Joint ML estimation of I/Q mismatch, DC offset, carrier frequency, and channel for direct-conversion receivers

This paper proposes a technique to jointly estimate and compensate for I/Q mismatch, DC offset and carrier frequency offset in frequency selective channel environments for direct-conversion receivers. It estimates the unknown parameters in a joint maximum-likelihood (ML) sense from a training sequence. In addition, a design method of training sequences is proposed to minimize the mean-square errors (MSEs) of the proposed estimates. It is shown through simulation that the proposed technique can provide significant performance improvement over the existing schemes, and the proposed method of designing training sequences is effective as well.