Integrated Adaptive Compression
Abstract
The objective of the Integrated Sensing and Processing (lSP) program was a holistic approach to the design of systems. I) SAIC designed a Joint Source-Channel Coding algorithm, designed to integrate the source data with channel coding. The work resulted it an algorithm that both incorporated a feedback of sensor information into the processing chain and broke through the traditional processing flow of optimizing the compression algorithms based on separate black boxes of source compression and channel coding and modulation. 2) SAIC also designed the Fast Adaptive Modulation algorithm that demonstrated a novel method of nonlinear optimization applied to end-to-end optimization of a radio communications system. Optimization is in least-logs" rather than the more familiar "least-squares" optimization. Least-logs has the advantage that it is less sensitive to outliers than least-squares optimization. As such, it can be applied to curve-fitting and curve-finding applications in noisy images and displays. This methodology measures features, not pixels, with the resulting SNR gain due to integration over the feature.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 31, 2005
- Accession Number
- ADA433740
Entities
People
- Hanna Witzgall
- J. S. Goldstein
- Michael D. Zoltowski
- Robert R. Greene