TL;DR
This paper bridges the gap between subjective driver feedback and objective physiological measurements to quantify "steering feel." By analyzing EEG and EMG signals during driving maneuvers, researchers found correlations between mental/physical workload and perceived vehicle handling, offering a novel approach to evaluating and designing better steering systems.
What Problem Does This Paper Address?
Evaluating "steering feel" – that intangible sense of connection and control a driver has with their vehicle – has traditionally relied on subjective ratings. While these ratings are crucial, they suffer from inconsistency and discreteness due to individual differences in drivers' mental and physical states. Existing objective methods often fail to directly map vehicle handling metrics to these subjective feelings. This paper tackles this challenge by proposing a method to objectively quantify steering feel by analyzing the driver's psychophysiological responses, specifically their mental and physical workload, during driving tasks.
Methods
The study employed a combination of subjective and objective measurement techniques.
- Subjective Evaluation: Drivers rated their subjective experience of steering feel after completing a series of driving tests. A multi-level evaluation structure, inspired by analytic hierarchy process, was used to break down steering feel into components like maneuverability, returnability, on-center steering performance, and lane-change ability.
- Objective Physiological Measurement:
- EEG (Electroencephalogram): Used to assess mental workload. Spontaneous EEG signals were recorded, with a focus on the alpha wave band, which is known to be affected by mental workload.
- EMG (Electromyogram): Used to assess physical workload. Surface EMG signals were collected from shoulder muscles (deltoid, triceps) to measure muscle activity and co-contraction.
- Steering Effort Sensor: A low-mass, moment of inertia sensor was mounted on the steering wheel to measure steering torque and angular position.
- Driving Task: A double-lane change test, standardized by ISO/3888-1, was used. This test was chosen to reflect steering accuracy and comfort. Drivers performed sets of tests at different speeds (40, 60, and 80 km/h).
- Data Analysis:
- EEG signals were processed using Independent Component Analysis (ICA) to remove artifacts like ECG.
- EMG signals were normalized by maximum voluntary contraction (MVC) to account for individual differences. Muscle activity and co-contraction were analyzed.
- ANOVA was used to determine the statistical significance of correlations between physiological features and vehicle types.
Datasets
Not discussed in the paper. The study involved a controlled experiment with human participants and vehicle testing, rather than using pre-existing datasets.
Key Insights and Ideas
- Physiological Correlates of Steering Feel: The study demonstrates that physiological signals (EEG and EMG) are not just passive responses but actively correlate with subjective perceptions of steering feel. This opens the door to objective quantification.
- Mental and Physical Workload Linkage: Both mental workload (indicated by EEG alpha band activity) and physical workload (indicated by shoulder muscle activity and co-contraction) were found to be linked to maneuverability and lane-change ability. This suggests that how a driver feels the steering is a complex interplay of cognitive and physical effort.
- Vehicle Type Matters: Different vehicle types elicited distinct physiological responses, even during the same driving task. This highlights how vehicle dynamics directly influence the driver's experience and workload.
- Muscle Co-contraction as a Drivability Indicator: Muscle co-contraction, particularly in shoulder muscles, emerged as a promising indicator for quantifying vehicle drivability, especially lane-change ability. Lower co-contraction was associated with better subjective ratings for lane-change ability.
- Beyond Subjectivity: The research provides a pathway to move beyond the limitations of purely subjective evaluations, offering a more reliable and consistent method for assessing steering feel.
Key Results
- EEG Alpha Band: The amplitude of the alpha band in EEG signals was significantly affected by the type of test vehicle. Specifically, the power spectrum of the alpha band was slightly decreased when driving vehicle 3 compared to other vehicles.
- Muscle Activity: Test vehicle 2 exhibited the minimum muscle activity while showing the highest maneuverability rating. Conversely, test vehicle 3 had the maximum muscle activity and the lowest maneuverability rating.
- Muscle Co-contraction: Test vehicle 1 showed the minimum muscle co-contraction, while test vehicle 3 consistently had the maximum. Muscle co-contraction was found to be closely related to subjective lane-change ability ratings.
- Subjective Ratings: Test vehicle 2 generally received higher subjective ratings for maneuverability and lane-change ability compared to test vehicle 3. Test vehicle 1 performed well in accuracy and rolling feel.
Limitations and Future Work
- Limited Number of Drivers: The study involved only six experienced male drivers. The authors acknowledge that a larger sample size (at least 10 drivers) would enhance reliability and reduce discreteness in subjective evaluations.
- Focus on Specific Muscles: While shoulder muscles were analyzed, future work could explore other muscle groups involved in steering.
- Generalizability: The findings are based on a specific driving task (double-lane change) and a limited set of vehicles. Future research should investigate a wider range of driving scenarios and vehicle types.
- Future Directions: The authors suggest exploring more parameters and verifying their effects to further refine the quantitative evaluation of steering feel. They also aim to use these physiological features as indicators of different aspects of vehicle performance.
Summary updated