Robust target localization in the absence of signal propagation models

This paper considers a problem of target or source localization based on the measurements from a mobile sensor or a network of sensors. The assumption is that the received signal strength is strictly monotonic with the distance between the target and the sensor. No explicit signal propagation model, neither the structure nor the mathematical description, is assumed or used in localization. The contribution of the paper is to show that this mere knowledge of monotonicity suffices in locating the unknown target in the absence of noise. Further in the presence of noise, robust localization algorithms are developed and analyzed by exploiting various forms of angular, temporal and spatial averages. The asymptotic convergence results have been established in the presence of noise and further the finite step performance results are developed.