An Ant-like Task Allocation Model for a Swarm of Heterogeneous Robots

摘要

This paper addresses the issue of applying decentralised task allocation and task switching mechanisms in heterogeneous groups of robots in order to increase their ability to respond to task demand effectively. Our work is strongly inspired by the behaviour of eusocial insects (typically ants) and their behaviour of switching tasks in order to meet the changing demand. The objective of this paper is threefold: 1) identification of task allocation and task switching mechanisms in robots inspired by ants like red harvester ants, Pogonomyrmex barbatas, 2) developing a simple model of these mechanisms for use with a heterogeneous group of simulated robots and 3) implementing a decentralised and adaptive mechanism of updating thresholds for heterogeneous groups of simulated robots. The paper extends the use of threshold based mechanisms in homogeneous robots into the realms of heterogeneous groups of robots. Experimental results show that the incorporation of task switching mechanisms in specialised groups of robots improves the foraging efficiency and swarm energy significantly.