Smooth Trajectory Tracking Using Longitudinal Distance Constraint for A 4WS4WD Unmanned Ground Vehicle*

Abstract

The model-based trajectory tracking method of four wheel steering and four wheel drive (4WS4WD) unmanned ground vehicles (UGVs) needs to establish accurate vehicle kinematics and dynamics models, which is computationally complex thus functioning poorly in real-world scenarios. The geometry-based method, i.e., the pure pursuit method, uses the position information to regulate vehicle s steering angle, which has good real-time performance in practical applications. However, the existing pure pursuit method only considers the look-ahead distance constraint, ignoring the influence of the longitudinal distance which affects the smoothness of the vehicle trajectory. It may cause the vehicle to oscillate back and forth around the target trajectory. This work proposes an improved pure pursuit method with longitudinal distance constraint (LDC) to avoid the vehicle oscillation effectively. Additionally, to alleviate the cornering cutting problem, a fuzzy- logic based controller is proposed to dynamically adjust the look-ahead distance according to the curvature. A comparative experiment between the proposed method and the conventional pure pursuit method was conducted. Results show that the proposed method has better trajectory smoothness and tracking accuracy than the traditional pure pursuit method.