AAS 96-110

DERIVATION AND OPTIMALITY OF DISTRIBUTED FILTERING ALGORITHMS BASED ON MINIMUM VARIANCE METHODS

N. Nabaa and R. H. Bishop, The University of Texas at Austin; J. R. Carpenter, NASA Johnson Space Center

Abstract

The problem of fusing data from a distributed network of sensors has received considerable attention. In estimation theory, measurements from several sensors can be processed either centrally, by a Kalman filter, or in a distributed fashion, where local sensor information is processed by local Kalman filters and a central processor combines the local estimates. This paper presents new distributed filtering algorithms based on minimum variance methods. These algorithms are compared to each other and to other methods, in terms of derivation, optimality, transmission and computation requirements, for linear systems. The extension to non-linear systems is also discussed.