SBCL - improved centroid estimation

Localization of sensor nodes is one of the key issues in wireless sensor networks. It is a precondition for a variety of applications, as well as geographic clustering and routing. A simple approach for coarse grained localization is centroid localization (CL) which was firstly presented by Bulusu et al. and assumed regularly arranged beacons. Unfortunately, CL has a biased error whenever exactly three beacons are in range or the beacons are situated at the edges of the sensor field. In this work, we investigate the bias and its impacts on localization error. With our shape-based CL algorithms, we present possibilities to reduce this error and improve the localization accuracy of CL. However, since our new approaches introduce more complexity to CL, an analysis of the complexity of the algorithms is presented.