Lateral control of autonomous vehicles with sliding angle reconstruction

Lots of satisfactory results of high-precision lateral control have been reported with the assumption that vehicles move without sliding. But this assumption never comes true under real working conditions. More and more anti-sliding controllers have been designed which heavily relied on sophisticated control laws. Although the previous works can actually improve the guidance accuracy, the complexity and the rigorous requirements for the controller abilities make such controllers not very realistic for actual applications. In this paper a kinematic model is built which takes sliding effects into account by introducing two additional tire sliding angles. Based on this model an anti-sliding controller is designed. But unfortunately its efficiency completely depends on the estimation of the sliding parameters which cannot be directly measured by sensors. To overcome this problem an adaptive observer is designed using Lyapunov methods. With this observer, the cornering stiffness parameters instead of the sliding angles are estimated. The Lyapunov stability theory guarantees that the estimated value of the cornering stiffness would converge to its real value when the persistent excitation (PE) condition is satisfied. Consequently the sliding angles are reconstructed precisely. Simulation and experimental results show that the sliding effects can be compensated effectively by the combination of the anti-sliding controller and the sliding angle reconstruction.