Location estimation

Abstract

In recent years, there has been increasing interest in WSNs in academic, industrial, and commercial sectors. In wireless sensor networks, wireless nodes are equipped with measuring devices in addition to wireless communication capabilities. With such networks one can monitor habitats of interest, detect intruders within private premises, or explore unknown environment (Mainwaring et al. 2002; Wang et al. 2003; Akyildiz et al. 2002). It is often useful to know the relative or absolute nodal locations in order to improve the quality of service delivered. Nodal location information is important in the following two cases: Location-based routing: As the transmission range of each node is limited, intermediate nodes are often required to forward data packets. When a packet is forwarded, its route may be chosen according to the relative positions of the nodes. Consequently, estimating the relative positions of the nodes is important for route correctness and efficiency. Location-based routing has been widely studied, most of which determine the route by making use of the network topology as obtained in the intermediate nodes and the location of the destined node (Karp and Kung 2000; Bose and Morin 2000). Location-based services: This refers to offering services to users depending on their locations. These services include identifying the location of a patient in a hospital, locating the parking position of ones car, monitoring a variety of ambient conditions, etc. The main concern in such a system is hence the estimation error in terms of both absolute and angular distance between nodes, as well as the distance error between the estimated and real locations. There has been much work on location estimation. Here we survey various techniques of estimating nodal locations applicable in wireless sensor networks. These techniques form the basics for location estimation. We consider a wireless sensor network as a network consisting of many stationary sensor nodes (in the order of hundreds to thousands of nodes) closely placed together with wireless communication capability. Depending on the specific technique a node may also be equipped with some extra capabilities such as angle measurement, GPS, etc. We also allow a small proportion of entities, which are more powerful than the other sensor nodes in the network, whose role is to provide some special features so as to facilitate the location estimation process. We discuss the state-of-the-art beyond what was presented in (Hightower and Borriello 2001), and evaluate the pros and cons of each of them. We first go over a number of location estimation techniques where a node either gathers information from neighbouring entities to estimate its own location or shares information with each other to estimate locations cooperatively. We classify the estimation approaches into four categories, namely, distance-based, angle-based, pattern-based and connectivity-based, and provide examples in each category. We then provide a comparison between these techniques in terms of estimation steps required, message complexity, power consumption, density dependency and special hardware requirement. © Springer-Verlag Berlin Heidelberg 2007.