Target Parameter Estimation with Distributed Sensors
Abstract
The primary focus of this research was the problem of detecting targets and estimating their locations with sensor arrays. Particular attention was paid to the performance requirements in the strategic defense scenario which dictated high probability of detection and high accuracy in localization in the presence of coherent interference. As the luxury of off-line computation was not available, computational efficiency in algorithm design was stressed. Since the sensors may be geographically distributed, the problem of distributed computing and computational resource allocation were also addressed. Statistical algorithms that were considered in this regard, for example hidden Markov model- based algorithms were also used for problems in vision and pattern recognition. The theoretical developments have extended our understanding of the degree of data-reduction, without loss of information as a preprocessor for detection and estimation. The new algorithms have generalized signal-subspace detectors/ estimators to beam-space and wideband processing. Neural network formalism was also used both for the estimation problem and for issues related to the channel bandwidth allocation and storage.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1992
- Accession Number
- ADA264706
Entities
People
- M. Kaveh
Organizations
- University of Minnesota