基于全局轨迹的六足机器人运动规划算法

Motion Planning Algorithm Based on Overall Trajectory for Locomotion of 6-legged Robot

为解决基于步态的规划方法难以适应复杂自然地形的问题,一种基于全局轨迹的规划算法被提出。机器人实现机动的基本运动单元是一次迈步,算法沿着运动轨迹估计下一步所能到达的位置,规划在当前位置和目的位置间机器人各关节的轨迹。在算法中机器人的一次迈步动作被分解成支撑腿移动身体和迈步腿向前摆动2个部分,采用抛物线参数曲线来描述迈步过程中机器人质心和摆动足的运动轨迹,从而避免了传统方法中迈步轨迹与参考坐标系相关的问题,便于实现在线规划。最后在文中给出了具体运动规划算法,并通过了仿真验证。

A traditional motion planning method based on gaits is difficult to adopt to complicate nature terrain. In order to solve the problem, a planning algorithm based on overall trajectory was proposed. The basic unit of robot maneuver was a walking step. The proposed method was to estimate position of robot in the next moment along a given overall trajectory, and then to plan the joints' trajectory of robot. In this algorithm, the motion of a single step was decomposed into two parts, the motion of moving robot body with support of legs and the motion of swinging legs forward, which performed in same time. A parametric parabola was used to describe the trajectory of the motions to avoid the problem that trajectories depended on a reference coordinate system in traditional method. It is convenient to realize online trajectory planning. A detail motion planning algorithm was discussed and proved by a simulation.