Comparative Performance Analysis of GFDM and OFDM Systems Using LS-Based Channel Estimation

Abstract

One of the new promising multicarrier waveform candidates beyond-OFDM wireless systems is Generalized Frequency Division Multiplexing (GFDM), which offers flexibility, enhanced spectral containment and low-latency transmission. In this work, the performance of a GFDM transceiver system using pilot-assisted LS channel estimation is evaluated in AWGN and Rayleigh fading channels. The analyzed system uses 32 subcarriers, 16 subsymbols, a QAM of order up to 8, a cyclic prefix of 10% and a pulse-shaping filter with roll-off =0.1. We evaluate the performance of GFDM in terms of power spectral density (PSD), bit error rate (BER), symbol error rate (SER) and normalized mean square error (NMSE) and compare it with a conventional OFDM system under similar simulation scenarios. These results demonstrate that GFDM provides improved spectral localization and reduced out-of-band (OOB) emissions compared to OFDM. Nonetheless, OFDM outperforms GFDM in terms of BER, SER and NMSE under the explored channel conditions thanks to its orthogonal format, while GFDM practices a larger self-interference together with approximation due to LS-based channel estimation. These results suggest that GFDM is a good model for spectrally efficient wireless systems, but additional advances in receiver design, equalization and channel estimation techniques are needed to optimize its error-rate performance.