Abstract
In this paper, a blind method to detect and suppress narrow-band interference (NBI) for multi-band orthogonal frequency-division multiplexing (MB-OFDM) based ultra-wideband (UWB) communications is proposed. The method utilizes mean-square-error (MSE) statistics in the frequency domain to detect subcarriers damaged by NBI. Approximating the MSE statistics by a decision-feedback (DF) scheme creates a blind solution without extra pilots or any other training sequences. After simplification, the computation overhead of the proposed DF method can be accomplished by only a multiplier and nine adders. Therefore, this novel method requires low computation complexity and is suitable for hardware implementation. The detection result is beneficial for band-group selection and interference detection and avoidance (DAA) mechanisms for MB-OFDM systems. Simulation results show that conventional spreading methods may fail to combat NBI as the signal-to-interference ratio (SIR) decreases. Rejecting interfered subcarriers according to the proposed method can improve system performance, reducing the bit error rate (BER) from around 10 -2 to below 10 -4 . © 2010 IEEE.