MOTION PLANNING OF A WALKING ROBOT USING ATTITUDE GUIDANCE

摘要

This paper addresses an improvement in navigability of a walking robot through the development of a simple method for measuring and correcting tracking errors while traversing uneven terrains. The proposed motion planning method utilizes a modular sensory system to measure the drift experienced by individual footholds in terms of translational and rotational drift errors. The method then computes the required corrections to prevent or reduce drift propagation resulting from the terrain irregularities by influencing the gait generation scheme. Landmark navigation approach is implemented here using a ripple gait to evaluate the locomotion of the robot in terms of tracking errors. Dynamic simulations are conducted to evaluate the success of the controller while real world experiments show significant improvement in the robot's navigability for path following missions in uneven terrains.