction head, U-turn performance prediction head and left turn performance prediction head; then the straight performance prediction head, lane change performance prediction head, sidewalk passing performance prediction head, meeting performance prediction head, overtaking performance prediction head, U-turn performance prediction head and left turn performance prediction head each perform the corresponding straight performance, lane change performance, sidewalk passing performance, meeting performance, overtaking performance, U-turn performance and left turn performance score prediction according to their own input vectors to obtain the corresponding seven categories of scores and input them into the comprehensive scoring module; finally, the comprehensive scoring module performs weighted sum calculation on the seven categories of scores to obtain the corresponding first evaluation score and output it.

[0107] Figure 5 The module structure diagram of a processing device for evaluating the driving control performance of a vehicle provided in the second embodiment of the present invention, the device is a terminal device or a server that implements the aforementioned method embodiment, and can also be a device that enables the aforementioned terminal device or server to implement the aforementioned method embodiment, for example, the device can be a device or chip system of the aforementioned terminal device or server. Figure 5 As shown, the processing device for evaluating the driving control performance of a vehicle provided in the second embodiment of the present invention includes: a model building and training module 201 and a model application module 202.

[0108] The model building and training module 201 is used to build a processing model for evaluating the vehicle's driving control performance based on the vehicle's collected data record sequence, which is recorded as the corresponding first evaluation model; and to build the corresponding first data set through data collection; and to train the first evaluation model based on the first data set.

[0109] The model application module 202 is used to receive a first collected data record sequence of any vehicle after the model training is completed; and the first evaluation model evaluates the driving control performance of the current vehicle according to the first collected data record sequence to obtain a corresponding first evaluation score.

[0110] A processing device for evaluating the driving control performance of a vehicle provided in an embodiment of the present invention can execute the method steps in the above method embodiment. Its implementation principle and technical effect are similar and will not be repeated here.

[0111] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the model building and training module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The function of the above-mentioned module is determined. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in software form.

[0112] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital signal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module above is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0113] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the above method embodiments are generated. The above computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above-mentioned computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the above-mentioned computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.) methods. The above-mentioned computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The above-mentioned available medium can be a magnetic medium (such as a floppy disk, a hard disk, a tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.

[0114] Figure 6 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. The electronic device may be a terminal device or a server for implementing the method of the aforementioned embodiment, or may be a terminal device or a server for implementing the method of the aforementioned embodiment connected to the aforementioned terminal device or server. Figure 6As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver 303. Various instructions may be stored in the memory 302 to complete various processing functions and implement the processing steps described in the aforementioned embodiment method. Preferably, the electronic device involved in the embodiment of the present invention also includes: a power supply 304, a system bus 305 and a communication port 306. The system bus 305 is used to realize the communication connection between components. The above-mentioned communication port 306 is used for connecting and communicating between the electronic device and other peripherals.

[0115] exist Figure 6 The system bus 305 mentioned in the figure can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include random access memory (RAM), and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk storage.

[0116] The above-mentioned processor can be a general-purpose processor, including a central processing unit CPU, a network processor (Network Processor, NP), a graphics processing unit (Graphics Processing Unit, GPU), etc.; it can also be a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0117] It should be noted that an embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the method and processing process provided in the above embodiments.

[0118] The embodiment of the present invention provides a processing method, device, electronic device and computer-readable storage medium for evaluating the driving control performance of a vehicle. It can be seen from the above invention content that the embodiment of the present invention constructs and trains an end-to-end driving control performance evaluation model; based on the evaluation model, the seven types of performance (straight performance, left turn performance, sidewalk passing performance, lane change performance, meeting performance, overtaking performance, U-turn performance) of the vehicle can be instantly scored according to a real-time / historical sequence of collected data records of the vehicle, and the seven types of scores are weighted summed to output a corresponding total performance score. The evaluation model of the embodiment of the present invention can not only get rid of the manual operation process, but also perform real-time evaluation during the vehicle driving process, which not only improves the real-time nature of the evaluation, but also improves the evaluation efficiency.

[0119] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0120] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0121] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating the driving control performance of a vehicle, characterized in that: The method comprises: Constructing a processing model for evaluating the driving control performance of the vehicle according to the vehicle's collected data record sequence as a corresponding first evaluation model; constructing a corresponding first data set through data collection; and training the first evaluation model based on the first data set; After the model training is completed, a first collected data record sequence of any vehicle is received; and the first evaluation model evaluates the driving control performance of the current vehicle according to the first collected data record sequence to obtain a corresponding first evaluation score.

2. The method for evaluating the driving control performance of a vehicle according to claim 1, characterized in that: The model input end of the first evaluation model is used to receive the collected data record sequence, and the model output end is used to output the corresponding evaluation score; The first evaluation model is used to evaluate the driving control performance of the vehicle according to the collected data record sequence input by the model and output a corresponding evaluation score; The collected data record sequence includes multiple collected data records; the collected data record includes a timestamp t i , coordinate c i , vehicle speed i , acceleration a i , longitudinal acceleration ay i , lateral acceleration ax i 、Steering wheel angle ω i and driving distance s i , 1≤index i≤N, N is the total number of records in the acquisition data record sequence; the time difference between every two adjacent timestamps △t=t i+1 -t i fixed and divisible by 0.3 seconds; the coordinate c i , the vehicle speed v i , the acceleration a i For the vehicle at the time stamp t i The real-time coordinates, real-time vehicle speed and real-time acceleration on the vehicle; the longitudinal acceleration ay i and the lateral acceleration ax i is the acceleration a i The longitudinal and lateral components of the steering wheel angle ω i The vehicle steering wheel is at the time stamp t i The real-time turning angle on the axis; the corresponding driving distance s when the index i is 1 i＝1 is 0, and the corresponding driving distance s when the index i is greater than 1 i>1 For vehicles from timestamp t i＝1 To the current timestamp t i>1 Relative driving distance; The first evaluation model includes a feature recognition module, a straight-ahead performance prediction head, a lane-changing performance prediction head, a sidewalk passing performance prediction head, a meeting performance prediction head, a overtaking performance prediction head, a U-turn performance prediction head, a left-turn performance prediction head and a comprehensive scoring module; The input end of the feature recognition module is connected to the model input end, and the first, second, third, fourth, fifth, sixth and seventh output ends are respectively connected to the input ends of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head; the first, second, third, fourth, fifth, sixth and seventh input ends of the comprehensive scoring module are respectively connected to the output ends of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head; The feature recognition module is used to perform feature item recognition processing according to the collected data record sequence to obtain corresponding thirteen types of feature recognition items; and perform prediction head input vector preparation processing according to the preset correlation relationship between the feature recognition items and the predicted performance items and the thirteen types of feature recognition items to obtain corresponding seven types of prediction head input vectors; and send the seven types of prediction head input vectors to the corresponding straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head respectively; the thirteen types of feature recognition items include average speed, number of sudden accelerations, number of sudden decelerations, speed variation coefficient, acceleration smoothness, longitudinal acceleration smoothness, lateral acceleration smoothness, initial maximum steering wheel angle per 100 meters, ratio of sudden lane changes, maximum speed, relative displacement of the first deceleration point, maximum steering wheel speed and number of positive and negative changes in steering wheel angle; the seven types of prediction head input vectors include first, second, third, fourth, fifth, sixth and seventh prediction head input vectors; The straight-ahead performance prediction head, the lane-changing performance prediction head, the sidewalk-passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left-turn performance prediction head are each implemented based on a type of nonlinear regression model; the nonlinear regression model at least includes a decision tree model, a random forest model, a gradient boosting tree model and an MLP model; The straight performance prediction head is used to predict the straight performance score of the vehicle according to the first prediction head input vector to obtain the corresponding first prediction score and send it to the comprehensive scoring module; the lane change performance prediction head is used to predict the lane change performance score of the vehicle according to the second prediction head input vector to obtain the corresponding second prediction score and send it to the comprehensive scoring module; the sidewalk passing performance prediction head is used to predict the sidewalk passing performance score of the vehicle according to the third prediction head input vector to obtain the corresponding third prediction score and send it to the comprehensive scoring module; the meeting performance prediction head is used to predict the meeting performance score of the vehicle according to the fourth prediction head input vector to obtain the corresponding fourth prediction score and send it to the comprehensive scoring module; the overtaking performance prediction head is used to predict the overtaking performance score of the vehicle according to the fifth prediction head input vector to obtain the corresponding fifth prediction score and send it to the comprehensive scoring module; the U-turn performance prediction head is used to predict the U-turn performance score of the vehicle according to the sixth prediction head input vector to obtain the corresponding sixth prediction score and send it to the comprehensive scoring module; the left turn performance prediction head is used to predict the left turn performance score of the vehicle according to the seventh prediction head input vector to obtain the corresponding seventh prediction score and send it to the comprehensive scoring module; The comprehensive scoring module is used to calculate the weighted sum of the first, second, third, fourth, fifth, sixth and seventh prediction scores to obtain the corresponding evaluation score and output it; evaluation score = w1×first prediction score+w2×second prediction score+w3×third prediction score+w4×fourth prediction score+w5×fifth prediction score+w6×sixth prediction score+w7×seventh prediction score, w1, w2, w3, w4, w5, w6, w7 are seven preset weighting coefficients, and the sum of the seven weighting coefficients is 1.

3. The method for evaluating the driving control performance of a vehicle according to claim 2, characterized in that: The feature item recognition processing is performed according to the collected data record sequence to obtain corresponding thirteen types of feature recognition items, specifically including: The feature recognition module records all the vehicle speeds v in the collected data sequence. i Perform mean calculation to obtain the corresponding average vehicle speed; and record the maximum vehicle speed v in the collected data record sequence. i as the corresponding maximum speed; And the data collection sequence exceeds 3m / s 2 The longitudinal acceleration ay i The number of rapid accelerations is counted and the statistical result is used as the corresponding rapid acceleration number; The acquisition data record sequence is sequentially divided into a plurality of first subsequences with a division time of 0.3 seconds, and the first and last timestamps t i The time interval is 0.3 seconds; and all the first subsequences are traversed in one round; and in this round of traversal, the first subsequence currently traversed is used as the corresponding current subsequence; and all the longitudinal accelerations ay in the current subsequence are traversed in one round. i Are they all less than -4.5m / s? 2 Identify, if yes, record the current subsequence as the corresponding first abnormal subsequence; and count the total number of the first abnormal subsequences at the end of this round of traversal and use the statistical result as the corresponding number of rapid decelerations; And record all the vehicle speeds v in the acquisition data sequence i The standard deviation of the speed is calculated to obtain the corresponding speed standard deviation; and the corresponding speed variation coefficient is calculated based on the speed standard deviation and the average vehicle speed = speed standard deviation / average vehicle speed; And record each acceleration a in the acquisition data sequence i , each of the longitudinal accelerations ay i and each of the lateral accelerations ax i The first-order derivative of is calculated to obtain the corresponding first derivative, second derivative and third derivative; and the standard deviation of all the first derivatives obtained is calculated and the calculation result is used as the corresponding acceleration smoothness; and the standard deviation of all the second derivatives obtained is calculated and the calculation result is used as the corresponding longitudinal acceleration smoothness; and the standard deviation of all the third derivatives obtained is calculated and the calculation result is used as the corresponding lateral acceleration smoothness; The first driving distance s greater than 100 meters in the acquisition data recording sequence is recorded as i The corresponding collected data record is recorded as the corresponding current end record; and all the steering wheel angles ω between the first collected data and the current end record are recorded. i The maximum value among is extracted as the corresponding initial maximum steering wheel angle of 100 meters; And based on all the coordinates c in the acquisition data record sequence i The total number of lane changes during vehicle driving is counted with a preset high-precision road map to obtain a corresponding first total number; the driving trajectory of each lane change is extracted to obtain a corresponding first lane change trajectory; each of the first lane change trajectories is identified as a sudden lane change trajectory, and if so, the current first lane change trajectory is recorded as a corresponding abnormal lane change trajectory; the total number of all the abnormal lane change trajectories obtained is counted to obtain a corresponding second total number; and the corresponding ratio of the number of sudden lane changes is calculated based on the first and second total numbers = (second total number / first total number) × 100%; And record the first acceleration a less than 0 in the acquisition data recording sequence i The corresponding collected data record is recorded as the corresponding first deceleration record; and the driving distance s of the first deceleration record is recorded as i as the corresponding relative displacement of the first deceleration point; And record each time stamp t in the acquisition data sequence i Calculate the absolute value of the steering wheel speed on the steering wheel, and use the maximum absolute speed as the corresponding maximum steering wheel speed; and counting the number of positive and negative switching of the steering wheel angle in the collected data record sequence and using the statistical result as the corresponding number of positive and negative changes of the steering wheel angle; The obtained average speed, the number of sudden accelerations, the number of sudden decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the initial maximum steering wheel angle per 100 meters, the ratio of the number of sudden lane changes, the maximum speed, the relative displacement of the first deceleration point, the maximum steering wheel speed and the number of positive and negative changes in the steering wheel angle constitute the corresponding thirteen categories of feature identification items.

4. The method for evaluating the driving control performance of a vehicle according to claim 2, characterized in that: The first prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the initial maximum steering wheel angle per 100 meters; The second prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the ratio of the number of rapid lane changes; The third prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the maximum speed and the relative displacement of the first deceleration point; The fourth prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the maximum steering wheel speed; The fifth prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness and the ratio of the number of rapid lane changes; The sixth prediction head input vector is composed of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness, the lateral acceleration smoothness, the maximum speed, and the number of positive and negative changes in the steering wheel angle; The seventh prediction head input vector consists of the average speed, the number of rapid accelerations, the number of rapid decelerations, the speed variation coefficient, the acceleration smoothness, the longitudinal acceleration smoothness and the lateral acceleration smoothness.

5. The method for evaluating the driving control performance of a vehicle according to claim 2, characterized in that: The first data set includes multiple first data records; the first data record includes a first training data record sequence, a first label score, a second label score, a third label score, a fourth label score, a fifth label score, a sixth label score and a seventh label score; the first training data record sequence includes multiple first training data records, and the first training data record has the same data format as the collected data record; the first, second, third, fourth, fifth, sixth and seventh label scores are the label scores of the corresponding straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head respectively; the first, second, third, fourth, fifth, sixth and seventh label scores are scores obtained by other manual scoring methods or machine scoring methods.

6. The method for evaluating the driving control performance of a vehicle according to claim 5, characterized in that: The training of the first evaluation model based on the first data set specifically includes: Step 61, based on a preset first segmentation ratio, the first data set is divided into two sub-data sets recorded as a corresponding first training set and a first evaluation set; Wherein, both the first training set and the first evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first segmentation ratio; Step 62, perform a round of traversal on all the first data records of the first training set; and in this round of traversal, use the first data record currently traversed as the corresponding current test record; and input the first training data record sequence of the current test record into the first evaluation model for processing to obtain the output scores of the seven prediction heads as the corresponding first, second, third, fourth, fifth, sixth and seventh scores; and bring the first score and the first label score of the current test record into the preset first model loss function; and bring the second score and the second label score of the current test record into the preset second model loss function; and bring the third score and the third label score of the current test record into the preset third model loss function; and bring the fourth score and the The fourth label score of the current test record is brought into the preset fourth model loss function; and the fifth score and the fifth label score of the current test record are brought into the preset fifth model loss function; and the sixth score and the sixth label score of the current test record are brought into the preset sixth model loss function; and the seventh score and the seventh label score of the current test record are brought into the preset seventh model loss function; and a round of optimization is performed on the model parameters of the straight performance prediction head, the lane change performance prediction head, the sidewalk passing performance prediction head, the meeting performance prediction head, the overtaking performance prediction head, the U-turn performance prediction head and the left turn performance prediction head in a manner that minimizes the first, second, third, fourth, fifth, sixth and seventh model loss functions; Wherein, the first, second, third, fourth, fifth, sixth and seventh model loss functions are implemented based on the L1 loss function or the L2 loss function; Step 63, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and input the first training data record sequence of the current evaluation record into the first evaluation model for processing to obtain the output scores of the seven prediction heads as the corresponding eighth, ninth, tenth, eleventh, twelfth, thirteenth and fourteenth scores to form a corresponding first prediction vector; and the first, second, third, fourth, fifth, sixth and seventh label scores of the current evaluation record form a corresponding first label vector; and the first prediction vector and the first label vector form a corresponding first prediction-label pair; and at the end of this round of traversal, bring all the obtained first prediction-label pairs into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value; Wherein, the first model evaluation function is RMSE function; Step 64, identifying whether the first evaluation value satisfies a preset first evaluation value range; if the first evaluation value does not satisfy the first evaluation value range, returning to step 61 to continue training; if the first evaluation value satisfies the first evaluation value range, stopping the mode training and confirming that the model training is completed.

7. A device for executing the processing method for evaluating the driving control performance of a vehicle according to any one of claims 1 to 6, characterized in that: The device comprises: a model building and training module and a model application module; The model building and training module is used to build a processing model for evaluating the driving control performance of the vehicle according to the vehicle's collected data record sequence, which is recorded as the corresponding first evaluation model; and to build a corresponding first data set through data collection; and to train the first evaluation model based on the first data set; The model application module is used to receive a first collected data record sequence of any vehicle after the model training is completed; and the first evaluation model evaluates the driving control performance of the current vehicle according to the first collected data record sequence to obtain a corresponding first evaluation score.

8. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 6; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 6.

Images