GFSK Demodulation Using Sequential Monte Carlo Technique

This letter applies the Sequential Monte Carlo (SMC) approach along with a low complex detection algorithm to non-coherently demodulate a Gaussian frequency shift keying (GFSK) signal in Bluetooth receivers. Sequential importance resampling (SIR) is used to estimate the state of the GFSK signal, while a low complex detection algorithm is used to decide on the received bit. SIR is the original and most used particle filtering algorithm, which approximates the filtering distribution by a weighted set of particles. The bit error rate (BER) analysis is used as the performance metrics of the proposed receiver at the physical layer. Applying SMC technique to GFSK demodulator shows significant performance improvement compared to several other existing techniques that are presented in the literature. In addition, it does not require a pre-detection filter, which is usually used to reject out-of-band interference. Simulation results show approximately 6 dB of BER improvement over the commonly used limiter discriminator with an integrator (LDI) receiver.