A Distributed Range-Free Localization Algorithm Based on Clustering for Wireless Sensor Networks

Nodes self-localization algorithm is one of the wireless sensor networks´ key techniques. The novel distributed range- free algorithm is based on clustering technique of data mine. Each regular node and its random two connected anchors construct a triangle. The regular node can calculate its pair of possible conjugate positions by the two anchors coordinates and edge scale of the triangles which may be the scale of the least hops between the three angles instead. Clustering selects the most similar cluster from all possible positions. The centroid of the subset should be used as the position of the regular node. Simulations by OPNET show that precision can be increased to approach 1/4 of radio range and better than that of most representative traditional ones, such as DV-hop.