Bhaskar Ramamurthi

A novel multiple-input multiple-output (MIMO) channel-diagonalization design is proposed for non-simultaneous two-way relaying (NS-TWR). Unlike conventional TWR, the base station in NS-TWR serves two different users - a transmit-only user and a receive-only user. The receive-only user experiences back-propagating interference (BI). The proposed design, which uses linear receivers, cancels the BI and diagonalizes the end-to-end MIMO channels. The diagonalized NS-TWR overcomes the restrictive antenna configurations of existing designs and yields substantial sum-rate improvement over them.

We investigate a joint design of linear precoders and receivers for multiple-input multiple-output non-simultaneous two-way relaying (NS-TWR). Unlike conventional two-way relaying, the base station in NS-TWR performs two-way relaying with two different users - a transmit-only user and a receive-only user (RUE). The RUE experiences back-propagating interference (BI). The proposed design cancels this BI and provides beamforming gain over existing designs. For NS-TWR, we maximize the weighted sum-rate (WSR) through joint power allocation, by solving a sequence of geometric programs. The precoder and the receiver designs as well as the power allocation program are then extended for a multi-user system with multiple transmit-only and receive-only users. With exhaustive simulations, we show that the proposed design provides significantly better WSR than the existing ones. The proposed design is also evaluated in a cellular framework using realistic path loss models, to assess the system-level performance gain achievable.

Two-way relaying (TWR) improves the spectral efficiency of cellular systems by enabling data exchange between two nodes via a relay station (RS) in two channel uses. It is assumed in TWR that a user has data to send and receive from the base station (BS) at the same time. This paper considers a non-simultaneous traffic scenario observed in cellular systems, where one user sends data to the BS and another user receives data from the BS. We consider a novel multiple-input multiple-output (MIMO) TWR protocol to enable communication between two users and the BS in two channel uses. By designing a linear precoder at the decode-And-forward RS, we maximize the sum-rate of this protocol by formulating a sequence of semidefinite programs. The sum-rate performance of this protocol is evaluated in two coverage-limited scenarios in a cellular framework by extensive numerical simulations. It is shown that the protocol results in a dramatic performance gain over conventional protocols.