Agile & Dexterous Robotics Lab
At the Agile and Dexterous Robotics Laboratory (ADRL) we develop model-based and model-free control and machine learning methods that enable robots with arms and legs to roam and manipulate dynamice and complex environments robustly and with agility. These environments require generalized approaches to impedance and whole body control. To achieve robust performance we complement robots with adaptive and learning methods. We devlop reinforcement learning algorithms, whereby a robot improves its performance in much the same way humans learn highly developed skills from experience. Our research has applications in mobile manipulation, grasping, legged locomotion, prosthetics, field/space robotics, bio-mechanics and bio-inspired robotics.