9460878 Nardi This Small Business Innovative Research Phase I project deals with development of novel neural algorithms for adaptive demodulation of spread spectrum communications. Direct-Sequence Spread-Spectrum Multiple-Access (DS/SSMA) has long been the technology of choiae in a wid evariety of satellite and military radio networks where multiple-access capabilities, anti-jamming, low probability of interceptm, and dynamic topologies are important. DS/SSMA technology is emerging as a likely modulation choice in emerging space and third-generation commercial mobile radio networks (e.g. personal communication systems) and in indoor wireless communications. The performance degradation casused by the dissimilarity of the received powers of a set of users who transmit simultaneously through the same channel. Multiuser detection promises substantial gains in bit-error-rate, which in turn afford the possibility of eliminating the need for power control and a significant reduciton in the number of chips- per-symbol (spreading ratio) necessary to sustain a desired level of multiuser capability. Neural algorithms are naturally suited to the training and adaptation of multiuser detection algorithms to realize those gains in practice. The proposed project will develop neural-network-based adaptive signal processing architectures for near-far-resistant demodulation of spread- spectrum multiple-access signals. This approach will lead to significant advances in the ability to use direct-sequence spread- spectrum formats in emerging radio networks ***
