Mobility assisted localization for robotic sensor networks

The accurate localization of sensor nodes in a mobile robotic sensor network is important in many surveillance applications. Based on our previous work in sensor localization, this paper introduces a mobility assisted localization approach guided by network rigidity for mobile robotic sensor networks. The rigidity test of a network is based on a pebble game algorithm, that decides whether the network is uniquely localizable or not. If its local network is not uniquely localizable, the mobile node circles around its original position to collect at least two extra ranging data with each of its neighbors while recording its trajectory using odometer. We prove that the extra data with robot trajectory will make the local network uniquely localizable. The local maps of all the nodes are then merged into a globally consistent map. We evaluate the effectiveness of the proposed approach using two examples of low-degree networks and four types of topologically different networks.