Fei Wang and Shaojie (Kyle) Xu have been chosen for the 2018 Qualcomm Innovation Fellowship. They are one of eight teams from United States universities chosen for this honor; the acceptance rate for this competition was 4.6 percent.
Wang and Xu are Ph.D. students in the Georgia Tech School of Electrical and Computer Engineering (ECE). Wang is advised by Hua Wang, who leads the Georgia Tech Electronics and Micro-System Lab (GEMS) and holds the Demetrius T. Paris Junior Professorship in ECE. Xu is advised by Justin Romberg, who is the associate chair for research in ECE and holds the Schlumberger Professorship. They will also be advised by a mentor from Qualcomm on their project.
The title of Wang’s and Xu’s award-winning project proposal is "An Artificial-Intelligence (AI) Assisted Mm-Wave Multi-Band Doherty Transmitter with Rapid Mixed-Mode In-Field Performance Optimization and Digital Pre-Distortion Compensation.” Wireless transmitters govern the energy efficiency, spectral efficiency, and the communication quality-of-service (QoS) in most wireless communication links. This is particularly true in future mm-wave 5G MIMO systems that rely on large numbers of high-performance transmitters working in close conjunction with antenna arrays and other electronics.
Doherty transmitter architecture is gaining increasing attention due to its high energy efficiency, large modulation bandwidth, low baseband processing overhead, and wideband mm-Wave carrier coverage. However, Doherty transmitters require extensive performance tuning in practice, and despite their wide utility infrastructure/base-stations, their application in mobile platforms still remain elusive.
To address these issues, Wang and Xu have proposed an AI-assisted assisted “smart” mm-Wave Doherty TX architecture that achieves high-performance adaptive operation enabled by a built-in machine-learning core for dynamic, Doherty-specific performance optimization and digital pre-distortion (DPD) compensation. This project is one of the first research explorations on leveraging AI to radically enhance the performance and reliability of RF/mm-wave front-end electronics in in-field and on-the-fly operation environments.