Improving Performance of Digital Mobile Fronthaul Employing 2-D Vector Quantization With Vector Linear Prediction

Abstract

In this paper, a two-dimensional (2-D) vector quantization with vector linear prediction (VLP-VQ) is proposed to improve the transmission performance of the digital mobile fronthaul (MFH). Compared with the conventional vector quantization based on k-means clustering algorithm, the VLP technique can decrease the mean square error of the orthogonal frequency division multiplexing (OFDM) samples by around 5-50% and thereby greatly reduce the quantization noise at the transmitter. In the proof-of-concept experiment, the proposed scheme is demonstrated in a 25-km 10 Gbaud/λ four-level pulse amplitude modulation intensity modulation and direct detection system. In the 5G new radio environment, the experimental results proved the ability of the proposed digital MFH to support 27, 20, 16, and 13 100-MHz OFDM with 4, 16, 64, and 256 quadrature amplitude modulation based antenna-carrier containers at quantization bits per sample of 3, 4, 5, and 6. Compared with the conventional 2-D vector quantization, the error vector magnitude performance of the proposed scheme is significantly improved with the same number of quantization bits.