Dynamic Modeling Analysis of Optimal Motion Planning and Control for High-speed Self-driving Vehicles

Abstract

Motion planning and control for high-speed self-driving vehicles is complicated since it involves highly nonlinear constraints such as sideslip and rollover. Inappropriate handling of these constraints, particularly under complex terrain conditions, will lead to loss control of vehicles or even fatal accidents. Through studies on the effect of road curvature and terrain to the vehicle steering characteristics and handling stability, a simplified equal dynamic model is developed for high-speed self-driving vehicles. And a model discretization method with variable time-steps in the prediction horizon is proposed to ensure immediate dynamical response while attaining a long enough prediction horizon and computational feasibility. Taking dynamical safety related concerns such as sideslip and rollover into account, vehicle handling stability constraints, including sideslip envelope and zero moment point, are concluded through vehicle steady-state analysis. A model predictive control problem is formulated and solved to find the optimal motion trajectory and control sequence, satisfying the road environmental constraints while ensuring handling stability. Simulation results validated the capability of the proposed approach under the influence of complex road curvature and terrain.