Integration of Machine Learning with Dynamics and Control: From Autonomous Cars to Biomechatronics

Researchers at the University of Waterloo are strategically combining artificial intelligence (AI) methods with dynamics and model-based control approaches, to solve real-world problems that are too difficult or computationally intensive to tackle with either AI or conventional methods on their own. Our goal is to combine the best features of machine learning and model-based control theories, in contrast to computer scientists that are seeking purely AI solutions to dynamics and control problems. This article presents our recent work towards this goal, with an emphasis on two very active research applications: Autonomous Cars and Biomechatronics.