Mosaic localization for wireless sensor networks

Abstract

According to its importance as one of the most important source in several issues of realizing ubiquitous computing including context-awareness, service discovery, and so on, various localization algorithms for Wireless Sensor Networks (WSN) have been proposed during past several years. However, the limited resources of a sensor node, including memory, battery, computing, and capability have reinforced the possible architectural choices of the localization of sensor networks to have GPS like position aware devices. It is thus natural for finding location of the sensor node by using GPS like position aware devices. However, when the systems asked to operate over long time, the localization which uses GPS device shows several drawbacks which are critical in WSN. To solve the drawbacks, we propose the Mosaic localization system for WSN. It achieves global localization with much reduced packet transmission by reducing the number of flooding packets. Our key concepts of the algorithm are inspired by the procedure solving a jigsaw puzzle of the art Mosaic in a sense that global location is obtained by matching process (shifting, tilting, and flipping) of the local relative locations. The local relative locations of one-hop neighbors of each node can liken to the piece. It can be acquired by one-hop ranging between a node and its neighbors and among neighbors. With these local relative locations, the Mosaic Localization system builds global location of a node by pieces matching process using position of anchor nodes. As a result, the lifetime of the network increases while not losing the accuracy compared to the flooding based approach because the mosaic localization include only one time broadcasting process at the end. We proved this result by MATLAB simulation with some simulation parameters at section 5.