Publisher's Synopsis
Presenting a single probability-based information-theoretic model for addressing issues in data fusion and sensor management for multi-sensor systems in general and decentralized systems in particular, this text develops mutually consistent data fusion architectures and algorithms. The algorithms are various architectural forms of the information filter and a corresponding Bayesian classification algorithm. Of significance is the normative sensor management method, making use of information-based utility functions.;An actual application to mobile robot navigation is described and a model for a Tracking Sonar sensor with its capacity to focus attention on given features is given. Throughout the text discussion, actual results and data are provided.
