MMSE based K-best algorithm for efficient MIMO detection

MIMO detector is the key component of modems used in multi-carrier wireless devices for LTE-A and 5G networks. The complexity of the detector is usually higher than for most other parts of the system. Therefore, it is considered as the most critical modem part. Maximum Likelihood (ML) detector is known to give the optimal solution, but at the expense of exponential complexity. Thus, it is mostly used only as a reference algorithm but not as an implementable component. The main target of this MIMO detector study is to find an optimal tradeoff between performance and complexity. Analytical derivations defining a new efficient MIMO detection scheme named MMSE-K Best are presented in the paper. Proposed detector design demonstrates performance approaching ML while its complexity remains close to relatively simple linear detectors, such as Minimum Mean Square Error (MMSE). © 2017 IEEE