Research on driving fatigue detection model in plateau based on steering wheel indicators

The low-pressure and low-oxygen environment in plateau regions significantly increases drivers’ psycho-physiological stress, affecting their driving behavior and road safety. This study constructed a simulation model using the UC/win-road simulator to replicate a plateau highway scenario. A fatigue detection model was also developed based on characteristic parameters from the steering wheel indicator. Twenty-six drivers (18 males, 8 females) from Lhasa, located at an altitude of 3650 m, participated in a simulation experiment. Heart rate variability (HRV) was chosen as the characteristic parameter from electrocardiogram signals to assess driving fatigue. Previous studies helped determine a threshold range, defining three fatigue states in the simulation: alertness, mild fatigue, and severe fatigue. Ten steering wheel angle indicators were extracted for time and frequency domain feature analysis. One-way ANOVA and collinearity tests identified feature indicators showing significant differences between fatigue states, which were then used as characteristic parameters for fatigue detection. An ordered logistic model was used to build the driving fatigue detection model. The driver’s fatigue state, determined by HRV, was the dependent variable, while steering wheel indicator parameters served as independent variables. The research results provide technical support and serve as a reference method for real-time detection of driving fatigue in plateau regions.