Robust Learning-Based ML Detection for Massive MIMO Systems with One-Bit Quantized Signals
share
In this paper, we investigate learning-based maximum likelihood (ML) detection for uplink massive multiple-input and multiple-output (MIMO) systems with one-bit ADCs. To resolve large dependency of learning-base detection on the training length, we propose two learning-based one-bit ML detection methods: a biased-learning method and a dithering-and-learning method. The biased-learning method keeps likelihood functions with zero probability from wiping out the obtained information through learning, thereby providing more robust detection performance. Extending the biased method to the system with the knowledge of signal-to-noise ratio, the dithering-and-learning method estimates more likelihood functions by adding dithering noise to the quantization input. The proposed methods are further improved by adopting post likelihood function update, which exploits correctly decoded data symbols as training pilot symbols. Simulation results validate the detection performance of the proposed methods in symbol error rate.
READ FULL TEXT