Robust Uplink Carrier Frequency Offset Estimation in MIMO-OFDM Systems

Carrier frequency offset estimation in MIMO systems is much more difficult than that in single antenna systems, especially in uplink transmission in a full-multiplexing transmission mode system. That is because in MIMO uplink transmission systems, different mobile nodes use different oscillators (even with different Doppler Shifts), each introducing one unique carrier frequency offset. In addition, more than one transmit antenna may be equipped in one mobile node. If different training sequences (TS) are transmitted at different transmit antennas, self-interference appears. This kind of self-interference makes carrier frequency offset estimation even more complicated. In a multi-user environment with full-multiplexing transmission mode, when performing uplink synchronization, training sequences transmitted by different users may overlap with each other at the receiver (base station). Co-Channel Interference may also result in some overlap of training sequences when performing uplink synchronization. This kind of overlap may degrade the estimation performance or even cause some estimators' failure. How to mitigate the overlap distortion is critical in MIMO uplink synchronization. This paper first analyzed the inter-TS-interference in MIMO uplink synchronization. For an unbiased carrier frequency, offset estimator, its Cramer-Rao Lower Bound (CRLB) is also derived. Two Lemmas without proof are also provided for MIMO uplink synchronization. With received Signalto- Interference-plus-Noise Ratio (SINR) known at the receiver, a joint carrier frequency offset estimator, named as Joint Best Linear Estimator (JBLUE), is also proposed. Simulation results prove that a great performance improvement with regard to estimation accuracy can be expected by using JBLUE when performing MIMO uplink synchronization.