Road surface recognition based on online estimation of soil characteristic parameters

Aiming at the problem that the traditional visual and radar methods are easily interfered by the dynamic changes of the environment and have insufficient accuracy in the identification of complex soft and muddy pavements, a pavement identification method based on the online estimation of soil mechanical characteristic parameters is proposed. By analyzing the sensitivity of wheel-ground coupling force to soil parameters, the dominant parameters are screened for dynamic identification, and the rest of the parameters are simplified based on the empirical values; the triangular approximation method is adopted to deal with the stress expression, avoiding the complex solution of the closed-form force/moment integral equation. Aiming at the time-varying characteristics of soil parameters, a strong tracking traceless Kalman filter is designed to realize online estimation of parameters. Simulation and real vehicle tests show that the proposed method can accurately recognize the soft pavement type without additional sensing equipment, and the dominant parameter estimation errors are less than 5%.