Driving behavior big data-based driver speed control ability analysis method and device

Abstract

The invention relates to the technical field of intelligent traffic and driving behavior analysis, in particular to a driving behavior big data-based driver speed control ability analysis method and device. A data acquisition and preprocessing step: acquiring related data of a driver in a driving process through a vehicle-mounted sensor, and processing the acquired data; in the feature extraction step, key features used for analyzing the speed control ability of the driver are extracted from the preprocessed data; in the model construction and evaluation step, a machine learning model is adopted to process the extracted features, a model capable of evaluating the speed control ability of the driver is constructed, and a speed control ability score is calculated; in the dynamic evaluation and feedback step, grade evaluation of the driver is generated based on a calculation result and a specific function, and personalized feedback suggestions are given in combination with a driving scene.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation and driving behavior analysis, and particularly to a method and device for analyzing the speed control ability of drivers based on big data of driving behavior. Background Art

[0002] In the field of modern transportation, traffic safety has always been of utmost importance, and the driving behavior of drivers plays a crucial role in traffic conditions. Traditional means of evaluating driving behavior mainly rely on manual observation and simple qualitative analysis. For example, traffic police visually judge the driving operations of drivers on the road, or score the drivers' behaviors based on limited rules of thumb. However, this method has many drawbacks. On the one hand, the efficiency of manual evaluation is extremely low, making it difficult to meet the evaluation needs of a large number of drivers. In the current situation of increasing traffic flow, it cannot meet the requirements of rapid and comprehensive evaluation. On the other hand, the accuracy of qualitative analysis is severely limited, being greatly affected by the subjective factors of evaluators. Different evaluators may have completely different judgments on the same driving behavior, resulting in the lack of reliability and consistency of evaluation results.

[0003] With the rapid development of information technology, intelligent transportation systems have emerged and gradually become the core development direction in the transportation field. The emergence of big data technology has brought new opportunities for driving behavior analysis, enabling the collection and storage of a large amount of driving behavior data. At the same time, the continuous progress of artificial intelligence technology has made it possible to deeply mine and accurately analyze these data. Among many aspects of driving behavior, the speed control ability of drivers has a crucial impact on traffic safety and driving comfort. Stable and reasonable speed control can effectively reduce the probability of traffic accidents and improve the riding experience of passengers in the vehicle.

[0004] Therefore, it is of extremely important practical significance to develop an efficient and accurate method and device for analyzing the speed control ability of drivers based on big data of driving behavior, which not only helps to improve the level of traffic safety management but also provides strong support for the further improvement of intelligent transportation systems. Summary of the Invention

[0005] In view of the deficiencies existing in the current relevant prior arts, the present invention provides a method and device for analyzing the speed control ability of drivers based on driving behavior big data, which are efficient, accurate and have low hardware dependence, aiming to optimize the algorithm calculation efficiency, improve the evaluation accuracy, and expand its applicability in low-cost devices. Through innovative designs and technical improvements in multiple aspects, this patent has successfully achieved efficient, accurate and low-hardware-dependent analysis of the speed control ability of drivers, providing a brand-new and highly practical solution for driving behavior evaluation in the field of intelligent transportation, and is expected to play an important role in many fields such as intelligent driving assistance systems, driver behavior monitoring, and intelligent vehicle systems, promoting the continuous development and progress of intelligent transportation technologies.

[0006] To achieve the above object, the first aspect of the present invention provides a method for analyzing the speed control ability of drivers based on driving behavior big data, which is characterized in that it at least includes the following steps:

[0007] Step S1: Data collection and preprocessing step, collecting relevant data of the driver during driving through in-vehicle sensors, and processing the collected data, including one or more of filling in missing data, smoothing noise, cleaning, denoising and standardization processing;

[0008] Step S2: Feature extraction step, extracting key features for analyzing the speed control ability of drivers from the preprocessed data, and the features include but are not limited to one or more features related to driving trajectory, speed gear matching, steering wheel operation and acceleration;

[0009] Step S3: Model construction and evaluation step, using a machine learning model to process the extracted features, constructing a model capable of evaluating the speed control ability of drivers, and calculating the speed control ability score;

[0010] Step S4: Dynamic evaluation and feedback step, generating a level evaluation of the driver based on the calculation results and a specific function, and giving personalized feedback suggestions in combination with the driving scenario.

[0011] Further, in step S1, it also includes:

[0012] Step S11: The in-vehicle sensors collect data at a predetermined frequency, and the collected data includes speed, gear and steering wheel angle data;

[0013] Step S12: Use the cubic spline interpolation algorithm to fill in missing data, use the Savitzky-Golay filter to smooth noise, and clean and denoise the data through data cleaning and denoising rules, and use the standardization method to convert the data into a standard form that meets the requirements of subsequent analysis.

[0014] Further, in step S2, it also includes:

[0015] Step S21: The training trajectory score is determined by comparing the actual driving path of the driver with the standard trajectory, so as to evaluate the driving performance of the driver on the specified path;

[0016] Step S22: The duration of speed-gear mismatch is cumulatively calculated according to the matching model of vehicle gear and speed, and this model takes into account the basic performance parameters of the vehicle;

[0017] Step S23: The number of positive and negative changes in the steering wheel angle is obtained by counting the positive and negative changes in the steering wheel angle within a unit time;

[0018] Step S24: The starting deceleration position calculates the position information of the vehicle when it enters the deceleration state from the stable state according to the law of speed change;

[0019] Step S25: The acceleration smoothness is calculated according to a specific formula, and the formula is: where Δa i = Δa i+1 - Δa i represents the acceleration change between consecutive time points, and a max is the maximum absolute value of the acceleration change. The smoothness range is [0, 1], and the higher the value, the smoother the acceleration;

[0020] Step S26: The maximum absolute value of the steering wheel angle directly obtains the maximum value of the steering wheel angle.

[0021] Furthermore, in step S3, it also includes:

[0022] Step S31: An evaluation is carried out using a model that fuses the Gradient Boosting Decision Tree GBDT and the Random Forest RF, and the speed control ability score is calculated according to a specific weight combination formula;

[0023] The model formula is S = 0.6·GBDT(x) + 0.4·RF(x),

[0024] where x is the feature vector containing the extracted features, and S is the speed control ability score;

[0025] Step S32: During the model training process, according to the distribution characteristics of the training data and the evaluation accuracy requirements, the hyperparameters of the GBDT model and the RF model are adjusted and optimized to improve the model performance.

[0026] Furthermore, in step S4, it also includes:

[0027] Step S41: Calculate the driver level probability based on the Softmax function according to the formula;

[0028]

[0029] where z i is the score of the i-th level after the eigenvector is calculated by the model;

[0030] Step S42: Comprehensively consider the road type, traffic conditions, and surrounding environment information of the current driving scenario, and combine the calculated level probabilities to generate targeted personalized feedback suggestions. The generation of the feedback suggestions also refers to the driver's historical driving data and the average performance of drivers of the same type.

[0031] Furthermore, it also includes: Step S5: Data storage and transmission step. Set up a data storage module to store the collected raw data, processed data, extracted features, model training process data, evaluation results, and feedback suggestions. At the same time, configure a data transmission module to achieve data interaction with external devices or systems. The external devices or systems include intelligent driving assistance systems, traffic management department servers, and vehicle remote monitoring centers to achieve data sharing and collaborative work.

[0032] Furthermore, it also includes: Step S6: User interface display step. Set up a user interface module to display the speed control ability score, level evaluation results, personalized feedback suggestions, and statistical analysis charts of historical driving data of the driver. The user interface module has good interactivity and visualization effects, which is convenient for the driver to intuitively understand their own speed control ability status and improvement direction. And the user interface module can adaptively adjust the display layout and content according to the needs of different users and device types.

[0033] The second aspect of the present invention provides a device for analyzing the speed control ability of a driver based on driving behavior big data. This device is used to implement the above method and includes:

[0034] A data collection and preprocessing unit for performing the functions of Step S1, that is, collecting relevant data of the driver during driving through in-vehicle sensors and processing the collected data, including but not limited to filling in missing data, smoothing noise, cleaning, denoising, and standardization processing;

[0035] A feature extraction unit for performing the functions of Step S2, that is, extracting key features for analyzing the driver's speed control ability from the preprocessed data. These features include but are not limited to features related to driving trajectory, speed gear matching, steering wheel operation, and acceleration;

[0036] A model construction and evaluation unit for performing the functions of Step S3, that is, using a machine learning model to process the extracted features, constructing a model capable of evaluating the driver's speed control ability, and calculating the speed control ability score;

[0037] A dynamic evaluation and feedback unit, which is used to perform the function of step S4, that is, to generate a driver's level evaluation based on the calculation results and a specific function, and give personalized feedback suggestions in combination with the driving scenario.

[0038] Furthermore, it also includes: a data storage and transmission unit, which is used to perform the function of step S5, that is, to set up a data storage module to store the collected raw data, processed data, extracted features, model training process data, evaluation results and feedback suggestions, and at the same time configure a data transmission module to achieve data interaction with external devices or systems. The external devices or systems include intelligent driving assistance systems, traffic management department servers, vehicle remote monitoring centers, etc., so as to achieve data sharing and collaborative work.

[0039] Furthermore, it also includes: a user interface display unit, which is used to perform the function of step S6, that is, to set up a user interface module to display the speed control ability score of the driver, the level evaluation result, personalized feedback suggestions, and statistical analysis charts of historical driving data. The user interface module has good interactivity and visualization effects, which is convenient for the driver to intuitively understand their own speed control ability status and improvement direction, and the user interface module can adaptively adjust the display layout and content according to the needs of different users and device types.

[0040] The present invention adopts the above technical solutions and has at least the following beneficial effects:

[0041] With the help of big data analysis and machine learning technologies, the present invention realizes a comprehensive evaluation of the driver's speed control ability and analyzes the driving behavior from multiple perspectives. Through in-depth analysis of multiple features such as the training trajectory score, the duration of speed-gear mismatch, and the change in the steering wheel angle, the present invention can more accurately evaluate the driver's speed control ability and provide personalized driving optimization suggestions accordingly. Using the Softmax function to output the probability distribution, the evaluation result is more intuitive, which helps to improve the driver's safety and driving comfort.

[0042] The present invention proposes a method and device for analyzing the speed control ability of drivers based on driving behavior big data. Six key features are selected, namely, training trajectory score, duration of speed-gear mismatch, number of positive and negative changes in steering wheel angle, starting deceleration position, acceleration smoothness, and maximum absolute value of steering wheel rotation angle. The fusion technology of gradient boosting decision tree (GBDT) and random forest (RF) models is adopted to achieve efficient analysis of speed control ability; a multi-level evaluation mechanism is constructed by introducing dynamic thresholds and Softmax functions to output speed control ability scores and improvement suggestions; combined with data augmentation strategies and low-precision sensor optimization, the generality and robustness of the algorithm are significantly improved. The present invention is applicable to low-cost device environments and has advantages such as high computing efficiency, accurate evaluation, and low hardware dependence. The technical solution of the present invention can be applied to fields such as intelligent driving assistance systems, driver behavior monitoring, and intelligent vehicle systems, which helps to improve driver safety, driving comfort, and energy-saving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0044] Figure 1 It is a flowchart of the method for analyzing the speed control ability of drivers of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0046] As Figure 1 shown, in the first aspect of this embodiment, a method for analyzing the speed control ability of drivers based on driving behavior big data is provided, which is characterized by at least including the following steps:

[0047] Step S1: A data collection and preprocessing step, where relevant data of the driver during driving is collected through in-vehicle sensors, and the collected data is processed, including one or more of filling in missing data, smoothing noise, cleaning, denoising, and standardization processing;

[0048] Step S2: Feature extraction step, extracting key features for analyzing the driver's speed control ability from the preprocessed data, where the features include one or more features related to, but not limited to, driving trajectory, speed gear matching, steering wheel operation, and acceleration;

[0049] Step S3: Model construction and evaluation step, using a machine learning model to process the extracted features, constructing a model capable of evaluating the driver's speed control ability, and calculating a speed control ability score;

[0050] Step S4: Dynamic evaluation and feedback step, generating a level evaluation of the driver based on the calculation results and a specific function, and giving personalized feedback suggestions in combination with the driving scenario.

[0051] As a preferred implementation manner, this embodiment further includes in step S1:

[0052] Step S11: The vehicle-mounted sensor collects data at a predetermined frequency, and the collected data includes speed, gear, and steering wheel angle data; collecting the driver's driving behavior data under different road conditions, and collecting data such as speed, gear, and steering wheel angle at a frequency of 5 Hz per second through the vehicle-mounted sensor.

[0053] Step S12: Using the cubic spline interpolation algorithm to complement the missing data, using the Savitzky-Golay filter to smooth the noise, and cleaning and denoising the data through data cleaning and denoising rules, and converting the data into a standard form that meets the requirements of subsequent analysis by using the standardization method.

[0054] As a preferred implementation manner, this embodiment further includes in step S2:

[0055] Step S21: Training trajectory score: The training trajectory score is determined by comparing the driver's actual driving path with the standard trajectory to evaluate the driver's driving performance on the specified path;

[0056] Step S22: Duration of speed-gear mismatch: The duration of speed-gear mismatch is calculated cumulatively according to the speed-gear matching model of the vehicle, and this model considers the basic performance parameters of the vehicle;

[0057] Step S23: Number of positive and negative changes in the steering wheel angle: The number of positive and negative changes in the steering wheel angle is obtained by counting the positive and negative changes in the steering wheel angle within a unit time;

[0058] Step S24: Starting deceleration position: The starting deceleration position calculates the position information of the vehicle when it enters the deceleration state from the stable state according to the law of speed change;

[0059] Step S25: Acceleration smoothness: The acceleration smoothness is calculated according to a specific formula, and the formula is: where Δa i = Δa i+1 - Δa i represents the acceleration change between consecutive time points, a max is the absolute maximum value of the acceleration change, and the smoothness range is [0, 1]. The higher the value, the smoother the acceleration;

[0060] Step S26: The absolute maximum value of the steering wheel angle: The absolute maximum value of the steering wheel angle directly obtains the maximum value of the steering wheel angle.

[0061] As a preferred implementation, this embodiment further includes in step S3:

[0062] Step S31: Use a model that fuses the Gradient Boosting Decision Tree GBDT and the Random Forest RF to evaluate, and calculate the speed control ability score according to a specific weight combination formula;

[0063] The model formula is S = 0.6·GBDT(x) + 0.4·RF(x),

[0064] where x is the feature vector containing the extracted features, and S is the speed control ability score;

[0065] Step S32: During the model training process, adjust and optimize the hyperparameters of the GBDT model and the RF model according to the distribution characteristics of the training data and the evaluation accuracy requirements to improve the model performance.

[0066] As a preferred implementation, this embodiment further includes in step S4:

[0067] Step S41: Calculate the driver level probability based on the Softmax function according to the formula;

[0068]

[0069] where, z i is the score of the i-th level after the feature vector is calculated by the model;

[0070] Step S42: Comprehensively consider the road type, traffic conditions, and surrounding environment information of the current driving scenario, and combine the calculated level probability to generate targeted personalized feedback suggestions. The generation of the feedback suggestions also refers to the driver's historical driving data and the average performance of drivers of the same type. For example, prompt the driver to decelerate in advance or avoid sudden acceleration on certain sections.

[0071] As a preferred implementation, this embodiment further includes: Step S5: Data storage and transmission step. A data storage module is set to store the collected raw data, processed data, extracted features, model training process data, evaluation results, and feedback suggestions. At the same time, a data transmission module is configured to achieve data interaction with external devices or systems, and the external devices or systems include an intelligent driving assistance system, a traffic management department server, and a vehicle remote monitoring center to achieve data sharing and collaborative work.

[0072] As a preferred implementation, this embodiment further includes: Step S6: User interface display step. A user interface module is set to display the speed control ability score of the driver, the level evaluation result, personalized feedback suggestions, and the statistical analysis chart of historical driving data. The user interface module has good interactivity and visualization effects, which is convenient for the driver to intuitively understand their own speed control ability status and improvement direction. And the user interface module can adaptively adjust the display layout and content according to the needs of different users and device types.

[0073] In the second aspect of this embodiment, a driver speed control ability analysis device based on driving behavior big data is provided. This device is used to implement the above method and includes:

[0074] A data collection and preprocessing unit, which is used to execute the function of Step S1, that is, to collect relevant data of the driver during driving through in-vehicle sensors and process the collected data, including but not limited to filling in missing data, smoothing noise, cleaning, denoising, and standardization processing;

[0075] A feature extraction unit, which is used to execute the function of Step S2, that is, to extract key features for analyzing the driver's speed control ability from the preprocessed data. These features include but not limited to features related to driving trajectory, speed gear matching, steering wheel operation, and acceleration;

[0076] A model construction and evaluation unit, which is used to execute the function of Step S3, that is, to process the extracted features using a machine learning model, construct a model capable of evaluating the driver's speed control ability, and calculate the speed control ability score;

[0077] A dynamic evaluation and feedback unit, which is used to execute the function of Step S4, that is, to generate a level evaluation of the driver based on the calculation results and a specific function, and give personalized feedback suggestions in combination with the driving scenario.

[0078] This embodiment also includes: a data storage and transmission unit, which is used to perform the functions of step S5, that is, setting a data storage module to store the collected raw data, processed data, extracted features, model training process data, evaluation results, and feedback suggestions, and at the same time configuring a data transmission module to realize data interaction with external devices or systems, where the external devices or systems include an intelligent driving assistance system, a traffic management department server, a vehicle remote monitoring center, etc., so as to achieve data sharing and collaborative work.

[0079] This embodiment also includes: a user interface display unit, which is used to perform the functions of step S6, that is, setting a user interface module to display the speed control ability score of the driver, the level evaluation result, personalized feedback suggestions, and the statistical analysis chart of historical driving data. The user interface module has good interactivity and visualization effects, which is convenient for the driver to intuitively understand their own speed control ability status and improvement direction, and the user interface module can adaptively adjust the display layout and content according to the needs of different users and device types.

[0080] The present invention provides a method and device for analyzing a driver's speed control ability based on driving behavior big data. Six core features, namely, training trajectory score, duration of speed-gear mismatch, number of positive and negative changes in steering wheel angle, start deceleration position, acceleration smoothness, and maximum absolute value of steering wheel rotation angle, are selected, and the gradient boosting tree (GBDT) and random forest (RF) models are fused to achieve efficient analysis of speed control ability; a multi-level evaluation mechanism is constructed by introducing a dynamic threshold and a Softmax function, and a speed control ability score and improvement suggestions are output; combined with a data enhancement strategy and low-precision sensor optimization, the generality and robustness of the algorithm are significantly improved. The present invention is suitable for a low-cost device environment and has the advantages of high computing efficiency, accurate evaluation, and low hardware dependence. The technical solution of the present invention can be applied to fields such as intelligent driving assistance systems, driver behavior monitoring, and intelligent vehicle systems, which helps to improve the safety, driving comfort, and energy-saving efficiency of drivers.

[0081] Through big data analysis and machine learning methods, the present invention can comprehensively evaluate the speed control ability of drivers and evaluate their driving behaviors from multiple dimensions. By analyzing various features such as training trajectory score, duration of speed-gear mismatch, and steering wheel angle change, the speed control ability of drivers can be evaluated more accurately, and personalized optimization suggestions can be provided for drivers. By outputting a probability distribution through the Softmax function, the evaluation result is more intuitive and can help drivers improve safety and driving comfort.

[0082] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for analyzing the speed control ability of drivers based on driving behavior big data, characterized in that: At least include the following steps: Step S1: Data collection and preprocessing step. Collect relevant data of the driver during driving through in-vehicle sensors, and process the collected data, including one or more of filling in missing data, smoothing noise, cleaning, denoising, and standardization processing; Step S2: Feature extraction step. Extract key features for analyzing the driver's speed control ability from the preprocessed data. The features include, but are not limited to, one or more features related to driving trajectory, speed gear matching, steering wheel operation, and acceleration; Step S3: Model construction and evaluation step. Use a machine learning model to process the extracted features, construct a model capable of evaluating the driver's speed control ability, and calculate the speed control ability score; Step S4: Dynamic evaluation and feedback step. Generate a level evaluation of the driver based on the calculation results and a specific function, and give personalized feedback suggestions in combination with the driving scenario.

2. The method according to claim 1, wherein: In step S1, it also includes: Step S11: The in-vehicle sensors collect data at a predetermined frequency. The collected data includes speed, gear, and steering wheel angle data; Step S12: Use the cubic spline interpolation algorithm to fill in missing data, use the Savitzky-Golay filter to smooth noise, and clean and denoise the data through data cleaning and denoising rules. Use the standardization method to convert the data into a standard form that meets the requirements of subsequent analysis.

3. The method according to claim 1, wherein: In step S2, it also includes: Step S21: The training trajectory score is determined by comparing the driver's actual driving path with the standard trajectory to evaluate the driver's driving performance on the specified path; Step S22: The duration of speed-gear mismatch is cumulatively calculated according to the speed-gear matching model of the vehicle, and the model considers the basic performance parameters of the vehicle; Step S23: The number of positive and negative changes in the steering wheel angle is obtained by counting the positive and negative changes in the steering wheel angle per unit time; Step S24: Calculate the position information of the vehicle when it enters the deceleration state from the stable state according to the law of speed change for the starting deceleration position; Step S25: The acceleration smoothness is calculated according to a specific formula, and the formula is: where Δa i = Δa i+1 - Δa i represents the acceleration change between consecutive time points, and a max is the absolute value maximum of the acceleration change. The smoothness range is [0, 1], and the higher the value, the smoother the acceleration. Step S26: Directly obtain the maximum value of the absolute value of the steering wheel angle for the maximum value of the steering wheel angle.

4. The method according to claim 1, characterized in that: In step S3, it also includes: Step S31: Use a model that combines Gradient Boosting Decision Tree (GBDT) and Random Forest (RF) for evaluation, and calculate the speed control ability score according to a specific weight combination formula; The model formula is S = 0.6·GBDT(x) + 0.4·RF(x), where x is the feature vector containing the extracted features, and S is the speed control ability score; Step S32: During the model training process, adjust and optimize the hyperparameters of the GBDT model and the RF model according to the distribution characteristics of the training data and the evaluation accuracy requirements to improve the model performance.

5. The method according to claim 1, characterized in that: In step S4, it also includes: Step S41: Calculate the driver level probability according to the formula based on the Softmax function; where z i is the score of the i-th level after the eigenvector is calculated by the model; Step S42: Comprehensively consider the road type, traffic conditions, and surrounding environment information of the current driving scenario, and combine the calculated level probability to generate targeted personalized feedback suggestions. The generation of the feedback suggestions also refers to the driver's historical driving data and the average performance of drivers of the same type.

6. The method according to any one of claims 1 to 5, characterized in that, It also includes: Step S5: Data storage and transmission step. Set up a data storage module to store the collected raw data, processed data, extracted features, model training process data, evaluation results, and feedback suggestions. At the same time, configure a data transmission module to achieve data interaction with external devices or systems, where the external devices or systems include intelligent driving assistance systems, traffic management department servers, and vehicle remote monitoring centers, so as to achieve data sharing and collaborative work.

7. The method according to any one of claims 1 to 5, characterized in that, It also includes: Step S6: User interface display step. Set up a user interface module to display the speed control ability score of the driver, level evaluation results, personalized feedback suggestions, and statistical analysis charts of historical driving data. The user interface module has good interactivity and visualization effects, which is convenient for the driver to intuitively understand their own speed control ability status and improvement direction, and the user interface module can adaptively adjust the display layout and content according to the needs of different users and device types.

8. Driving behavior big data-based driver speed control ability analysis device, characterized in that: This device is used to implement the method described in any one of claims 1 to 7, and includes: A data acquisition and preprocessing unit, which is used to execute the function of step S1 in claim 1, that is, collect relevant data of the driver during driving through in-vehicle sensors, and process the collected data, including but not limited to filling in missing data, smoothing noise, cleaning, denoising, and standardization processing; A feature extraction unit, which is used to execute the function of step S2 in claim 1, that is, extract key features for analyzing the driver's speed control ability from the preprocessed data, and these features include but not limited to features related to driving trajectory, speed gear matching, steering wheel operation, and acceleration; A model construction and evaluation unit, which is used to execute the function of step S3 in claim 1, that is, use a machine learning model to process the extracted features, construct a model that can evaluate the driver's speed control ability, and calculate the speed control ability score; A dynamic evaluation and feedback unit, which is used to execute the function of step S4 in claim 1, that is, generate a level evaluation of the driver based on the calculation results and a specific function, and give personalized feedback suggestions in combination with the driving scenario.

9. The device according to claim 8, wherein: It also includes: A data storage and transmission unit, which is used to execute the function of step S5 in claim 6, that is, set up a data storage module to store the collected raw data, processed data, extracted features, model training process data, evaluation results, and feedback suggestions. At the same time, configure a data transmission module to achieve data interaction with external devices or systems, where the external devices or systems include intelligent driving assistance systems, traffic management department servers, vehicle remote monitoring centers, etc., so as to achieve data sharing and collaborative work.

10. The device according to claim 8, characterized in that: It also includes: A user interface display unit, which is used to execute the function of step S6 in claim 7, that is, set up a user interface module to display the speed control ability score of the driver, level evaluation results, personalized feedback suggestions, and statistical analysis charts of historical driving data. The user interface module has good interactivity and visualization effects, which is convenient for the driver to intuitively understand their own speed control ability status and improvement direction, and the user interface module can adaptively adjust the display layout and content according to the needs of different users and device types.

Images