Intention-based front-following control for an intelligent robotic rollator in indoor environments

This work presents an intention-based assistive controller for allowing a robot to follow a human while moving in the front. This task is particularly challenging in indoor environments, as there are situations that are undecidable, namely in junctions. We propose a novel local kinodynamic planner which concurrently detects discrete routes and continuous motion paths, providing path equivalence classes. Furthermore, we detail an intention recognition algorithm to make the robot take the user-intended turn. Our scheme is experimentally tested in a T-Junction set-up using human subjects, and the results are discussed.