Background
The growing demand for high data-rates transmission in modern wireless communication systems has motivated the development of broadband wireless technologies to support increased channel capacity with good spectral efficiency. A distinct feature in such systems is carrier aggregation – a configuration in which two or more component carriers are aggregated on the physical layer to support wider transmission bandwidths for higher data throughput. In practice, aggregation of more than two carriers at the transmitter becomes increasingly challenging due to the unrealistically large number of coefficients required for a multi-dimensional predistortion based on standard memory polynomial schemes.
Project Proposal
The goal of this project is to develop a flexible al low complexity neural network based algorithm to simultaneously linearize several channels that are transmitted concurrently through a single power amplifier, taking into account inband and crossband intermodulation distortions. The algorithm should outperform state-of-the-art memory-polynomial based predistortion approaches in terms of sampling rate requirement, bandwidth and implementation complexity. The algorithm will be developed using actual nonlinear power amplifier models, and then evaluated in real lab experiments using a broadband Wi-Fi transmitter.
Requirements
- Good programming capabilities (knowledge in Python – advantage)
- Optimization/Learning courses – advantage
RFIC/Analog circuit design courses – advantage
Instructors
Nimrod Ginzberg (nimrodg@tx.technion.ac.il)