Baseband Assisted RF Impairment Correction: The Case of IQ Imbalance

IQ imbalance is a serious RF front-end impairment and in this work, we study the issue of IQ imbalance correction using a combination of iterative and adaptive signal processing techniques. First, we propose an estimator based on EM algorithm for the case of a Bit Interleaved Coded modulation Iterative Detection (BICM-ID) receiver for burst mode communications. The resulting TURBO IQ Decorrelator is embedded in the BICM-ID loop and is non data aided. The performance of this estimator depends on the errors in the first iteration (due to decision feed back) and thus an adaptive filter is used for the purpose of reducing the bit errors on the first iteration. Adaptive decorrelators are a practical solution for correcting IQ imbalances, but long convergence times associated with adaptive algorithms are not practical in burst mode communications. Adaptive decorrelator working alone needs training phase for convergence, which means loss of power and bandwidth. This motivates one to find techniques to increase the rate of convergence of the learning algorithm for a given packet and one technique is recirculation of the received data over and over during the turbo iterations. We call such a decorrelator as an enhanced adaptive decorrelator. Furthermore, we show how the interaction of the embedded estimator and enhanced adaptive decorrelator working in a synergic manner can correct the crippling effects of IQ imbalances. The performance is simulated for 64QAM under flat fading channel conditions.