Gaussian Message Passing Detection With Constant Front-Haul Signaling for Cell-Free Massive MIMO

The cell-free concept has garnered great attention in the industry and academia and is envisaged to be a promising technology for beyond fifth generation (5G) networks. However, the extremely high front-haul and computational overhead of the original cell-free network with centralized or star topologies limits its practical adoption. In this work, we propose a novel non-linear sequential detection method based on the Gaussian message passing algorithm (GMP) for the sequential topology, i.e., radio stripe network, which requires
constant
front-haul signaling as the number of access points (APs) increases. Specifically, each AP iteratively calculates means and variances locally and fuses them with the received messages from the preceding AP based on the product principle of the multivariate Gaussian probability density function. The fused messages are forwarded to the next AP to calculate the
a priori
probability with constant front-haul signaling, which is independent of the number of APs. Simulation results demonstrate that the proposed sequential GMP detector achieves an attractive trade-off between bit error rate (BER) performance, computational complexity, and front-haul requirements.