Tracking methods for relative localisation

The multi-robot relative localisation problem asks whether it is possible for an autonomous vehicle to start at an unknown location and incrementally estimate its own position with reference to the relative locations of the other robots based only on local sensor information. As shown in our previous papers [13] and [14] the topic of relative localisation (RL) for heterogenous multi-robot systems in unknown environments is a challenging task. We used the information of SICK laser scanner systems to estimate the relative positions between the robots. An Extended Kalman Filter was used to combine this information into continuously updated position estimations. All robots of a group utilized this data to generate one common co-ordinate system. In this paper we would like to address the sub-problem of tracking which is inherent to RL.