Obstacle avoidance control of unmanned ground vehicle based on NMPC

Abstract

This paper presents a comprehensive control of the unmanned ground vehicle (UGV) about obstacle avoidance, including the lateral and longitudinal control. Firstly, local path is planned to realize obstacle avoidance of UGV. The direct collocation method is utilized to solve the online path planning problem. Then, based on the UGV's kinematic model, the nonlinear model predictive control (NMPC) is used to achieve the UGV's steering and speed control. In order to avoid rollover and slip, the input constraints are taken into account. Finally, the simulation results show that UGV can avoid the obstacle and track the reference commands.