Efficient MMSE design for joint MIMO processing in analog network coding schemes

A new generalized iterative approach (GIA) is proposed for jointly designing the precoders, relay processors and decoders for analog Network Coding schemes. The considered criterion is minimum mean square error (MMSE). The proposed approach is the same as our previously developed GIA except that it uses single user multiple-input multiple-output (MIMO) closed-form solutions for the precoders. This change reduces the number of design variables by two since the precoding Lagrange multipliers are no longer needed. The proposed approach is compared with our previously developed GIA design in terms of convergence rate, BER and MSE. Numerical results show that both GIA's achieve similar MSE and BER performances and that the proposed GIA may provide faster convergence speeds.