Behaviour Study of a Multi-Agent Mobile Robot System during Potential Field Building

Abstract: In this paper a multi-agent based mobile robot simulation system will bepresented where the behaviour of the system is studied with different number of agents (1,
3,6) and also with different number of ultrasonic range sensors on agents (8 or 16 US
sensors on individual agents). The task of the autonomous agents is to create the potential
field (PF) of an unknown environment. The classic problems of PF building, like oscillation
and trapping, are not the focus of the article, but instead, the article is concerned with the
agents’ self-organizing ability where self-organizing is controlled by a genetic algorithm
(GA). The GA is equipped with two fitness functions where one “maintains” the distances
between certain agents (spat distr), while another “watches” the area coverage (area
cover). In fact, the paper can be divided into three main parts. The first part describes the
ultrasonic sensing and range measuring with systematic errors, the potential field (PF)
building and the moving strategies. The second part contains description of the GA, the
operation of the GA, the structure of the system, the fitness functions and a general systemerror
determination. In the final third part, the obtained results are analyzed and presented
in the appendices.