Effects of frequency offset compensation error on channel estimation for OFDM system under mobile radio channels

• Published in: Signal processing Vol. 83; no. 12; pp. 2621 - 2630
• Main Authors: Park, Seung Young, Seo, Bo Seok, Kang, Chung Gu
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.12.2003; Elsevier Science
• Subjects: Applied sciences; Channel estimation; Detection, estimation, filtering, equalization, prediction; Exact sciences and technology; Frequency offset; Information, signal and communications theory; OFDM; Pilot carrier arrangement; Rayleigh fading; Signal and communications theory; Signal, noise; Telecommunications and information theory; Pilot carrier arrangement; Frequency offset; Rayleigh fading; OFDM; Channel estimation; Performance evaluation; Mobile radiocommunication; Time variable channel; Bit error rate; Rayleigh channels; Pilot carrier arrangement: Channel estimation; Multipath channels; Orthogonal frequency division multiplexing; System identification; Least mean squares methods
• ISSN: 0165-1684; 1872-7557
• DOI: 10.1016/j.sigpro.2003.08.002

In this paper, we analyze the effects of frequency offset compensation on the pilot-assisted channel estimation for the orthogonal frequency division multiplexing system under a time-variant multipath fading channel. The analysis focuses on the pilot arrangement of a pilot-assisted channel estimation scheme, in which the minimum mean-square error channel estimation is employed to exploit both time- and frequency-domain correlation characteristics of the frequency response of time-variant multipath fading channels. We also investigate the effect of the length of the training signal used for frequency offset estimation on the channel estimation and subsequently, on the bit error rate performance, subject to varying signal-to-noise power ratio under mobile radio channel. Numerical results show that the maximum allowable pilot carrier spacing in time-domain is critically affected by the frequency offset error. In addition, a proper length of the training signal must be configured to suppress the noise influence while reducing a systematic error due to the time-variant behavior of the channel.