Adaptive Antenna Algorithms Using Successively Re-encoded Data for GSM

Abstract

Adaptive antenna algorithms and a multistage receiver architecture are presented for the Global System for Mobile communications (GSM) where the beamformer weights are successively modified using re-encoded data. Adaptive filtering of the training sequence is utilized to refine the beamformer weights and to compensate for different sampling phase offsets of the cochannel signals. A multichannel matched filter with a sliding data window is also examined for burst detection. The receiver consists of cascaded stages such that the received antenna data are processed iteratively to improve the reliability of the final bit decisions. The performance of the multistage receiver is evaluated via computer simulations using simulated and real GSM data collected by an array of eight antennas.

A GMSK Receiver with Beamformer Weight Refinement

“…However, that beamforming algorithm assumes a known channel to construct the variable reference signals. Instead of this assumption, we propose a practical beamforming algorithm that utilizes filtered training sequences [14]. Specifically, a three-tap constrained adaptive filter is INTERFERENCE implemented as shown in Fig.…”