Digital Front End Algorithms for Sub-Band Full Duplex

Abstract: Sub-band full duplex is a new communication scheme technology, where a single frequency band is partitioned into sub-bands for downlink (DL) and up-link(UL) transmissions, and both can take place simultaneously. The idea behind the sub-band full duplex development is to improve the throughput, and coverage and reduce the latency of the UL communication by allowing the UL reception during the DL transmission. Several integrated ways enable sub-band frequency isolation, such as antenna’s spatial isolation and signal processing techniques to mitigate interferences. The main challenge for a gNodeB, capable of a full duplex, is self-interference mitigation. A self-interference mitigation technique enables a radio transceiver to transmit and receive simultaneously on a single channel. During the design process, the Digital Front End (DFE) of the Radio Near algorithm (RNA) is considered the main part of the study (with certain limitations) as it involves the algorithms with the use of MATLAB, which are used to mitigate the distortion caused by radio hardware imperfections. Signal processing algorithms are performed on the transmitter side and receiver side of DFE. The transmitter side includes the Crest Factor Reduction (CFR) block followed by Digital Pre-Distortion (DPD) block before the Power amplifier (PA) to linearize PA output. The transmit path also includes the feedback path for DPD which is used to linearize PA. The algorithms for CFR block namely Turbo Clipping and Peak Cancelation Crest Factor Reduction (PC-CFR) were developed and compared for the performance in terms of Adjacent Channel Leakage Ratio (ACLR) and Error Vector Magnitude (EVM). The algorithm for DPD namely 2-stage frequency selective DPD is implemented along with the legacy DPD technique, and the performance was evaluated in terms of ACLR and EVM. Finally, the Self-interference cancelation (SIC) algorithm is implemented at the receiver chain side. Signal to Interference Noise Ratio (SINR) is measured for different Signal to Noise Ratios (SNRs) and different RF cancellations levels to evaluate the system performance at gNodeB.