The Research On Planning And Controlling Method Of Walking For Biped Robot

The controlling and planing method have always been a hot issue in the field ofRobotics, Usually the biped robot were treated like a Industrial Robot by the ZMPconstraint, the generated gait was more inefficient then human. So the main ideawas to find a controlling and planing method for biped robots by studying themechanism of human bipedalism to carry out a much more high efficient gait.1. Studying the kinematic of robot with D-H method, using the homogeneoustransformation for the forward kinematics, using a dimensionality reduction methodfor the inverse kinematics.2. Using Kane method for the forward dynamic of the robot, using theNewton-Euler for the inverse dynamic of the robot.3. Developing a dynamic simulation platform with PhysX and DirectX3D APIfor studying the robot, with high solving speed and high-precision.4. The common principle of Bipedalism from all kinds of researches on humanbipedalism will be a significant guidance to robot walking. We found that energeticoptimum is a key principle to human bipedalism. A simple bionic control methods ispresent to test the feasibility that using the similar method on robot.5. There’re many traditional gait planning method based on the linear invertedpendulum or table-cart model, by which the mechanical energy of robot during theswinging phase was fluctuant and lead to a much more inefficient gait compare tohuman, to solve this we introduced a new gait planning method based on an invertedpendulum with a virtual pivot, by which the mechanical energy of robot wereapproximate conservative and the ZMP were traveling form the heel to toe, like theground reaction force mechanism of bipedalism.