| Biped robot is one of the most typical representatives of robots and it involves many disciplines.Its complex structure and dynamic model also brought great difficulty to control.Early biped robots mostly used static walking,which was slower and could only adapt to flat terrain.In order to improve the motion performance of biped robots,dynamic walking is needed.Based on the planar biped robot and "WUKONG-Ⅱ" biped robot platform,this thesis designs a dynamic walking control algorithm for biped robot based on virtual force model,and also do experiment on the adaptive control of the uneven terrain.This thesis mainly contribute to following area:1)Building virtual force control models.Firstly,a motion state machine based on dynamic walking is designed,and virtual force models are established for the single leg support phase and the two leg support phase.Then it analyzes the constraints between virtual forces and the relationship between composition and distribution.Based on the virtual force model control,the robot's attitude and height are stably controlled.2)Biped robot walking speed control algorithm.First,the Kyman filter is used to fuse the gyroscope data with the kinematics solution data.Then based on the foothold selection model proposed by Raibert,a foothold selection model is designed,and the forward speed is controlled by this model.With this method,the robot's forward speed can reach 2km/h.3)Adaptive control method for uneven terrain.The influence of uneven terrain on walking control is mainly reflected in height control.This method initializes the expected height of the robot in the world coordinate system,and then plans the trajectory of the swing leg in different situations,so that the height of the robot in the world coordinate system remains stable.In this way,the robot can walk on uneven terrain. |
PDF Full Text Request