Simulated performance of local maximum likelihood detection with interferences of unknown signature waveforms

This paper presents the simulation results for the group blind local maximum likelihood (GB-LML) detection (Y. Sun, July 2000), (Y. Sun, Dec. 2001) method with interference of unknown signal waveforms. Typical multiuser detection in code division multiple access (CDMA) systems has prior knowledge of signature waveforms of all users. Recent blind detection technologies focus on demodulating the desired users' information with the prior knowledge of signature waveforms of such users, but with other users' signature waveforms unknown. In the battle field, peer to peer digital communication often comes with the interfering signal without prior knowledge of signature waveforms, moreover with some hostile interfering signals. This technology also considers the commercial CDMA network with the inter and intra cell interference. In CDMA network, the base station has the knowledge of signature waveforms of all users within the cell, but has little knowledge of the signature waveforms outside the cell, especially with users switching between different base stations. In this paper, the simulation result showed that the local maximum likelihood detection (Y. Sun, July 2000), (Y. Sun, Dec. 2001) could achieve good performance, which exceeds that of traditional blind linear multiuser detectors with lower complexity.