Research On Free Gait Planning And Control For Quadruped Robot Walking Over Terrain That Contains Obstacles

Legged structure enables quadruped robot to take advantage of the discrete regions that can support walking in the terrain, so it can adapt well to irregular or rugged terrain, which makes quadruped robot has broad application prospects both in military and civilian applications. When walking on complex terrain, quadruped robot usually uses free gait, which can adjust the leg sequence and stride according to the terrain, thus its good terrain adaptability is well played. In this paper, the free gait planning problem for quadruped robots walking over terrain that contains obstacles is studied. In order to reduce the dimension of planning space, the obstacle avoidance gait planning problem is decomposed into planning the free gait sequence in the high level, and planning the foot collision free-trajectory in the low level, and then a gait planning algorithm based on A* and a foot trajectory generation algorithm aiming at optimizing both the running time and joint energy cost are designed. Finally, the free gait control algorithm is studied in the dynamic simulation environment.Firstly, combing the gait characteristics of quadruped robot, a simplified quadruped model is propped, and a discrete representation method of gait state is designed. The continuous gait planning problem is then changed to the discrete gait sequence planning problem. The gait state transition rules are designed after considering the obstacle, stability and foot work space constraints. Then the discrete gait is represented by a graph, and a gait planning algorithm aiming for minimum gait steps is designed. Simulation results show that the gait sequence planned by this algorithm is stable and feasible; the leg sequence and stride are optimally regulated; the footholds do not fall on the obstacles, so the robot crosses the obstacles and the gait sequence length is optimized.Secondly, after analyzing the constraint conditions of foot trajectory, the trajectory planning method based on optimization target is adopted, and the trajectory running time and total energy consumption are chosen as a comprehensive optimization objective; the piecewise polynomial curve is built in the joint space of the legs, and a foot trajectory generation algorithm based on genetic algorithm is designed. The simulation results show that the foot trajectory planned by this algorithm is continuous and smooth;do not collision with the obstacles,and the running time and energy consumption of the trajectory are both reduced.Finally, the control algorithm for quadruped robot’s free gait is studied. An algorithm of regulating center of the robot body is analyzed, and finite state machine based on control logic is taken to determine the state of gait and motion sequences. Simulation results in ADAMS-Simulink show that the quadruped robot crosses the obstacles with the planned gait, and the stability margin achieves setting expectations, and the footholds position error is small with stable body posture, thus the feasibility of planning gait and the effectiveness of the control algorithm are verified.