Abstract
Wheeled mobile robots with multiple joint legs have excellent geometric adaptability and stability. Specifically, ones with 3DOF SCARA legs are tolerant of vertical force and may allocate wheels at flexible position and steering angle within mobility range of joints. The steering angle, though, depends on both ground velocity field at wheel position and joint angles, while they correlate with each other. Moreover, even within reach of leg, wheel position and steering angle are strictly constrained if the ranges of joint angles were limited. To resolve these problems, in this paper, a nonlinear tracking control method for wheeled mobile robots with legs and an optimal leg allocation method using variable transformation and quasi-Newton method are presented. The nonlinear tracking controller renders the robot to track the desired trajectory with designated exponential convergence rate. The optimal leg allocation method can assign wheel position at the nearest of the desired point by feasible joint angles, while steering angles are consistent with ground velocity field.
