Tire-road surface characteristics estimation for skid-steered wheeled vehicle

Skid-steered wheeled vehicles have gained extensive application across multiple domains owing to their outstanding maneuverability, with the demand for intelligent features continuously growing. The achievement of such intelligence is critical to the accuracy of the control, which is significantly influenced by the stiffness of the tire and the adhesion coefficient of the road. However, current estimation methods face challenges such as inadequate precision and monolithic validation techniques. To address these issues, we propose a hybrid off-line and on-line estimation approach. Initially, a dynamic model for multi-axle vehicles and a brush-based tire model were constructed. Following this, an Extended Forgetting Factor Recursive Least Squares (EFRLS) for estimating the road adhesion coefficient, alongside an adaptive genetic algorithm (AGA) for estimating tire parameters, was developed. Ultimately, joint simulations using Tracksim and SimLink, along with real-world vehicle tests, were performed. The estimated coefficient of road adhesion was found to remain stable within the reference range, while the discrepancy between the tire stiffness values obtained from the simulation using the estimated parameters and those from the real vehicle tests was of the order of.