Noncoherent Joint Decision-Feedback Detection in Multi-User Massive MIMO Systems

Noncoherent detection in multi-user massive MIMO uplink systems is studied. In particular, noncoherent approaches based on decision-feedback equalization (DFE) are presented and assessed. On the one hand, sorted decision-feedback differential detection (DFDD) is attractive for the detection of a particular user over a transmission burst. On the other hand, noncoherent DFE (nDFE) over the users, purely based on statistical channel knowledge, is rewarding. Up to now, DFDD has been used inside the nDFE scheme; the transmission bursts of the users are detected in sequence and the sortings of nDFE and DFDD are optimized individually. In this paper, we present a noncoherent detection scheme where joint user/temporal sorting is carried out. The suited choice of the receive windowing is discussed. Numerical results quantifying the performance gains over stateof- the-art noncoherent schemes and coherent BLAST, taking the non-perfect channel knowledge due to finite-length training sequences into account, are presented.