Capacity for suboptimal receivers for coded multiple-input multiple-output systems

The optimal detector for coded multiple antenna systems is too complex to be implemented. Even approximations to the optimal decoder, such as iterating between detection and decoding, are too complex for many practical systems. Thus, we consider a number of spatial stream decouplers with the assumption of separate decoding and decoupling. The linear zero-forcing decoupler, the linear minimum mean-squared error (LMMSE) decoupler, the successive soft/hard interference canceller, and the parallel soft/hard interference canceller are investigated. The capacity of the channel including these constrained receiver structures is analyzed. It is shown that no other decoupler can achieve larger capacity than the LMMSE decoupler in a sense that it achieves the maximum sub-channel capacities. This contrasts the general belief that interference cancellation schemes are better than linear filters. This is because this general belief is based on the assumption of perfect interference cancellation, which, in practice, requires joint decoding/decoupling. We discuss the likelihood value derivation from the decouplers, which can be used for decoding error correction codes. The performance of parallel transmission of turbo-coded symbols is given to support the constrained capacity analysis.