Modeling and Mitigation of Nonlinear Distortions by using Neural Networks for LTE-A Wireless Transmitters
Abstract
A two-box DPD system based on the cascade of a memory predistortion model and a memoryless predistortion model is proposed. The memory predistortion model is designed by using an ARVTDNN, while the memoryless predistortion model is designed by using a memoryless ARVTDNN.
Considering the signal at the output of the PA linearized by the proposed two-box DPD system, one will notice that its ACPR is demonstrated by measurement results to have a better performance by 3-5 dB than that of an existing two-box polynomial-based DPD system. Most importantly, the two-box polynomial-based DPD system does not meet the spectrum emission mask of -45 dBc. In addition, the proposed two-box ARVTDNN-based DPD system meets the ACPR requirements for the observation bandwidth of as low as 155 MHz.
Description
Keywords
Engineering, Engineering--Electronics and Electrical
Citation
Wang, D. (2016). Modeling and Mitigation of Nonlinear Distortions by using Neural Networks for LTE-A Wireless Transmitters (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/26939