Research On Walking Control For Humanoid Robot

Humanoid walking robot compared with other multi-legged robot has a smaller volume, lighter weight, more flexible movement, and closer to the characteristics of human walking, and the best ability to adapt the environment, so humanoid robot has broader application prospects. In order to let robot have a good stability in the walking process, its gait planning and control have become the focus of the research. On the basis of collection, collation and analysis of relevant literature on humanoid robot,this paper does a research on gait planning and walking control for humanoid robot.The research work as follows:1. Gait planning and optimization base on the trajectory planning of the hip jointFirst using inverted pendulum model plans the trajectory of the robot center of gravity, then considering the trajectory of the center of gravity trajectory function as the reference variable plans the trajectory of the ankle joint’s movement.Under the conditions satisfy the geometric constrains and physical structure, determine the knee’s movement.After that,the gait planning of the humanoid has been completed.To guarantee robot’s walking stability,and accordinig to the characteristics of the motion of a humanoid robot in a single step cycle,using a high-function integrating walking chracterristics gets the gait optimization objective function.Finally,using the advanced genetic algorithm to optimize the trajectory parameters of the hip,which increases stability margin of the robot walking simulation to achieve the stability of the robot’s walking.2. Walking control method based on the inner fuzzy control and the outer arm swingingProposed robot walking is controlled by the inner and outer loop control method.The inner loop control uses fuzzy controller to adjust the robot’s stability and the outer loop controller is necessary to compesate totorque basing on arms’swing.First to develop the fuzzy rules according to the constraints of the joint movement of the robot,design a stable fuzzy controller,and simulate in Matlab and NAO robot’s simulation software.Then in Matlab,when the arms are swinging and not the yaw torque of the robot simulation experiments are been done to verify the control method validity.At the last of the paper the author gets a summary of the full text, illustrates the main findings about the subject,and mades the subject’s prospect which needs to be improved.