Motion Control And Online Footstep Replanning For Humanoid Robots

There are various achievements in the area of humanoid robots motion control after decades of research work, but it’s still a challenge to control these robots to walk stably and flexibly. The control of bipedal walking is a complicated task. This thesis first reviews the development of humanoid robots around the world, and lists the most important achievements in biped motion control. Then the research work on online footstep replanning and motion control of humanoid robots based on the project and platform provided by the lab is fully discussed. The main contributions are as follows:1. A reference path generation algorithm is proposed. The robot is dealt as a mass point, and then a time optimal control method is designed, which considers the velocity and acceleration limits to generate a smooth trajectory. Finally the trajectory is transformed to get a reference footstep path.2. An online footstep adjustment and motion control algorithm is proposed. Based on the reference path, the state of ZMP, COM and footsteps are fully discussed to give strategies to adjust those factors and control the balance of the robot. Finally through the preview controller, a smooth COM trajectory can be generated.3. A software system for the humanoid motion control is designed. Besides the basic motion control module, a monitor module and decision-making modules are designed to make them work together to give a complete solution of humanoid motion control. And they are applied on two applications.