Model-Free Detection and Following of Moving Objects by an Omnidirectional Mobile Robot using 2D Range Data

Abstract

The ability of mobile robots to detect and track moving targets is frequently considered as an important behavior that is a necessary pre-requisite while developing autonomous robots intended to coexist with humans or other dynamic objects in a shared environment. If the robot can track dynamic objects this is further simplifying the development of autonomous navigation and obstacle avoidance behaviors as well as the robot ability to operate within multi-agent formations. The approach proposed in this investigation allows a KUKA youBot omnidirectional mobile platform, equipped with a 2D LiDAR sensor to detect and follow a moving target (i.e. another mobile robot) in a shared indoor environment. It uses a nonparametric representation of the dynamic objects that is independent of their classes and shapes. The target following holonomic mobile robot is also maintaining a pre-specified relative position and orientation with respect to the tracked dynamic object by controlling its own linear and angular velocity.