In this SuperUROP project, Joe Griffin and I will research if the control approaches that Robot Locomotion Group has developed for an torque-controlled hydraluic robot like Atlas are still feasible on a relatively weak robot powered by electric motors. The significance of the project is not only to verify the mathematical abstraction that is made on the controllers and path planners on the control suite – Drake, but also to realize what are the required parameters when defining a humanoid robot considering of its power and primitive control input. From this understanding, we will compare several state-of-the-art algorithms on their efficiency on the robots

I have been working with the Robot Locomotion Group throughout the summer on restoring Hubo and familiarizing the Drake system. Hubo and the motion capturing system has been updated to be in sync with Atlas’s setup. Currently, I’m working on having Hubo catching with a quadratic programming controller with the Vicon motion capture.