Events Calendar

ECE PhD Thesis Defense: Timur Zirtiloglu

Title: Co-Design of Radio Frequency Architectures and Communication Algorithms for Energy-Efficient Wireless Systems

Presenter: Timur Zirtiloglu

Advisor: Professor Rabia Tugce Yazicigil

Chair: Professor Selim Ünlü Committee: Professor Rabia Tugce Yazicigil, Professor Ajay Joshi, Professor David Starobinski, Doctor Andreia Cathelin (STMicroelectronics), Professor Negar Reiskarimian (MIT)

Google Scholar: https://scholar.google.com/citations?user=3FIYhVEAAAAJ&hl=en

Abstract: The evolution of wireless communication continues to push radio systems toward more demanding operating conditions. Future wireless systems are expected to deliver high data rates, low latency, and broad coverage, all while maintaining strict power efficiency, robustness against interference and errors, and the ability to support multiple frequency bands and communication standards. This places significant stress on Radio Frequency (RF) front-end architectures, which are often limited by energy efficiency, linearity, and various other hardware limitations. These challenges lead to a growing performance gap between the capabilities predicted by numerical algorithmic evaluations and the performance reported by practical RF hardware implementations. This thesis claims that this performance gap can be addressed by joint co-design of RF front-end architectures and communication algorithms, substantially boosting system performance and enabling more energy-efficient, resilient, and adaptive wireless systems and integrated circuits (IC) design. To support this claim, this dissertation presents two key contributions: 1) a Task-Specific Multiple-Input Multi-Output (MIMO) receiver that focuses on adaptive signal recovery under front-end hardware constraints, and 2) an energy-efficient Optimal Modulation Transmitter that enhances error correction through non-uniform symbol constellations.

The first contribution introduces a Task-Specific MIMO receiver developed for reliable signal reception with low power consumption under specific task constraints, such as limited quantization of Analog-to-Digital Converters (ADCs), the presence of strong spatial blockers, and the finite resolution of analog pre-processing elements.

The second contribution presents an Optimal Modulation Transmitter that utilizes non-uniform symbol constellations and variable-length bits-to-symbol mapping, assisted by a Guessing Random Additive Noise Decoding (GRAND)-inspired algorithm on the receiver side to enhance error rate performance. The energy efficiency of the system is supported by the integration of deep power back-off (PBO) efficiency enhancement techniques.

The presented receiver and transmitter implementations demonstrate the capability to advance the state-of-the-art systems toward next-generation wireless front-end architectures.