基于Adams和Matlab联合仿真的双足机器人越障研究

与轮行机器人相比,双足机器人具有更灵活的机械结构,具有跨越静态或动态障碍物的能力,使其可以在更复杂的环境中工作;以往的双足机器人路径规划控制策略只能解决静止或以可预测速度运动的障碍物的越障问题,提出了一种基于模糊Q学习算法的路径规划策略,在Adams软件中建立机器人的三维虚拟样机模型,在Matlab软件中设计控制器,进行联合仿真;仿真结果表明所设计的控制策略可以有效地克服机器人在线学习时间长的问题,并且可以成功跨越速度不可预测的运动障碍物,有很好的鲁棒性。

Approach of Biped Robot to Step over Obstacle Based on Adams and Matlab United Simulation

Comparing with wheeled robot,the biped robot has more flexible mechanical system.Its ability to step over static or dynamic obstacles allowing it can work in more complex environment as well.The previous control strategy that has been adopted to solve the path planning problem of biped robot can only handle the static or predictable obstacle avoiding problems.A footstep planning approach which is based on the fuzzy Q-Learning algorithm has been introduced in this paper.The three-dimension virtual model of biped robot has been developed with the Adams software and control system has been designed with Matlab software.By united simulation,the results show that the proposed control strategy can overcome the problem of long time on-line learning,and the robot can step over the dynamic obstacle with random velocity successfully.