Long term optimisation of a mobile robot with proprioceptive perception

Abstract

<p>Robots expended a significant portion of their energy on mobility therefore it is important to optimise paths for the terrain. Our previous work has demonstrated this optimisation in simulation and this paper is concerned experimental validation using a small scale robot. An energy-constrained robot can maximise its time in the field by taking paths that minimise its energy expenditure. We have shown that by recording motor power usage this can be used to generate maps, explore and ultimately converge to minimum energy tours of the environment. The comparison of simulated and experimental results on small scale robot shows equivalent convergence to optimal paths demonstrating the validity of the simulation.</p>