Quasi-Optimal Processing in Spread Spectrum Environments.
Abstract
The derivations of robust Locally Optimum (LO) detection algorithms (high interference/signal ratios) with and without memory, and their performance in a spread spectrum communications system, are the subject of this report. There are three forms of LO detection considered: (1) detection with memory; (2) detection in independent and identically distributed (iid) noise; and (3) memoryless detection. Several algorithms are investigated, all using the usual assumption that the small signal corrupting the noise estimate is acceptable. Simulation results show, however, that this assumption is not always valid, even in high interferencel signal ratio environments. Detailed investigations focused on regions of poor performance, using ideal signal-free estimates of the noise as a baseline. The ideal estimator yielded a truly robust detector over a wide range of scenarios. Recommended research should thus be focused on means for obtaining true noise estimates in practical environments.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1996
- Accession Number
- ADA321811
Entities
People
- Donald R. Ucci
- John Tanas
- William R. Jacklin
Organizations
- Illinois Institute of Technology