An improvement to a geolocation algorithms in wireless sensor networks

Location awareness is important in many wireless sensor networks. In sensor networks because of the size, weight, battery life and price some nodes (e.g. GPS-equipped nodes) are self-located but all other nodes have to determine their locations by communication with location-aware neighbors. The classical distance-based technique for location determination is triangulation/lateration (referred to as 3NA for short), which determines the location of a node by communication with three neighbors. In 2004 Barbeau et al proposed an algorithm, referred to as 2N+aB for short, that can locate a location-unknown node using two location-aware neighbors plus certain information from a third adjacent location-aware node. In this paper an improvement to Barbeau et al's algorithm, referred to as 2N+cB for short, is proposed. Our computer simulations show that for randomly deployed low-density sensor networks 2N+cB can increase the number of located location-unknown nodes by at least 31% compared with 2N+aB, and the average increase achieves 43.6%. Besides, we give an analysis of 3NA, as an example for the analysis of 2N+aB and 2N+cB, and our simulation result well matches the theoretical analysis.