Performance optimisation of mobile robots in dynamic environments

Abstract

This paper presents a robotic simulation system, that combines task allocation and motion planning of multiple mobile robots, for performance optimisation in dynamic environments. While task allocation assigns jobs to robots, motion planning generates routes for robots to execute the assigned jobs. Task allocation and motion planning together play a pivotal role in optimisation of robot team performance. These two issues become more challenging when there are often operational uncertainties in dynamic environments. We address these issues by proposing an auction-based closed-loop module for task allocation and a bio-inspired intelligent module for motion planning to optimise robot team performance in dynamic environments. The task allocation module is characterised by a closed-loop bid adjustment mechanism to improve the bid accuracy even in light of stochastic disturbances. The motion planning module is bio-inspired intelligent in that it features detection of imminent neighbours and responsiveness of virtual force navigation in dynamic traffic conditions. Simulations show that the proposed system is a practical tool to optimise the operations by a team of robots in dynamic environments.