Spectrum Sensing of Unsynchronized OFDM Signals for Cognitive Radio Communications

Abstract

Sensing of orthogonal frequency division multiplex- ing (OFDM) signals at low signal-to-noise ratio (SNR) is of great importance for cognitive radio (CR) communications due to the wide applications of OFDM in many existing and evolving broadband wireless communications. In-band pilots, multiplexed with the data-carrying subcarriers, provides one distinct feature of OFDM signals. The objective of this paper is to investigate the application of in-band pilots in OFDM signal spectrum sensing, particularly when synchronization is not achievable at very low SNR. In this paper, the proposed sensing technique is to match the received OFDM signals with a self-defined pilot tone filter. Due to the extremely low requirement on sensing SNR for CR applications, synchronization of the sensing receiver and OFDM signals usually is not achievable. As a result, carrier frequency offset (CFO) and timing offset have to be dealt with during the sensing process. With the hybrid domain signal processing, the proposed sensing technique is very robust to both timing and frequency offset. We also apply the proposed technique to DVB-T to verify our proposal and subsequent analysis. Robust performance with very low false alarm probability and short sensing time were achieved under various channel conditions including Rician and Rayleigh channels, in the presence of random timing offset and CFO.