Random Matrix Theoretic Algorithms for Optimal, Adaptive Fusion of Multi-Modal Information Sources
Abstract
Project Summary: Fusion of diverse information sources can potentially yield significant detection and classification performance gains and provide robustness relative to sensing using just a single modality. There are significant unmet challenges to multi-modal fusion in environments where the per-modality SNR is low and time-varying. This is especially the case in US Navy operating environments where the physically driven, time-varying characteristics of the ocean and/or jamming by the adversary introduces distortion/interference such that at various time instances, different modalities will have higher SNRs and the identity of the modalities with better SNRs will keep changing. The technical challenge is to automatically compute the weighting coefficient that is to be assigned to each modality modalities with greater informational content should receive a higher weight while modalities with lower information content should receive a lower weight or not be used at all.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2016
- Source ID
- N000141512141
Entities
People
- Raj Rao Nadakuditi
Organizations
- Board of Regents of the University of Michigan
- Office of Naval Research
- United States Navy