Joint Multi-Source Localization and Environment Perception in Wireless Sensor Networks

This paper addresses the problem of joint multi-source localization and environment perception in wireless sensor networks (WSNs) based on received signal strength (RSS). Contrary to the traditional RSS approaches which assume known number of sources and known parameter of environments, we propose to estimate the sources with unknown number and the unknown environmental parameter simultaneously. By assuming that the sources are sparse in the sensing field, a non-convex l(1) regularized least squares problem is formulated. This non-convex problem is then decomposed to two simple subproblems, one for multi-source localization and another for environment perception. A joint multi-source localization and environment perception algorithm is proposed, with extensive simulations to validate its effectiveness.