Simulation method and system for fatigue detection of steeringwheel grip force based on dil

Abstract

The invention provides a simulation method and system for fatigue detection of steering wheel grip force based on DIL, which belongs to the field of automobile auxiliary driving and driving safety technology, including the construction of driver-in-the-loop system platform and virtual driving scene library, the integration of hardware and software to realize system construction, the construction of fatigue driving detection module, the design of experimental scheme, the establishment of fatigue driving sample data set, the construction of fatigue state recognition model, and finally the verification of experimental results, so as to comprehensively evaluate and optimize the fatigue driving detection system. The invention adopts the above-mentioned simulation method and system for fatigue detection of steering wheel grip force based on DIL, which improves the accuracy of fatigue detection, avoids the danger of collecting driving data under actual driving conditions, and improves the efficiency of driving experiment data collection. (Figure 1)

Description

A Steering Wheel Grip Fatigue Detection Simulation Method and System Based on DIL Technical Field

[0001] This invention relates to the field of automotive driver assistance and driving safety technology, and in particular to a method and system for simulating steering wheel grip fatigue detection based on DIL. Background Technology

[0002] Existing vision-based fatigue detection technologies are susceptible to environmental factors such as changes in lighting and driver posture. Most research on physiological signal-based detection technologies still requires wearing complex devices, which are not only cumbersome to wear but may also cause unnecessary interference to the driver's driving experience and may even affect driving safety to some extent.

[0003] While existing methods for fatigue detection based on steering wheel grip force have achieved some research results, they face numerous challenges in practical application. The most prominent issue is the strong subjectivity of fatigue state labeling, which can lead to significant discrepancies in how different researchers or labelers assess the fatigue level of the same driver. Furthermore, due to the complexity of steering wheel grip force data, accurately extracting fatigue-related features from this data and constructing a high-precision fatigue recognition algorithm are also key problems that urgently need to be solved in current research.

[0004] With the continuous development of DIL (Driver-In-the-Loop) technology, by combining virtual driving scenario modeling software with the actual driving steering wheel, a simulation testing system that simulates real driving scenarios to the greatest extent can be built. This system can collect steering wheel grip force driving data in real time, providing a rich data source for the construction of fatigue detection models.

[0005] Against this background, the present invention proposes a novel steering wheel grip fatigue detection simulation method and system based on DIL. Summary of the Invention

[0006] The purpose of this invention is to provide a steering wheel grip fatigue detection simulation method and system based on DIL (Displacement Injection) to overcome the limitations of existing technologies, improve the accuracy of fatigue detection, and avoid the potential dangers of collecting driving data under actual driving conditions through DIL system simulation testing with unlimited test scenarios.

[0007] To achieve the above objectives, this invention provides a steering wheel grip fatigue detection simulation method and system based on DIL, comprising the following steps:

[0008] S1. Establish a driver-in-the-loop system bench: establish the hardware-in-the-loop component and the software-in-the-loop component;

[0009] S2. Establish a virtual driving scenario library: Establish a virtual driving scenario library that includes different road scenarios and external environments;

[0010] S3. Establish a driver-in-the-loop system: Integrate the hardware-in-the-loop and software-in-the-loop components to realize a driver-in-the-loop system;

[0011] S4. Build a fatigue driving detection module: Add a flexible pressure sensor, EEG device, signal acquisition and detection module to the driver-in-the-loop system to build a fatigue driving simulation system.

[0012] S5. Design a simulated driving experiment: Design two experimental schemes: normal driving and fatigue driving.

[0013] S6. Establish a fatigue driving sample dataset: Combine subjective experiments with the Stanford Sleepiness Scale and objective evaluation of EEG signals to complete the labeling of different fatigue states;

[0014] S7. Establish a fatigue state identification model;

[0015] S8. Verify the experimental results.

[0016] Preferably, in step S1, the specific steps for establishing the driver-in-the-loop system test bench are as follows:

[0017] S11. Hardware in the loop: Complete the construction of driver operation device, image simulation display device, sound simulation device and seat hardware, and simulate vehicle driving state through dynamic driver simulator to realize hardware in the loop;

[0018] S12. Software in the Loop: Achieve software in the loop by building scenario, sensor simulation software, and vehicle dynamics simulation software.

[0019] Preferably, in step S6, the process for determining the optimal threshold of the EEG signal is as follows:

[0020] S61. Preprocessing: Bandpass filtering, trappass filtering, rereference, downsampling rate, and ICA preprocessing are performed on the EEG signal.

[0021] S62. Feature Extraction: Extract the power spectral density features of the preprocessed EEG signal;

[0022] S63. Determine the optimal threshold: By performing wavelet transform denoising, pre-labeling, ROC analysis, and calculating the confusion matrix on the extracted power spectral density features, the optimal threshold of the EEG signal is determined.

[0023] Preferably, in step S7, the step of establishing the fatigue state identification model is as follows:

[0024] S71. Data preprocessing: Perform linear interpolation, Kalman filtering, and normalization on the grip strength data;

[0025] S72. Feature Extraction: Extract time-domain features that reflect driver fatigue from the fatigue driving sample dataset;

[0026] S73. Feature Selection and Optimization: Normalize, perform correlation analysis, significance test, feature effect size analysis, significant feature selection, smoothing filtering, and effect analysis on the extracted time-domain features;

[0027] S74. Model Training and Optimization: Based on the selected time-domain features, the model is trained and optimized to obtain the optimal fatigue state recognition model.

[0028] Preferably, in step S8, the step of verifying the experimental results is as follows:

[0029] S81. Model Deployment: Deploy the fatigue state recognition model in the driver-in-the-loop system;

[0030] S82. Experimental verification: Conduct simulated driving experiments on different drivers and identify fatigue state based on the fatigue state recognition model;

[0031] S83, Human-Machine Interface Adjustment: Based on the identified fatigue state, the vehicle's human-machine interface is dynamically adjusted, and the warning module responds accordingly.

[0032] Preferably, the grip force data is derived from the pressure value of a flexible pressure sensor arranged on the steering wheel.

[0033] Therefore, the present invention employs the above-mentioned steering wheel grip fatigue detection simulation method and system based on DIL, and the technical effects are as follows:

[0034] 1. High-precision simulation: By combining virtual driving scenarios with the actual steering wheel, it simulates real driving scenarios to the greatest extent and improves the accuracy of fatigue detection;

[0035] 2. High safety: System simulation testing avoids the dangers of collecting data during actual driving, improving data collection efficiency;

[0036] 3. Timely warning: Effectively identifies driver fatigue, provides timely warnings, and dynamically adjusts the human-machine interface to improve driving safety. Attached Figure Description

[0037] Figure 1 is a schematic diagram of the overall framework of the driver-in-the-loop system test bench;

[0038] Figure 2 shows the creation of the virtual driving scenario library;

[0039] Figure 3 shows the driver-in-the-loop system creation diagram;

[0040] Figure 4 shows the creation of the fatigue driving sample dataset;

[0041] Figure 5 is a flowchart for determining the optimal threshold of EEG signals;

[0042] Figure 6 shows the establishment of the fatigue state identification model;

[0043] Figure 7 shows the driver's fatigue state when gripping the steering wheel in a circular motion. Detailed Implementation

[0044] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0045] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0046] Example

[0047] Figure 1 is an overall framework diagram of the driver-in-the-loop system bench of this invention. Constructing a complete driver-in-the-loop system bench requires the development of both hardware and software components. The specific method is as follows: A data acquisition module captures torque, angle, and force sensor data in real time when the driver operates the steering wheel, brake, and accelerator pedals. This data is processed by a PC and a real-time simulation platform, combined with rolling resistance and wind resistance information, to calculate the vehicle's motion state. A scene simulation model renders the driving scene based on the vehicle's state, providing feedback to the driver through image and sound simulators, and using PreScan software to simulate sensor functions. The main controller generates control signals to drive the electric cylinders and suspension, simulating real vehicle motion while simultaneously providing force feedback and vibration sensing. The entire system is highly integrated, providing the driver with an immersive driving simulation experience and effectively supporting the research and application of steering wheel grip fatigue detection.

[0048] Figure 2 shows the creation of a virtual scenario library in driving simulation. This scenario library is based on actual road driving scenarios, and its main creation process is as follows: By installing a data acquisition device on a real vehicle, data of different driving scenarios are collected. After processing, the data is divided into vehicle driving status and driving scenario images. After filtering, the image data is combined with a high-precision map library to construct a virtual driving scenario library, providing realistic driving feedback for the driving simulation platform and improving the realism of driving simulation and the accuracy of fatigue detection simulation.

[0049] Figure 3 illustrates the driver-in-the-loop system creation diagram, providing a closed-loop description of the driver, the established driving platform, and the driving simulation system. The main creation process is as follows: The driver interacts with the driving simulation software on the driving platform by manipulating the pedals and steering wheel, performing real-time operations and receiving visual feedback. The driver can customize virtual scenarios, such as road types and weather, to achieve realistic feedback interaction between simulated and real-world driving, meeting the requirements of driving experiments.

[0050] Figure 4 illustrates the creation of the fatigue driving sample dataset. The specific creation process is as follows: Based on a driver-in-the-loop simulation system, a fatigue driving simulation system was built using a flexible pressure sensor on the steering wheel and an EEG signal acquisition device. The experiment design included two parts: normal driving and fatigue driving, conducted in the morning when drivers were energetic and in the afternoon after sleep restriction, respectively. During data collection, the Stanford Sleepiness Scale and EEG signal analysis were combined to label different fatigue states of drivers using a comprehensive subjective and objective evaluation method, thus establishing a fatigue driving sample dataset that distinguished between conscious, mild fatigue, and moderate and severe fatigue states.

[0051] Figure 5 shows the flowchart for determining the optimal threshold for EEG signal features. The specific process is as follows: EEG signal data acquisition covers key brain regions, and noise and artifacts are removed through preprocessing. Power spectral density features strongly correlated with fatigue are extracted. After wavelet transform denoising, the changing trend of the power spectral density features is combined with the subjective rating results of the Stanford Somnolence Scale to complete the pre-labeling of EEG signals under different fatigue states. The corresponding fatigue threshold for EEG signals is determined through ROC analysis. Through confusion matrix verification, with an overall classification accuracy greater than 95% as the standard, if the standard is not met, a new search is conducted. Finally, the optimal threshold is determined to classify conscious, mild, moderate, and severe fatigue states, thereby achieving fatigue driving detection.

[0052] Figure 6 illustrates the establishment of the fatigue state recognition model. The specific establishment process is as follows: In the data processing stage, linear interpolation, Kalman filtering, and normalization techniques are first used to effectively address missing values, noise, and individual differences in the grip strength data. Subsequently, time-domain features are extracted, and features closely related to fatigue level are accurately selected through correlation analysis and significance testing. To ensure the integrity and smoothness of feature information, moving average smoothing, wavelet denoising, and Savitzky-Golay filtering are used for feature optimization. Finally, based on these carefully processed feature parameters, a fatigue state recognition model is constructed and optimized using random forest, one-dimensional convolutional neural network (1D-CNN), long short-term memory network (LSTM), and bidirectional LSTM. Validation set tuning is then used to strive for the best fatigue detection performance.

[0053] Figure 7 shows the driver-in-the-loop (HIL) simulation of steering wheel grip fatigue. The specific verification process is as follows: A fatigue state recognition model is deployed in the driver-in-the-loop simulation system to identify different fatigue states and dynamically adjust warning measures and the human-machine interface. Mild fatigue triggers a slight warning, while moderate and severe fatigue triggers multiple mandatory measures. The display is simplified or controls are adjusted according to the degree of fatigue to improve driving safety.

[0054] Therefore, this invention employs the aforementioned steering wheel grip fatigue detection simulation method and system based on DIL (Distributed Inertial Vehicle). By combining virtual driving scenarios with actual steering wheels, a simulation testing system is constructed to achieve a high degree of simulation and unlimited testing of real driving scenarios, facilitating steering wheel grip fatigue identification. Simultaneously, a fatigue state recognition model provides driver warnings, dynamically adjusting the human-machine interface and significantly improving driving safety and experience.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method and system for simulating steering wheel grip fatigue detection based on DIL, characterized in that, Includes the following steps: S1. Establish a driver-in-the-loop system bench: establish the hardware-in-the-loop component and the software-in-the-loop component; S2. Establish a virtual driving scenario library: Establish a virtual driving scenario library that includes different road scenarios and external environments; S3. Establish a driver-in-the-loop system: Integrate the hardware-in-the-loop and software-in-the-loop components to realize a driver-in-the-loop system; S4. Build a fatigue driving detection module: Add a flexible pressure sensor, EEG device, signal acquisition and detection module to the driver-in-the-loop system to build a fatigue driving simulation system. S5. Design a simulated driving experiment: Design two experimental schemes: normal driving and fatigue driving. S6. Establish a fatigue driving sample dataset: Combine subjective experiments with the Stanford Sleepiness Scale and objective evaluation of EEG signals to complete the labeling of different fatigue states; S7. Establish a fatigue state identification model; S8. Verify the experimental results.

2. The steering wheel grip fatigue detection simulation method and system based on DIL according to claim 1, characterized in that, In step S1, the specific steps for establishing the driver-in-the-loop system test bench are as follows: S11. Hardware in the loop: Complete the construction of driver operation device, image simulation display device, sound simulation device and seat hardware, and simulate vehicle driving state through dynamic driver simulator to realize hardware in the loop; S12. Software in the Loop: Achieve software in the loop by building scenario, sensor simulation software, and vehicle dynamics simulation software.

3. The steering wheel grip fatigue detection simulation method and system based on DIL according to claim 1, characterized in that, In step S6, the process for determining the optimal threshold of the electroencephalogram (EEG) signal is as follows: S61. Preprocessing: Bandpass filtering, trappass filtering, rereference, downsampling rate, and ICA preprocessing are performed on the EEG signal. S62. Feature Extraction: Extract the power spectral density features of the preprocessed EEG signal; S63. Determine the optimal threshold: By performing wavelet transform denoising, pre-labeling, ROC analysis, and calculating the confusion matrix on the extracted power spectral density features, the optimal threshold of the EEG signal is determined.

4. The steering wheel grip fatigue detection simulation method and system based on DIL according to claim 1, characterized in that, In step S7, the steps for establishing the fatigue state identification model are as follows: S71. Data preprocessing: Perform linear interpolation, Kalman filtering, and normalization on the grip strength data; S72. Feature Extraction: Extract time-domain features that reflect driver fatigue from the fatigue driving sample dataset; S73. Feature Selection and Optimization: Normalize, perform correlation analysis, significance test, feature effect size analysis, significant feature selection, smoothing filtering, and effect analysis on the extracted time-domain features; S74. Model Training and Optimization: Based on the selected time-domain features, the model is trained and optimized to obtain the optimal fatigue state recognition model.

5. The steering wheel grip fatigue detection simulation method and system based on DIL according to claim 1, characterized in that, In step S8, the step of verifying the experimental results is as follows: S81. Model Deployment: Deploy the fatigue state recognition model in the driver-in-the-loop system; S82. Experimental verification: Conduct simulated driving experiments on different drivers and identify fatigue state based on the fatigue state recognition model; S83, Human-Machine Interface Adjustment: Based on the identified fatigue state, the vehicle's human-machine interface is dynamically adjusted, and the warning module responds accordingly.

6. The steering wheel grip fatigue detection simulation method and system based on DIL according to claim 4, characterized in that, The grip strength data comes from the pressure values ​​of a flexible pressure sensor located on the steering wheel.