Navid Nourani-Vatani, Michael Bosse, Jonathan Roberts and Matthew Dunbabin
Conference Name & Place
Australasian Conference on Robotics and Automation (ACRA), Auckland, New Zealand
There is a need for systems which can autonomously perform coverage tasks on large outdoor areas. Unfortunately, the state-of-the art is to use GPS based localization, which is not suitable for precise operations near trees and other obstructions. In this paper we present a robotic platform for autonomous coverage tasks. The system architecture integrates laser based localization and mapping using the Atlas Framework with Rapidly-Exploring Random Trees path planning and Virtual Force Field obstacle avoidance. We demonstrate the performance of the system in simulation as well as with real world experiments.