Nearly Lossless Zero Shot Learning for Robot Navigation Using Declarative Re-Planning

Navigation for mobile robots is critical for the deployment and use of robots. Without the ability to interact with the world a robot cannot bring to bear its utility. This work further explores the properties of a mobile robot navigation method called declarative re-planning (DRP). DRP is a hybrid method that leverages classical path planning and robot controllers combined with machine learning (ML) based local path planner for navigation. This work demonstrates how DRP is effective at zero shot transfer learning. DRP is able to train a navigation policy on one robot, and be used on other robots while remaining equally effective at reaching its goal. The other robots do not need to have the same properties, actuators, or number of actuators. DRP also enables simulation acceleration by allowing lower fidelity trajectory calculations to be used without effecting the outcome of the navigation trajectory.