Improved LPC-Based Fronthaul Compression With High Rate Adaptation Resolution

Abstract

Abstract-This paper presents a fronthaul signal compression scheme based on linear prediction coding (LPC) adapted to orthogonal frequency division multiplexing (OFDM) signals. The proposed method is capable of providing fine tuning of the compression factor, which is an alternative to legacy compression methods that tune the compression factor by changing the discrete number of bits of the quantizer and, consequently, are only able to do so with coarse resolution.

A Survey of the Functional Splits Proposed for 5G Mobile Crosshaul Networks

“…In addition, a maximum jitter of about 40 μs can be tolerated in the fronthaul. The simulations in [78] experiment with linear predictive coding for fronthaul compression, the authors conclude that the proposed method allows fine regulations between the achievable compression factor and the corresponding compression Signal to noise Ratio (SNR) or Error Vector Magnitude (EVM).…”

“…Please note that some of the references for this section are placed here due to estimated locations of the split, where option 7-1 was found most suitable. [5], [12], [26], [27], [29], [30], [31], [33], [34], [38], [41], [48], [49] Simulations: [44], [53], [55], [58], [60], [61], [75], [76], [77], [78] Practical experiments: [79], [80] In this functional split, the Fast Fourier Transformation (FFT) is included locally in the DU. Due to the Fourier transformation the data to be transmitted over the fronthaul interface is represented by subcarriers.…”

“…The project ended in 2017 and has contributed to standardization along with providing several white papers and journal papers [132]. Some of the most recent literature provided by this project are: [63], [78], [96], [100], [134].  The 5G xhaul is an H2020 project considering a Dynamically Reconfigurable Optical-Wireless Backhaul/Fronthaul with Cognitive Control Plane for Small Cells and Cloud-RANs.…”

Analysis of Baseband IQ Data Compression Methods for Centralized RAN

“…overload distortion. UQ performance could be enhanced by dynamically adapting the input signal level to have a trade-off between granular and overload distortions [9]. The loading factor γ is defined as…”

“…The aim was to exploit time correlated samples and led to improved performance however at the expense of substantial increase of computational complexity. In [8] and [9] authors explored the correlation between samples by the well known linear predictive coding, hereby trying to solve the complexity issue introduced by vector quantization. This approach remains however restricted to uplink signal.…”