IPAB Workshop - 21/03/2019

Title:

A generic model-based control and learning framework for manipulation and locomotion

Abstract:

One of my research interest is on developing a unified theory for robot control that generalizes over different motion. This is done by decomposition of control law into smaller orthogonal spaces, so each can be treated as a smaller control problem. To make robots safer in an environment with human, I also investigate how to make robots more versatile under external disturbances. For this, we explicitly model the external force acting on the robot, so we can compensate it accordingly. In order to adopt a motion into a slightly different scenario, I will also discuss some work the estimation of movement constraints imposed by the tasks or the environment. The applications of this framework were verified on both manipulation and locomotion on real robotic platforms.