Biped robot Gaiting Using Trajectory Planning and Robust PID Control System

• Main Authors: Kao-wei Lee, 李國威
• Other Authors: Ren C. Luo
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/67052874432499642563

碩士 === 國立中正大學 === 電機工程所 === 96 === In this thesis, the goal is to design a biped robot and simulation in MSC Visual Nastran 4D. Traditionally, the walking trajectory planning has two kinds of methods: one is cubic polynomial curve, and the other one is cycloid cubic. Even though, cubic polynomial curve can have more closes to the movement orbit of the waist as human being during walking locomotion. But the shortage of cubic polynomial curve are calculate complicatedly and difficult in asking solve. Therefore, we adapt the method of cycloid cubic for plan the walking trajectory of our biped robot model in this thesis. Because the cycloid cubic method in computation is easier than cubic polynomial curve, so take cycloid cubic method to design our motion trajectory. Put to use the geometry method and kinematics concept to design biped robot model in this thesis. In the first instance, use the forward kinematics to define whole biped robot coordinate, and then by means of inverse kinematics to derive the moment trajectory of each joint. In the case of calculating the angles have some methods, such as pseudo inverse, generalized inverse transform, geometry, optimal perturbation, and null-space method. But owing to the geometry has more batter efficiency than others. Then use Z.M.P to calculate the compensation angle to achieve the torso compensation and keep the whole biped robot balance. Because adaptive control has develop ripe, this thesis introduce a Robust Adaptive PID controller to achieve tracking trajectory, and uses nonlinear controller to control the biped robot without derive out the dynamics equation. Then utilize SOLIDWORKS to design biped robot model and realize control system in SIMULINK, and use the control architecture in SIMULINK combine with MSC Visual Nastran 4D in the end. Furthermore, consider the assumption previously such as friction, reacting force… Eventually, introduce the biped robot model to MSC. And obtain a satisfied simulation results.