A Decentralized Reciprocity Calibration Approach for Cooperative MIMO

As a key technology to improve the coverage and throughput of distributed wireless network, cooperative multiple-input multiple-output (CO-MIMO) system jointly manipulates the precoders of multiple distributed base stations (BSs) to serve multiple mobile stations, simultaneously. However, in time division duplexing mode, independent ambiguity multiplicative factors induced by practical transceiver radio frequency circuits will break the reciprocity between the uplink–downlink channels, resulting in the failure of cooperative precoding in multiple BSs. In this paper, we develop an efficient decentralized reciprocity calibration approach with low overhead and complexity for CO-MIMO systems. Specifically, the calibration problem of CO-MIMO is first modeled as a least square (LS) problem. Based on the decomposition theory, the LS problem is decomposed into a master problem and several subproblems. In order to avoid gathering channel state information from distributed BSs, the master problem is solved jointly and iteratively at the BSs’ side, while the subproblems are solved independently at each BS. According to the decomposition results and the solving strategy, the calibration process is described as a decentralized algorithm. The simulation results demonstrate that better performance could be achieved compared with the centralized approach.