Locomotion strategies for legged robotics platform

Abstract

Locomotion of legged robots is one of the major research projects in robotics since legged robot can traverse more robust terrain than other robots. Different types of gaits and trajectories of legged robots can be used to help human do certain task. The most widely used approaches for locomotion strategies these days are the model predictive control approaches. Unfortunately, this approach only works well for short duration dynamic motion planning. A recently developed framework called Trajectory Optimization for Walking Robot can be used to generate motion for longer time horizon with better performance than model predictive control. In this project, we tried to implement and extend this framework to create a silent gait for walking and trotting of the legged robot and to make it able to generate a stair climbing trajectory for legged robot.