Quantization Noise Shaping for LTE Fronthaul Downlink

Abstract: The modern mobile networks such as the Cloud Radio Network Access (C-RAN) are developing to deal with upcoming requirements in the current generation $4$\textsuperscript{th} Generation (4G) and the upcoming generation i.e. the $5$\textsuperscript{th} Generation (5G). One of the challenges in the modern communication systems is the required high data traffic specially at the fronthaul. Fronthaul in Long Term Evolution (LTE) contains the Baseband Unit (BBU) and Remote Radio Unit (RRU). These two units are connected using a protocol called by Common Public Radio Interface (CPRI). The CPRI is containing a set of fiber links used in the fronthaul to transport data between BBU and RRU in both Uplink (UL) and Downlink (DL). The fiber optic cables have a limited amount of data to transport. This leads to a limitation of transmitting higher data rates in both DL and UL directions. To solve this problem, a compression algorithm is needed to compress the data before transmission. In this thesis a compression algorithm is introduced to compress the complex baseband LTE DL signal for different bandwidths at the BBU. The concept of quantization is used to obtain the compression. The uniform quantization is used to reduce the number of bits. To improve the results of quantization, the concept of Quantization Noise Shaping (QNS) is introduced. The QNS system is based on Noise shaping method. In this, the feedback filtering system is used to shape the quantization noise. This gives a one more bit reduction compared to quantization. Furthermore, we used the oversampling concept by oversample the complex baseband signals to get better results. This gives better shaping and same bit reductions. The simulation is done using MATLAB LTE System Toolbox which offers LTE test complex baseband signals and built in functions. The results are evaluated according to $3$\textsuperscript{rd} Generation Partnership Project (3GPP) standardization by measuring the Error Vector Magnitude (EVM) and Signal to Quantization Noise Ratio (SQNR).