Dynamic And Stable Gait Planning Of Miniaturization Humanoid Robot

Different from the stable supporting of wheel type, pedrail type, multiped type and other kind of robots when they are moving, stabilization of humanoid walking has always been the main technical difficulty of humanoid robot, especially the fast dynamic walking. To solve this problem, reasonable structure and control system of humanoid robot are necessary, besides, gait planning is also the key. So, gait planning has become one of the hot topics about humanoid robots. This paper mainly talked about the dynamic gait planning of own miniaturization humanoid robot.First, according to the structure, size and freedom of the miniaturization humanoid biped robot, we built the mathematic model of the robot in its front and side plane, and calculated the kinematics and reverse kinematics equations, established the basic for gait planning.Second, we use the ZMP (zero-moment point) as a criterion, to plan the dynamic walking gait of the humanoid robot. First, the stabilization of humanoid robot has been discussed when walking, and then we educed the expressions of coordinate of ZMP. Second, we analyzed the restrictions which are used to restrict the walking motion of humanoid robot. Last, using three-order spline interpolation function as a tool, we adopted five-point method to make some parameters of dynamic gait which have great influence on biped walking as variables, and got varied and adaptive gait.Then, using Matlab GA (Genetic Algorithm) toolbox, taking the maximum of stability margin of ZMP and the minimum of torque of joints as aims, we adopt multi-aim method to optimize the parameters of gait. At last, we did the simulation with the optimized parameters in environment of Matlab and ADAMS, and the simulation results proved that the method of gait planning which we designed can achieve the aim of humanoid robot's stable dynamic walking.In the end, we designed the experiment of humanoid robot's dynamic walking, the result of experiment proved the correctness of principle of gait planning. Besides, we have done the early research on the humanoid robot's complex actions planning such as push-up, and did the simulation and experiment, then we got the action which we planned and proved the correctness of our planning.