Optimized motion strategy for active target localization of mobile robots with time-varying connectivity: Extended Abstract

This paper addresses optimal trajectory generation for active target positioning using a collective localization scheme under a time-varying observation topology. We show that a team of assisting robots using the optimal trajectories can improve the localization accuracy of leader robots whose commands are assigned by high-level tasks. We apply the standard centralized extended Kalman filter to estimate all robot positions by using distance-only relative measurements. In this work, we also explicitly consider the limits on the maximum ranging distance within which the robots are able to make pairwise measurements. The trace of the covariance sub-matrix corresponding to the leader robot's position estimate is selected as the optimization criterion. Simulation results are presented that demonstrate the applicability of this method and provide insights into the difficulties in optimizing this problem.