Distributed processing in estimation and detection

Abstract
New distributed forms for optimal estimation and detection problems that arise for instance in a distributed sensor network (DSN) are proposed. One of the prime motivations in a DSN is to try to minimize the communication load between different sites, This is achieved by preprocessing or summarizing the observations locally, e.g., at the sensor location, thus geographically "distributing" or "decentralizing" the algorithm. A batch method is described, as well as an on-line or real-time algorithm when sensor parameters are a priori unknown. After sensor failure optimality is preserved under these new constraints.

This publication has 0 references indexed in Scilit: