Dimitrios Koutsonikolas, Saumitra M. Das,Y. Charlie Hu, Yung-Hsiang Lu and C.S. George Lee

Mobile robot teams are particularly suited to many application scenarios where infrastructure is unavailable or damaged. For example, mobile robot teams can be useful for exploration in remote regions or search and rescue after a disaster. Localization of individual robots in these teams is essential for enabling many applications or improving the robot’s performance in particular tasks. However, in infrastructure-less application scenarios, conventional techniques for localization have many disadvantages such as cost, deployment time, inaccuracy and energy use. Thus, there is a need for a localization scheme that works in infrastructure-less scenarios and is low-cost, quickly deployable and energy-efficient while providing reasonable accuracy for the applications.

In this paper, we propose CoCoA, Coordinated Cooperative Ad-Hoc localization. In CoCoAonly a subset of the robots in the mobile robot team are equipped with external localization devices (e.g. GPS, or laser rangers with SLAM). Subsequently, while robots perform their tasks, the subset of robots with localization devices help to localize other robots, avoiding the need for static landmarks to be deployed. This is achieved using a modi-fied Bayesian inference-based localization algorithm previously proposed for localization in sensor networks. In addition, CoCoA coordinates this localization process using multicast to put wireless devices in sleep mode periodically which provides significant energy savings. Using detailed sim-ulations and localization models calibrated from experimental data, we find that CoCoA is effective in reducing energy consumption while providing good localization accuracy.