Antenna Array Absolute Self-Calibration and Application to Separate Tx/Rx Array Full Duplex MIMO

Full Duplex communication is one of the promising candidate technologies for 5G. As is well known, the key challenge is the mitigation of self interference at the receiver of the Full Duplex system from its own transmitter. With separate transmit (Tx) and Receive (Rx) antenna arrays, the transmitter can beamform in such a manner as to minimize the self interference. The number of constraints this poses on the Tx beamformer would be limited if the internal MIMO channel between Tx and Rx arrays would have limited rank. Hence, in this work, we focus on characterizing the internal channel between antenna subsets of an array. An impediment to the determination of the internal propagation channel is that the estimated (digital domain) channel also involves the impact of the Tx and Rx front ends. Hence, it becomes necessary to estimate these absolute (Tx and Rx) calibration parameters as well, for which we propose two approaches. In a first approach, we assume a parameterized path loss based model for the internal propagation channel and estimate in an alternating fashion the calibration and path loss parameters. In the second approach, we exploit the symmetry of the MIMO propagation channel and temporal variation of the calibration parameters to estimate all unknowns without any further model assumptions. We provide a measurement based comparison between the different estimation approaches and apply the calibration to exhibit the singular value profile of full duplex MIMO internal propagation channels in our testbed. The proposed absolute self-calibration has, of course, many other applications such as for the antenna array manifold in Direction-of- Arrival (DoA) estimation and DoA based beamforming.