Method and system for optimizing steering wheel grip feeling of vehicle based on deep learning

Abstract

The application provides a deep learning-based automobile steering wheel grip feeling optimization method and system, relates to the technical field of grip feeling optimization, and comprises the following steps: collecting road surface data of a driving road surface according to an integrated sensor to obtain road surface data; performing road condition analysis on fused road surface data obtained by data fusion on the road surface data to obtain a road condition; generating an adjustment strategy through the road condition, actively warning according to the road condition, generating a grip warning on a preset grip mode, and obtaining a warning grip parameter; adjusting a steering wheel parameter according to the warning grip parameter and obtaining user feedback information through interaction; evaluating the warning effect on the record, and performing integrated warning. Through the application, the technical problem that the existing technology cannot adjust the steering wheel feedback in real time and accurately according to complex road conditions, which further affects the perception of the driver on the road condition, can be solved, and the technical effect of enhancing the perception of the driver on the road condition is achieved.

Description

Technical Field

[0001] This application relates to the field of grip optimization technology, and in particular to a method and system for optimizing the grip of automotive steering wheels based on deep learning. Background Technology

[0002] With the continuous development of intelligent driving technology, optimizing the grip of the steering wheel has become an important direction for improving the driving experience and safety. Traditional steering wheel feedback systems usually rely on mechanical structures to provide fixed force feedback and vibration effects. However, with the diversification of driving needs and road conditions, this fixed feedback method is difficult to meet the safety warning needs of different driving scenarios.

[0003] Currently, existing electronic force feedback systems have low accuracy and sensitivity in responding to changes in road surface conditions, and cannot adjust steering wheel feedback in real time and accurately according to complex road conditions. For example, on wet, muddy, or rough roads, existing electronic force feedback systems may fail to enhance steering wheel feedback or vibration feedback in a timely manner, resulting in the driver lacking necessary road information perception and thus increasing driving hazards.

[0004] In summary, existing technologies suffer from the technical problem that electronic force feedback systems have low accuracy and sensitivity in responding to changes in road surface conditions. This results in the inability to adjust steering wheel feedback in real time and accurately according to complex road conditions, further affecting the driver's perception and reaction to road conditions and increasing the danger of driving. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for optimizing the grip of automotive steering wheels based on deep learning, in order to solve the technical problem in the prior art where the electronic force feedback system has low accuracy and sensitivity in response to changes in road surface, which makes it impossible to adjust the steering wheel feedback in real time and accurately according to complex road conditions, further affecting the driver's perception and reaction to road conditions and increasing the danger of driving.

[0006] In view of the above problems, this application provides a method and system for optimizing the grip feel of automotive steering wheels based on deep learning.

[0007] Firstly, this application provides a deep learning-based method for optimizing the grip feel of a car steering wheel, implemented through a deep learning-based car steering wheel grip feel optimization system. The method includes: collecting road surface data from integrated sensors to obtain road surface data; performing road condition analysis on fused road surface data obtained by data fusion of the road surface data to obtain road surface conditions; generating an adjustment strategy based on the road surface conditions to provide proactive safety warnings and generate grip warnings for preset grip modes, obtaining warning grip parameters; adjusting steering wheel parameters according to the warning grip parameters and interacting to obtain user feedback; evaluating the perception of the grip warning based on the user feedback information, obtaining real-time warning response records, combining the perception with the real-time warning response records to evaluate the warning effect, using the warning effect as the basis for integration activation, obtaining supplementary safety warning functions, adding the supplementary safety warning functions to the grip warning for integrated warning, and executing the integrated warning.

[0008] Secondly, this application also provides a deep learning-based automotive steering wheel grip optimization system for executing the deep learning-based automotive steering wheel grip optimization method described in the first aspect, comprising: a data acquisition module for collecting road surface data based on integrated sensors to obtain road surface data; a road condition analysis module for performing road condition analysis based on fused road surface data obtained by data fusion of the road surface data to obtain road surface conditions; a grip optimization module for generating an adjustment strategy based on the road surface conditions, providing active safety warnings based on the road surface conditions, and generating grip warnings for preset grip modes to obtain warning grip parameters; a parameter adjustment module for adjusting steering wheel parameters based on the warning grip parameters and interacting to obtain user feedback information; and a closed-loop control module for evaluating the perception of the grip warning based on the user feedback information, obtaining real-time warning response records, evaluating the warning effect by combining the perception and the real-time warning response records, using the warning effect as the basis for integration activation, obtaining supplementary safety warning functions, adding the supplementary safety warning functions to the grip warning for warning integration, and executing integrated warning.

[0009] The technical solution provided in this application has at least the following technical effects or advantages: Road surface data is obtained by collecting road surface data from integrated sensors; road condition analysis is performed on the fused road surface data obtained by data fusion of the road surface data to obtain road conditions; an adjustment strategy is generated based on the road conditions to provide active safety warnings and generate grip warnings for preset grip modes, obtaining warning grip parameters; steering wheel parameters are adjusted according to the warning grip parameters, and user feedback information is obtained through interaction; the perception of the grip warning is evaluated based on the user feedback information, real-time warning response records are obtained, and the warning effect is evaluated by combining the perception and the real-time warning response records; the warning effect is used as the basis for integrated activation to obtain supplementary safety warning functions; these supplementary safety warning functions are added to the grip warning for integrated warning, and integrated warning is executed. In other words, by implementing a dynamic adjustment system based on real-time road surface data, the force and vibration feedback of the steering wheel can be automatically adjusted precisely according to different road conditions, improving the response accuracy and sensitivity to road changes, thereby enhancing the driver's perception of road conditions, improving control precision, and enhancing driving safety.

[0010] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the deep learning-based method for optimizing the grip of a car steering wheel in this application.

[0013] Figure 2 This is a schematic diagram of the structure of the deep learning-based automotive steering wheel grip optimization system of this application.

[0014] Explanation of reference numerals in the attached figures:

[0015] Data acquisition module 11, road condition analysis module 12, grip optimization module 13, parameter adjustment module 14, closed-loop control module 15. Detailed Implementation

[0016] This application provides a deep learning-based method and system for optimizing the grip of automotive steering wheels. It addresses the technical problem in existing technologies where the low accuracy and sensitivity of electronic force feedback systems in responding to road surface changes prevent real-time and accurate adjustments to steering wheel feedback based on complex road conditions. This, in turn, negatively impacts the driver's perception and reaction to road conditions, increasing driving hazards. The application implements a dynamic adjustment system based on real-time road data, which can automatically and precisely adjust the force and vibration feedback of the steering wheel according to different road conditions. This improves the accuracy and sensitivity of response to road surface changes, thereby enhancing the driver's perception of road conditions, improving control precision, and increasing driving safety.

[0017] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0018] Example 1, please refer to the appendix. Figure 1 This application provides a deep learning-based method for optimizing the grip feel of automotive steering wheels, which is applied to a deep learning-based automotive steering wheel grip feel optimization system. The method includes the following steps:

[0019] Step 1: Collect road surface data based on integrated sensors to obtain road surface data.

[0020] Specifically, collecting road surface data based on integrated sensors first requires acquiring various road-related data through multiple sensors. For example, wheel speed sensors monitor changes in wheel rotation speed to determine if there is slippage or insufficient friction; temperature sensors measure road surface temperature to help assess the risk of icing; acceleration sensors detect vehicle vibrations; significant vibrations may indicate uneven road surfaces or obstacles; and cameras capture real-time images of the road ahead to analyze for obstacles such as water accumulation, mud, or snow. This sensor data collection provides detailed information about the current road surface conditions, supporting subsequent road surface recognition and feedback adjustments. For instance, if a temperature sensor detects an ambient temperature below zero degrees Celsius, and an acceleration sensor detects significant road surface vibrations, combined with wheel speed sensor slippage data, it can be inferred that the road surface ahead is icy and uneven.

[0021] Step 2: Based on the fused road surface data obtained by data fusion of the road surface data, road condition analysis is performed to obtain the road surface condition.

[0022] Specifically, data from different sensors is integrated to complement each other, providing more accurate and comprehensive information. For example, wheel speed sensors provide data on wheel slippage, acceleration sensors provide vibration information, and cameras capture water accumulation or obstacles on the road surface. After data fusion, this information can be considered simultaneously to obtain comprehensive road surface data. The fused data is called fused road surface data, which includes multi-dimensional information from various sensors. Next, road condition analysis is performed on the fused road surface data. Based on the integrated data, the current road surface condition is inferred by analyzing features such as wheel slippage, vibration frequency, temperature changes, and road texture. For example, if the wheel slippage rate is high, the acceleration sensor detects strong vibrations, and the temperature sensor shows a low temperature, it may indicate that there is icy road ahead. By analyzing the current road surface condition, such as wet, slippery, icy, muddy, or flat and dry, the vehicle's control strategy can be adjusted.

[0023] Step 3: Generate adjustment strategies based on the road conditions, provide proactive safety warnings based on the road conditions, generate grip warnings for preset grip modes, and obtain warning grip parameters.

[0024] Specifically, road conditions may include various states such as wet / slippery, dry, icy, or muddy, which can affect vehicle handling and steering wheel feedback. Based on the identified road conditions, corresponding adjustment schemes are developed to dynamically adjust the steering wheel's response force, feedback frequency, and vibration intensity according to different road types, reminding the driver to slow down. For example, on wet / slippery roads, steering wheel resistance and vibration feedback are increased to alert the driver to road conditions; while on flat, dry roads, steering wheel feedback force is reduced to provide a more relaxed driving experience. Active safety warnings are issued based on road conditions; for example, the steering wheel vibrates to remind the driver to pay attention to the turn at curves. Next, grip warnings are issued for preset grip modes. For example, during long-distance driving, the steering wheel vibrates to remind the driver to maintain a comfortable grip posture. By adjusting these feedback parameters, the final warning grip parameters are obtained, including the ideal feedback intensity, sensitivity, and vibration frequency of the steering wheel under different road conditions and driving modes, ensuring the driver receives the best driving experience in various situations.

[0025] Step 4: Adjust the steering wheel parameters according to the warning grip parameters and interact to obtain user feedback information.

[0026] Specifically, the system adjusts steering wheel parameters based on warning grip parameters, including steering wheel resistance, vibration intensity, and feedback force, to provide the driver with the optimal driving experience under different road conditions and driving modes. For example, warning grip parameters might require increased steering wheel resistance to allow the driver to better perceive the vehicle's driving status, while reduced resistance might make driving easier. The adjusted steering wheel parameters are then interactively communicated to the driver. The user perceives these changes through feel, vibration, and steering wheel feedback force, and relays these feedbacks, resulting in user feedback information. This feedback includes the driver's level of perception of road condition warnings and whether more or other warning methods are needed.

[0027] Step 5: Evaluate the perception of the grip warning based on the user feedback information, obtain real-time warning response records, combine the perception with the real-time warning response records to evaluate the warning effect, use the warning effect as the basis for integration activation, obtain supplementary safety warning functions, add the supplementary safety warning functions to the grip warning for warning integration, and execute integrated warning.

[0028] Specifically, the assessment is based on user feedback to evaluate the driver's perception of grip warnings, that is, whether the driver can promptly notice and understand warning signals such as vibrations or force feedback from the steering wheel. For example, if the steering wheel emits continuous vibrations as a reminder while driving on a slippery road, can the driver slow down in time?

[0029] Next, the system obtains real-time records of the driver's responses after receiving the warning signal, such as whether the driver slowed down or adjusted the direction. For example, if the driver promptly reduced speed and adjusted the steering wheel after receiving a vibration alert from the steering wheel at a curve, it indicates that the warning signal effectively alerted the driver and prevented a potential traffic accident.

[0030] Then, the effectiveness of the warning is evaluated by combining the perception of the grip warning with the driver's real-time response record. For example, if the driver can slow down in time after receiving the vibration reminder of the steering wheel multiple times, it means that the warning signal has effectively improved the driver's safety awareness and reduced the probability of traffic accidents.

[0031] Finally, based on the effectiveness of the warnings, a decision is made on whether to activate supplementary safety warning functions, such as lane departure warning or collision warning, and to integrate these functions with grip warning to form a more comprehensive warning system, providing drivers with more comprehensive safety protection. For example, if a driver receives not only a vibration alert from the steering wheel but also a lane departure warning while driving on a slippery road, they can more effectively avoid losing control of the vehicle or deviating from their lane, thereby improving driving safety.

[0032] The deep learning-based automotive steering wheel grip optimization method is applied to a deep learning-based automotive steering wheel grip optimization system. It can realize a dynamic adjustment system based on real-time road data, and can accurately and automatically adjust the force feedback and vibration feedback of the steering wheel according to different road conditions. This improves the response accuracy and sensitivity to road changes, thereby enhancing the driver's perception of road conditions, improving control precision, and enhancing driving safety.

[0033] Furthermore, this application also includes: extracting the collection timestamp of the road surface data, aligning the collection timestamps to obtain synchronized road surface data under the aligned timestamps; extracting features from the standardized synchronized road surface data to obtain road surface features; and fusing the road surface features to obtain the fused road surface data.

[0034] Specifically, the time point information for each data set is extracted from the road surface data collected by various sensors. A timestamp is a number that marks the exact moment the data was collected. Since different sensors may have different collection times, multiple collection timestamps are aligned, that is, the timestamps of all sensors are unified to the same moment, ensuring that the data from all sensors are synchronized. Under the aligned timestamps, synchronized road surface data is obtained; with all sensor data collected at the same time point, the corresponding road surface information is accurately matched and compared.

[0035] After synchronizing and aligning the data, the data is standardized to ensure they are on the same scale. All data is transformed into a uniform range or standard to avoid any data points having an excessive impact on the analysis results due to differing numerical ranges. Following standardization, feature extraction is performed on the synchronized data to extract useful information from the raw data. For example, features such as vibration intensity and frequency can be extracted from accelerometer data, or information such as road texture and obstacle type can be extracted from camera data. Feature extraction allows for a better understanding of road conditions.

[0036] This involves integrating and merging various feature data obtained from different sensors and feature extraction methods. Data fusion can improve the accuracy of analysis because different sensors and features can complement each other, providing more diverse information. For example, by fusing the slip ratio extracted by the wheel speed sensor and the temperature information extracted by the temperature sensor, it is possible to more accurately determine whether the road surface is wet or icy. Feature fusion methods can include weighted averaging, deep learning, etc., and the final product is fused road surface data that can better reflect complex road conditions.

[0037] By extracting and aligning the timestamps of the sensors, all data are synchronized at the same time, thus accurately reflecting the road surface information at the same moment. Then, these synchronized data are standardized, and useful features of each sensor, such as vibration intensity and road surface texture, are extracted. Finally, by fusing these features from different sources, a comprehensive road surface data is obtained, which can more accurately describe the current road surface condition, thus providing a reliable basis for subsequent analysis and decision-making.

[0038] Furthermore, this application also includes: identifying the fused road surface data based on a classification model to obtain the road surface type; calculating the friction coefficient of the road surface type through road surface friction mapping; performing coverage settlement based on the friction coefficient at the collection timestamp to obtain a friction coefficient sequence; and performing road condition analysis based on the friction coefficient sequence to obtain the road surface condition.

[0039] Specifically, classification models are methods that classify new data by learning patterns from existing data. By analyzing and fusing different features contained in road surface data, such as wheel speed, acceleration, vibration intensity, and road texture, classification models can determine the current road surface type.

[0040] The friction coefficient of a road surface is calculated using road surface friction mapping. The friction coefficient is a physical quantity that measures the magnitude of the frictional force between the road surface and the vehicle's wheels. Different types of road surfaces have different friction coefficients; a higher friction coefficient results in stronger grip between the wheels and the road surface, making it easier for the vehicle to maintain stability, while a lower coefficient makes it easier to slip. Friction mapping allows the calculation of the corresponding friction coefficient based on the identified road surface type. For example, icy and snowy roads typically have a lower friction coefficient, while dry highways have a higher coefficient. By comparing the road surface type with known friction coefficient data, the friction coefficient of the current road surface can be calculated.

[0041] The friction coefficient is overlaid and settled based on the data collection timestamps to obtain a friction coefficient sequence. Overlay settlement refers to aligning the calculated friction coefficients with the data collection timestamps and generating a corresponding friction coefficient for each timestamp. For example, if the road surface type at a certain moment is slippery, the friction coefficient at that moment is calculated through friction mapping and associated with the corresponding timestamp. By performing this operation at different timestamps, a friction coefficient sequence can be obtained, which is a series of friction coefficient data that varies with time, clearly depicting how road surface friction changes over time.

[0042] A friction coefficient sequence reflects changes in road surface friction characteristics over different time periods. For example, if the friction coefficient remains consistently low for a period, it indicates the presence of ice or water accumulation on the road surface. Analysis of the friction coefficient sequence allows for real-time assessment of whether the road surface is safe for driving or whether safety measures such as reducing speed are necessary. For instance, if the value in the friction coefficient sequence falls below a certain threshold, it indicates that the road surface is slippery or icy, thus alerting the driver to be more vigilant.

[0043] By identifying the fused road surface data through a classification model, the current road surface type can be accurately determined. Based on the road surface type, the corresponding friction coefficient is calculated through friction mapping, reflecting the friction force between different road surfaces and wheels. The friction coefficient is aligned with the collected timestamps to generate a friction coefficient sequence. Road condition analysis is performed based on the friction coefficient sequence to assess the current road surface condition in real time, thereby providing drivers with accurate road information, helping to optimize driving strategies, and ensuring driving safety.

[0044] Furthermore, this application also includes: aggregating wheel speed samples, acceleration samples, gyroscope samples, road visual samples, and environmental samples from a preset historical time period based on a preset sample threshold to obtain sample groups; encoding the road type of the sample groups to obtain road type sample groups; and training the classification model based on the wheel speed samples, acceleration samples, gyroscope samples, road visual samples, environmental samples, and road type sample groups.

[0045] Specifically, based on a preset sample threshold, wheel speed samples, acceleration samples, gyroscope samples, road vision samples, and environmental samples from a preset historical time period are aggregated to obtain sample groups. The preset sample threshold is a standard value set during data acquisition to determine the clustering criteria, and the sample data is clustered according to these criteria. For example, if the wheel speed variation range is within a certain value, similar samples are grouped together according to certain rules to form representative sample groups. These sample data may involve different sensors, such as wheel speed sensors, acceleration sensors, gyroscopes, road vision sensors, and environmental sensors.

[0046] Road surface type classification is performed on the aggregated sample data. Road surface type encoding refers to converting different road surface conditions, such as dry, wet, and icy / snowy, into a machine-recognizable coded form. For example, a wet / slippery road surface might be encoded as "1", an icy / snowy road surface as "2", and a flat, dry road surface as "0". The encoding process can be implemented through manual annotation or machine learning models. Based on features (such as acceleration, wheel speed, etc.) in the aggregated samples, the current road surface type is automatically determined, and a corresponding code is assigned to each sample. The road surface type sample group is a dataset containing road surface type codes, which facilitates subsequent model training and analysis.

[0047] A classification model is trained using samples from wheel speed, acceleration, gyroscope data, road surface vision, environmental data, and road surface type samples. Sample training involves inputting pre-labeled sample data into the machine learning model, which then learns how to determine the road surface type based on input features (such as wheel speed, acceleration, and road surface vision). The training process continuously adjusts the parameters of the machine learning model, enabling it to accurately predict road surface types when encountering new samples. For example, wheel speed samples may contain features of different wet and dry road surfaces; the machine learning model learns the differences between these two types during training. Ultimately, the trained classification model can predict the corresponding road surface type based on new input samples, providing real-time road surface classification information for autonomous driving or intelligent vehicle systems.

[0048] By aggregating various sensor data into sample groups using a preset sample threshold, and then encoding these groups according to road surface type, the samples are classified according to road surface type. Labeled samples are then used to train a classification model, enabling the model to learn how to predict road surface type based on different sensor data. Through this process, the current road surface type can be identified based on real-time sensor data, thus providing drivers with more accurate road condition information.

[0049] Furthermore, this application also includes: arranging the collected timestamps sequentially to obtain a timestamp sequence; and calculating the friction coefficient based on the fused road surface data corresponding to the timestamp sequence to obtain the friction coefficient sequence.

[0050] Specifically, a data acquisition timestamp refers to the specific time marker at which each data point was acquired. Data from all sensors is timestamped and arranged in ascending order of acquisition time to form a timestamp sequence. Because different sensors may acquire data at different times, organizing this data in precise chronological order ensures the timeliness and synchronization of the data, thereby generating an ordered timestamp sequence.

[0051] The coefficient of friction is a physical quantity used to describe the magnitude of friction between the road surface and the vehicle's wheels. Different road surface types have different coefficients of friction. By analyzing fused road surface data from a time-stamped sequence, the coefficient of friction at each time point is calculated. The fused road surface data includes data from various sensors, such as wheel speed, acceleration, and visual information, providing a comprehensive view of the road surface conditions. Based on the fused road surface data, combined with factors such as road surface type, vehicle motion, and external environment, the coefficient of friction is calculated. For example, the coefficient of friction is lower on wet roads and higher on dry roads. The calculated coefficients of friction are combined in the order corresponding to their timestamps to obtain a friction coefficient sequence, reflecting the changes in road surface friction at different time points.

[0052] By arranging the collected timestamps sequentially, the data from all sensors are arranged in chronological order to form a timestamp sequence. Based on these timestamp sequences and the corresponding fused road surface data, the friction coefficient at each time point is calculated, resulting in a friction coefficient sequence. This allows for tracking changes in road surface friction, providing crucial information for subsequent road condition analysis and driving decisions.

[0053] Furthermore, this application also includes: identifying road surface condition changes in the friction coefficient sequence based on a preset friction change threshold to obtain a road surface condition series; configuring a preset error time period, performing time series analysis on the road surface condition series within the preset error time period to obtain the road surface condition at the error time; and combining the road surface condition series and the road surface condition at the error time to obtain the road surface condition.

[0054] Specifically, road condition changes are identified based on a preset friction change threshold, resulting in a road condition series. The preset friction change threshold is a pre-defined standard used to determine if the change in friction coefficient is sufficiently large to identify whether the road condition or type has changed. For example, if the change in friction coefficient is significant and exceeds the preset friction change threshold, the road condition is considered to have changed, potentially from a dry surface to a slippery surface. Based on the changes in friction coefficient, different stages of road condition are identified and converted into a road condition series.

[0055] A preset error time period is configured, and time series analysis is performed on the road surface condition series within this preset error time period to obtain the road surface condition at the error time. The preset error time period refers to the range of time differences that the system allows for, used to predict road surface conditions in the future. The preset error time period is customized by those skilled in the art based on actual conditions. For example, due to sensor response speed or data processing delays, changes in the friction coefficient may lag behind changes in actual road surface conditions. The preset error time period provides a margin of error for predicting the road surface conditions of subsequent road segments, allowing adjustments to road surface condition changes within a certain timeframe. Within the preset error time period, combined with time series analysis methods, the changes in the road surface condition series are predicted, further confirming the trend of road surface condition changes and obtaining road surface condition prediction information for the future period.

[0056] By combining road surface condition series and error time road surface condition, the changes in friction coefficient and the adjustment of error time period can be comprehensively considered, thereby more accurately judging the road surface condition at the current time and for a period of time in the future.

[0057] By analyzing the friction coefficient sequence using a preset friction change threshold, changes in road conditions are identified, and a road condition series is generated. A preset error time period is configured to analyze the road conditions in the future, resulting in the road condition at the error time. By combining the road condition series with the road condition at the error time, the accuracy of road condition identification is further improved, ensuring that the system can correctly reflect the current road condition. This enables the system to identify road changes in a timely and accurate manner during actual driving, providing drivers with reliable road condition information.

[0058] Furthermore, this application also includes: identifying dangerous road conditions to obtain a hazard warning, and assessing the hazard coefficient based on the hazard warning; extracting force feedback parameters and vibration feedback parameters according to the preset grip mode based on the hazard coefficient; adjusting the force feedback parameters by extracting the force adjustment strategy of the adjustment strategy to obtain a force response setpoint; adjusting the vibration feedback parameters by extracting the vibration feedback parameters of the adjustment strategy to obtain a vibration response setpoint; and obtaining the warning grip parameter by combining the force response setpoint and the vibration response setpoint.

[0059] Specifically, the system analyzes the collected road conditions to identify potential hazards, such as slippery surfaces, curves, and potholes. If the system detects a slippery road surface, it will issue a slippery road warning.

[0060] Next, based on the hazard warning assessment hazard coefficient, the degree of danger of the current road surface condition is evaluated. For example, a slippery road surface may be assessed as moderately dangerous, while a curve may be assessed as highly dangerous.

[0061] Then, force feedback parameters and vibration feedback parameters are extracted based on the preset grip mode. The preset grip mode refers to the grip requirements pre-set during the design process according to different driving needs, such as comfort, sport, or economy modes. In the preset grip mode, the force and vibration feedback of the steering wheel will differ. Force feedback parameters refer to the resistance or pressure felt by the driver when operating the steering wheel; vibration feedback parameters refer to the vibration perception generated on the steering wheel, used to identify obstacles, bumps, or road friction. By using the preset grip mode, these two types of feedback parameters related to driving needs can be extracted to prepare for the next adjustment. For example, on a slippery road surface, heavier force feedback parameters and continuous vibration feedback parameters are extracted to remind the driver to slow down and maintain a stable grip posture.

[0062] The force adjustment strategy, derived from the extraction and adjustment mechanism, adjusts the force feedback parameters to obtain the force response setpoint. The adjustment strategy is a scheme that optimizes feedback based on current road conditions and driving mode. Specifically, the force adjustment strategy adjusts the force feedback parameters on the steering wheel to adapt to different driving conditions. For example, on a slippery road, the driver needs higher force feedback to enhance driving stability; while on a smooth highway, the force feedback can be appropriately reduced to make driving easier. By adjusting the force feedback parameters in this way, a force response setpoint is obtained, which is the ideal resistance value that the steering wheel should provide under the current conditions.

[0063] The vibration feedback parameters of the adjustment strategy are extracted and adjusted to obtain the vibration response setpoint. Adjusting the vibration feedback parameters refers to adapting the vibration intensity to changes in road conditions. For example, when an uneven road surface is detected, the vibration feedback intensity increases to alert the driver to poor road conditions; while on a smooth road surface, the vibration feedback decreases to make the driving experience smoother. By adjusting the vibration feedback parameters, the ideal intensity of the vibration feedback is determined according to different road conditions and driving modes, ultimately yielding the vibration response setpoint.

[0064] By combining the force response setting and vibration response setting, the warning grip parameters are obtained. The warning grip parameters refer to the optimal combination of force and vibration feedback that the steering wheel should provide under the current road conditions and driving mode, ensuring that the steering wheel's handling experience meets the driver's needs. For example, it provides stronger force feedback and vibration in Sport mode, while providing a more relaxed and softer grip in Comfort mode.

[0065] In summary, the deep learning-based automotive steering wheel grip optimization method provided in this application has the following technical effects: Road surface data is obtained by collecting road surface data from integrated sensors; road condition analysis is performed on the fused road surface data obtained by data fusion of the road surface data to obtain road conditions; an adjustment strategy is generated based on the road conditions to provide active safety warnings and generate grip warnings for preset grip modes, obtaining warning grip parameters; steering wheel parameters are adjusted according to the warning grip parameters, and user feedback is obtained through interaction; the perception of the grip warning is evaluated based on the user feedback information, and real-time warning response records are obtained; the warning effect is evaluated by combining the perception and the real-time warning response records; the warning effect is used as the basis for integration activation to obtain supplementary safety warning functions; these supplementary safety warning functions are added to the grip warning for integrated warning, and the integrated warning is executed. In other words, by implementing a dynamic adjustment system based on real-time road surface data, the force and vibration feedback of the steering wheel can be automatically adjusted precisely according to different road conditions, improving the response accuracy and sensitivity to road changes, thereby enhancing the driver's perception of road conditions, improving control precision, and enhancing driving safety.

[0066] Example 2: Based on the same inventive concept as the deep learning-based automotive steering wheel grip optimization method described in the previous examples, this application also provides a deep learning-based automotive steering wheel grip optimization system. Please refer to the appendix. Figure 2The system includes: a data acquisition module 11, which collects road surface data based on integrated sensors to obtain road surface data; a road condition analysis module 12, which performs road condition analysis based on fused road surface data obtained by data fusion of the road surface data to obtain road surface conditions; a grip optimization module 13, which generates an adjustment strategy based on the road surface conditions, provides active safety warnings based on the road surface conditions, and generates grip warnings for preset grip modes to obtain warning grip parameters; a parameter adjustment module 14, which adjusts steering wheel parameters based on the warning grip parameters and interacts to obtain user feedback information; and a closed-loop control module 15, which evaluates the perception of the grip warning based on the user feedback information, obtains real-time warning response records, evaluates the warning effect by combining the perception and the real-time warning response records, uses the warning effect as the basis for integration activation, obtains supplementary safety warning functions, adds the supplementary safety warning functions to the grip warning for warning integration, and executes integrated warnings.

[0067] Furthermore, the deep learning-based automotive steering wheel grip optimization system is also used for: extracting the collection timestamp of the road surface data, aligning the collection timestamps to obtain synchronized road surface data under the aligned timestamps; extracting features from the standardized synchronized road surface data to obtain road surface features; and fusing the road surface features to obtain the fused road surface data.

[0068] Furthermore, the deep learning-based automotive steering wheel grip optimization system is also used for: identifying the fused road surface data based on a classification model to obtain the road surface type; calculating the friction coefficient of the road surface type through road surface friction mapping; performing coverage calculation based on the friction coefficient at the collection timestamp to obtain a friction coefficient sequence; and performing road condition analysis based on the friction coefficient sequence to obtain the road surface condition.

[0069] Furthermore, the deep learning-based automotive steering wheel grip optimization system is also used for: aggregating wheel speed samples, acceleration samples, gyroscope samples, road visual samples, and environmental samples from a preset historical time period based on a preset sample threshold to obtain sample groups; encoding the road type of the sample groups to obtain road type sample groups; and training the classification model based on the wheel speed samples, acceleration samples, gyroscope samples, road visual samples, environmental samples, and road type sample groups.

[0070] Furthermore, the deep learning-based automotive steering wheel grip optimization system is also used to: sequentially arrange the collected timestamps to obtain a timestamp sequence; and calculate the friction coefficient based on the fused road surface data corresponding to the timestamp sequence to obtain the friction coefficient sequence.

[0071] Furthermore, the deep learning-based automotive steering wheel grip optimization system is also used for: identifying road condition changes in the friction coefficient sequence based on a preset friction change threshold to obtain a road condition series; configuring a preset error time period, performing time series analysis on the road condition series within the preset error time period to obtain the road condition at the error time; and combining the road condition series and the road condition at the error time to obtain the road condition.

[0072] Furthermore, the deep learning-based automotive steering wheel grip optimization system is also used for: identifying dangerous road conditions to obtain a hazard warning; assessing the hazard coefficient based on the hazard warning; extracting force feedback parameters and vibration feedback parameters based on the preset grip mode and the hazard coefficient; adjusting the force feedback parameters using the force adjustment strategy of the adjustment strategy to obtain a force response setpoint; adjusting the vibration feedback parameters using the vibration feedback parameters of the adjustment strategy to obtain a vibration response setpoint; and obtaining the warning grip parameter by combining the force response setpoint and the vibration response setpoint.

[0073] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The deep learning-based automotive steering wheel grip optimization method and specific examples in the aforementioned Embodiment 1 are also applicable to the deep learning-based automotive steering wheel grip optimization system of this embodiment. Through the foregoing detailed description of the deep learning-based automotive steering wheel grip optimization method, those skilled in the art can clearly understand the deep learning-based automotive steering wheel grip optimization system of this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0074] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0075] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for optimizing the grip feel of a car steering wheel based on deep learning, characterized in that, include: Road surface data is obtained by collecting road surface data from integrated sensors. Road condition analysis is performed on the fused road data obtained by data fusion of the road data to obtain the road condition. The road condition is used to generate an adjustment strategy, which provides active safety warnings based on the road condition and generates a grip warning for a preset grip mode, thereby obtaining the warning grip parameters. Adjust the steering wheel parameters according to the warning grip parameters and obtain user feedback information through interaction; The perception of the grip warning is evaluated based on the user feedback information, real-time warning response records are obtained, the warning effect is evaluated by combining the perception and the real-time warning response records, and the warning effect is used as the basis for integration. Supplementary safety warning functions are obtained, and the supplementary safety warning functions are added to the grip warning for warning integration, and integrated warning is executed. The fused road surface data includes: Extract the collection timestamp of the road surface data, and perform timestamp alignment on the collection timestamps to obtain synchronized road surface data under the aligned timestamps; Feature extraction is performed on the standardized synchronous pavement data to obtain pavement features; The road surface features are fused to obtain the fused road surface data; Obtain road conditions, including: The road surface type is obtained by identifying the fused road surface data based on a classification model. The friction coefficient of the road surface type is calculated using road surface friction mapping. The friction coefficient sequence is obtained by overwriting the data with the collected timestamp based on the friction coefficient. Road condition analysis is performed based on the friction coefficient sequence to obtain the road surface condition; Based on a preset friction change threshold, the friction coefficient sequence is used to identify road condition changes, resulting in a road condition series. Configure a preset error time period, and perform time series analysis on the road surface condition series within the preset error time period to obtain the road surface condition at the error time. The road surface condition is obtained by combining the road surface condition series and the road surface condition at the error time.

2. The method for optimizing the grip feel of a car steering wheel based on deep learning as described in claim 1, characterized in that, The classification model is obtained, including: Based on a preset sample threshold, wheel speed samples, acceleration samples, gyroscope samples, road visual samples, and environmental samples from a preset historical time are aggregated to obtain sample groups; The road surface type is encoded by the sample group to obtain the road surface type sample group; The classification model is obtained by training samples based on the wheel speed samples, acceleration samples, gyroscope samples, road surface visual samples, environmental samples, and road surface type samples.

3. The method for optimizing the grip feel of a car steering wheel based on deep learning as described in claim 1, characterized in that, The friction coefficient sequence is obtained, including: The collected timestamps are arranged sequentially to obtain a timestamp sequence; Friction coefficients are calculated based on the fused road surface data corresponding to the timestamp sequence to obtain the friction coefficient sequence.

4. The method for optimizing the grip feel of a car steering wheel based on deep learning as described in claim 1, characterized in that, The warning grip parameters are obtained, including: Hazard identification is performed on road conditions to obtain hazard warnings, and the hazard coefficient is assessed based on the hazard warnings; Based on the risk factor, force feedback parameters and vibration feedback parameters are extracted according to the preset grip mode; The force adjustment strategy of the adjustment strategy is extracted and the force feedback parameter is adjusted to obtain the force response setpoint; Extract the vibration feedback parameters of the adjustment strategy and adjust the vibration feedback parameters to obtain the vibration response setpoint; The warning grip parameters are obtained by combining the force response setting value and the vibration response setting value.

5. A deep learning-based automotive steering wheel grip optimization system, characterized in that, The steps for implementing the deep learning-based automotive steering wheel grip optimization method according to any one of claims 1 to 4 include: The data acquisition module is used to collect road surface data based on integrated sensors to obtain road surface data. The road condition analysis module is used to perform road condition analysis based on the fused road surface data obtained by data fusion of the road surface data, and to obtain the road surface condition. The grip optimization module is used to generate an adjustment strategy based on the road conditions, provide proactive safety warnings based on the road conditions, generate grip warnings for preset grip modes, and obtain warning grip parameters. A parameter adjustment module is used to adjust the steering wheel parameters according to the warning grip parameters and to obtain user feedback information through interaction. The closed-loop control module is used to evaluate the perception of the grip warning based on the user feedback information, obtain real-time warning response records, evaluate the warning effect by combining the perception and the real-time warning response records, and initiate the integrated warning based on the warning effect. It then obtains a supplementary safety warning function, adds the supplementary safety warning function to the grip warning for warning integration, and executes the integrated warning.

Images