Methods and experiments for hazardous area activities using a multi-robot system

Abstract

This work presents new methods and experiments for hazardous area activities using a multi robot system. In a special test environment various dangerous settings are reproduced. The experiments include the scanning for hazardous material and radiation. The paper will present empirical results on collective mapping of the sensor information. In addition, an approach to the problem of relative position estimation for multi robot systems is presented. The sensor information of the robots is utilized to estimate the relative positions between each other. An Extended Kalman Filter (EKF) is used to combine the gathered position information into one continuously updated position estimation. All robots of a group use these data in order to generate one common co-ordinate system. This co-ordinate system is a "relative" one, meaning that it has no fixed reference to global world co-ordinates. Preliminary results of experiments with real robots are presented.