Behavioural Approach for Multi-Robot Exploration

Abstract

Control strategies for mapping of unknown environments require a tradeoff between exploration and accuracy. One approach to balance these conflicting requirements is to use schema based behaviours to provide a means of exploration and manipulate the parameters controlling such schema to maintain the desired localisation and mapping accuracy. This paper provides such a behaviour based strategy for multi-robot exploration. Beyond the coordination of forces attracting agents to unexplored frontiers and repulsing against obstacles, as well as fellow agents, the proposed reactive approach is coupled with frontier based path planning to escape local minma otherwise encountered in a solely potential field based solution. The proposed behaviour based approach is designed so that it can be augmented to moderate movement based on criteria such as localisation accuracy in mapping applications. One planned application is the integration of proposed mechanism with probabilistic filtering for the mapping of indoor environments. 1