Open AccessOver‐sampling effect in distributed Alamouti coded OFDM with frequency offset
Author(s) -
Kim Bongseok,
Choi Kwonhue
Publication year - 2016
Publication title -
iet communications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.355
H-Index - 62
eISSN - 1751-8636
pISSN - 1751-8628
DOI - 10.1049/iet-com.2016.0083
Subject(s) - orthogonal frequency division multiplexing , fading , carrier frequency offset , fast fourier transform , computer science , algorithm , coherent sampling , sampling (signal processing) , frequency offset , electronic engineering , interference (communication) , channel (broadcasting) , mathematics , telecommunications , engineering , detector
Unlike the conventional distributed Alamouti coded orthogonal frequency division multiplexing (OFDM), over‐sampling substantially improves the performance of frequency asynchronous distributed Alamouti coded (FADAC) OFDM. This exclusive effectiveness of over‐sampling in FADAC‐OFDM comes from two factors. One is the unique characteristics of residual inter carrier interference term in FADAC‐OFDM and the other factor is moving away the cyclic harmonic inter‐carrier interference (ICI) by over‐sampling. In addition, the authors confirm that over‐sampling factor of 2, i.e. 2 N point fast Fourier transform (FFT) is sufficient to move the cyclic harmonic ICI. From this investigation, the authors propose 2 N point FFT FADAC‐OFDM. The proposed scheme achieves the significantly improved performance not only in the flat fading channel but also in the selective fading channels. For example, while the non‐oversampled FADAC‐OFDM has the unacceptable error rate level near to or above 0.1 at the band edges even with high signal to noise ratio, the proposed scheme achieves near intersymbol interference‐free performance. This leads the proposed scheme to achieve 50% reduction of the required number of null subcarriers to achieve ICI free performance compared with the non‐oversampled FADAC‐OFDM. Moreover, in terms of implementation feasibility and computational complexity, the proposed scheme is shown to be acceptable.