Neural Network-Based Digital Predistortion and Self-Interference Cancellation in a Quadrature Balanced Full Duplex Transmitter

Author(s):  
Erez Loebl ◽  
Nimrod Ginzberg ◽  
Emanuel Cohen
2021 ◽  
Vol 25 (1) ◽  
pp. 181-185
Author(s):  
Mohamed Elsayed ◽  
Ahmad A. Aziz El-Banna ◽  
Octavia A. Dobre ◽  
Wanyi Shiu ◽  
Peiwei Wang

Author(s):  
G. T. Watkins

Abstract Full duplex (FD) could potentially double wireless communications capacity by allowing simultaneous transmission and reception on the same frequency channel. A single antenna architecture is proposed here based on a modified rat-race coupler to couple the transmit and receive paths to the antenna while providing a degree of isolation. To allow the self-interference cancellation (SiC) to be maximized, the rat-race coupler was made tuneable. This compensated for both the limited isolation of the rat race and self-interference caused by antenna mismatch. Tuneable operation was achieved by removing the fourth port of the rat race and inserting a variable attenuator and variable phase shifter into the loop. In simulation with a 50 Ω load on the antenna port, better than −65 dB narrowband SiC was achieved over the whole 2.45 GHz industrial, scientific and medical (ISM) band. Inserting the S-parameters of a commercially available sleeve dipole antenna into the simulation, better than −57 dB narrowband SiC could be tuned over the whole band. Practically, better than −58 dB narrowband tuneable SiC was achieved with a practical antenna. When excited with a 20 MHz Wi-Fi signal, −42 dB average SiC could be achieved with the antenna.


2014 ◽  
Vol 3 (4) ◽  
pp. 405-408 ◽  
Author(s):  
Zhaojun He ◽  
Shihai Shao ◽  
Ying Shen ◽  
Chaojin Qing ◽  
Youxi Tang

Sign in / Sign up

Export Citation Format

Share Document