Wireless Sensor Localization With Distance Measurement Information

The dissertation would focus on two problems: localization in wireless sensor network and distance reconstruction. Sensors are often randomly deployed and in such case, determining each sensor's position is critical, which explains the recent attention given to the sensor localization problem. Previous methods of projection onto convex sets (POCS) overcome the multimodality problem that plagues earlier least-squares formulations.;Previous efforts in this direction require that the sensor be located in certain position in the plane. Here we propose a new algorithm which projects onto the boundary of convex sets, and features a computationally simple update procedure. The new algorithm has a unique solution no matter where the sensor is located at. Simulation results will be shown and conclusions would be given.;Sensor localization typically exploits distance measurements to infer sensor positions with respect to known anchor nodes. Missing or unreliable measurements for special nodes can impede such procedures, raising the problem of distance measurement reconstruction using distance information from other nodes. The problem has traditionally been approached through multidimensional scaling, and more recently through semidenite programming and low-rank matrix completion. Here we develop new iterative reconstruction algorithms which instead exploit inertia of key matrices, thereby encompassing stronger constraints than rank alone. This serves to overcome limitations observed in earlier competitive approaches which do not exploit the problem structure as well. Simulation examples illustrate the performance of the new algorithms.