Robot assisted localization of sensor networks guided by rigidity

Abstract

In this paper, we propose a robot assisted localization guided by rigidity checking. For a network to be localizable, it must be rigid, which means given the existing edge constraints, there is only one location for each node in it. Many networks, however, are not rigid. To localize these nonrigid network, a robot can be used to add additional distance measure between nodes so that the network becomes rigid and localizable. To facilitate the localization and reduce the travel cost of the robot, the rigid subregions of the network are first identified. The identification of rigid subregions provides a guide for the movement of robots. A brief introduction to rigidity theory is presented in this paper. Experiments that evaluate the performance of the proposed approach are presented. The experimental results prove that the proposed approach is sound and gives excellent results compared to previous approaches for localization.