Cooperative Localization in Wireless Sensor Networks: Centralized Algorithms

Abstract

The basic localization techniques known as triangulation and angulation were introduced in Chap. 2. In two-dimensional triangulation the location of a mobile device is computed by measuring its range from at least three base stations with known coordinates. Analogously, in two-dimensional angulation the mobile’s location is computed from the arrival angle from at least two stations. While these techniques are practical in Global Positioning Systems or cellular networks, in some networks connectivity of some nodes to even two base stations cannot be guaranteed. A prime example is in wireless sensor networks (WSNs). Because wireless sensors often have a deployment life of months or even years, battery conservation is critical to their operation. This prescribes transmitting infrequently and over short distances. To address the latter, nodes communicate between each other via short, multihop links to stations external to the network (Perkins 2001). The coordinates of the base stations are either hardwired or—because in most cases they are installed outside—can be determined through GPS. In contrast, many WSN applications—such as military or in emergency response—require on-the-fly setup, meaning sensor positions cannot be hardwired and, since sensors are battery constrained, they may not have sufficient power to receive GPS signals. As such, sensors must cooperate in order to extrapolate their locations through multihop links to the stations. This is the basis of cooperative localization.