Alternative Viterbi Detection Metrics for GFSK Receivers: A Hardware Reduction Approach

Abstract

Gaussian frequency shift keying (GFSK)is a digital modulation scheme with both attractive spectra and high power efficiencies. Modulation schemes with these characteristics often require a maximum likelihood sequence estimator (MLSE)implemented by means of Viterbi Algorithm (VA). The main drawback of VA is complexity implementation due to the filter bank needed to calculate the metrics, the large number of states required to describe the signal and the employment of squaring and multinpication operations in calculating metric cost. This paper discusses alternative metrics for VA in GFSK schemes that avoids the use of multiplication and squaring operations, achieving a reduction in implementation complexity. Moreover, the tradeoff between complexity and performance (in terms of bit error rate, BER)for the proposed metrics is shown under additive white Gaussian noise (AWGN).