Improving Kinodynamic Planners for Vehicular Navigation with Learned Goal-Reaching Controllers
Aravind Sivaramakrishnan,Edgar Granados,Seth Karten,Troy McMahon,Kostas E. Bekris,Aravind Sivaramakrishnan,Edgar Granados,Seth Karten,Troy McMahon,Kostas E. Bekris
This paper aims to improve the path quality and computational efficiency of sampling-based kinodynamic planners for vehicular navigation. It proposes a learning framework for identifying promising controls during the expansion process of sampling-based planners. Given a dynamics model, a reinforcement learning process is trained offline to return a low-cost control that reaches a local goal state ...
