ANFIS for parameter estimation in coherent communications phase synchronization

The use of soft computing techniques in coherent communications phase synchronization provides an alternative to analytical or hard computing methods. The paper discusses a novel use of Adaptive Neuro-Fuzzy Inference Systems (ANFIS) for phase synchronization in coherent communications systems utilizing multiple phase shift keying (M-PSK) modulation. A brief overview of the M-PSK digital communications bandpass modulation technique is presented and its requisite need for phase synchronization is discussed. We briefly describe the hybrid platform developed by J.-S. Jang et al. (1997) that incorporates fuzzy/neural structures, namely the Adaptive Neuro-Fuzzy Interference Systems (ANFIS). We then discuss the application of ANFIS to phase estimation for M-PSK. The modeling of an implicit phase estimation scheme for M-PSK symbols with unknown structure are discussed. Performance results from simulation of the above scheme are presented.