Estimation and compensation of I/Q imbalance in OFDM-based systems

Conventional OFDM communications systems employ a super-heterodyne architecture for RF-to-baseband conversion. Thanks to its lower complexity, Zero-IF, or equivalently direct-conversion architecture becomes a promising candidate and is getting more attention in recent years. However, one main drawback of this low-cost alternative is the introduction of I/Q imbalance, which seriously degrades the signal quality and hence the system performance. In the first part of this dissertation, we evaluate the impact of frequency-independent I/Q imbalance together with two other common front-end non-idealities, the carrier frequency offset (CFO) and the phase noise, on the average SINR performance of a coded OFDM system. In particular, an approximate closed-form expression is derived for multipath fading channels with arbitrary modulation format as a function of the front-end non-idealities. From the results derived herein, it is shown that these front-end non-idealities impose a cap or ceiling on the average SINR which cannot be boosted further by applying a higher SNR. For the second part of the thesis, a novel low-complexity estimation and compensation algorithm is proposed for handling the often overlooked but increasingly important case of joint frequency-dependent I/Q imbalance at both the transmitter and the receiver. Our simulation results showed that good error probability performance is obtained with low computational complexity using our proposed algorithm at various imbalance levels