Hybrid non-parametric belief propagation for localization in wireless networks

Automatic self-localization is becoming a critical feature for the effective use of ad-hoc wireless networks in military or civilian applications, especially in short to medium range GPS- denied environments. The use of a distributed cooperative system which optimizes both the information and the computational cost of each node in the networks is not only highly desirable but, in the case of a high noise environment, necessary. In this paper, we propose a novel hybrid method combining the advantages of non- parametric belief propagation (NPBP) and maximum-likelihood (ML) estimation (e.g., expectation maximization (EM)). In this fashion, a message passing system is designed to minimize both the complexity of the messages sent, and the number of message passing iterations required for the nodes to self-localize.