Distributed Control System And Motion Optimization For Legged Robot

Legged robot has long been a hot research field of robotics because of its potential to travel on the rugged, complex terrain. In recent years, new driving and control methods have promoted the legged robot to make great progress. The flexibility of robot could be improved by using SEA (Series Elastic Actuator) in joints. Therefore, a Distributed Control System of Legged Robot with SEA is designed in this thesis. In order to reduce internal collision in robot, the walking planning problem is redefined as a motor coordination problem. Then the motor coordination problem of a planar five-link legged robot is studied through motion optimization approach, the algorithm for solving the optimal coordinated trajectory is proposed and the coordination mechanism of partial results is also analyzed.The main contributions are as follows:1. A Distributed Control System of Legged Robot with SEA is designed by using MDA (Model Driven Architecture) approach, which has a servo frequency of 1 KHz. Practical application proves that it works well on a monopod with SEA;2. The motor coordination problem is solved by motion optimization approach. First, Lagrange dynamics equation is used for modeling. Then, an objective function is proposed by combining energy efficiency and attitude balance. Calculus of variations problem is transformed into a nonlinear programming problem by Raleigh-Ritz method. A fast algorithm to calculate the values of the objective function and gradient is designed. Finally, the optimal coordinated trajectory is got by a nonlinear programming problem solver;3. The motor coordination of legged robot is analyzed. The Adams and Matlab co-simulation experiments found that robot emerged leg retraction and leg back swing phenomenon after optimizing. The coordination mechanism of leg retraction is analyzed based on above experiments.