Radio Frequency Machine Learning

Summary

More about Professor Philip Leong Stay in touch

The aims of this research is to develop new FPGA-based systems that can localise, classify, detect, segment, and interpret radio signals with improved speed and accuracy using domain-specific architectures, with an ultimate goal of achieving RF scene understanding in the same way computer vision systems can understand video scenes.

Research Area:

Radio frequency, Machine learning, FPGA

Supervisor

Professor Philip Leong.

Research location

Electrical and Computer Engineering

Synopsis

Recent advances in machine learning have enabled computers to process and understand video scenes with excellent speed and accuracy. However, our ability to interpret radio frequency (RF) signals is not as mature, and many challenges in RF scene understanding remain unsolved research problems.

Field programmable gate arrays (FPGAs) are an ideal platform for implementing RF machine learning (RFML) systems as they can integrate a software-defined radio (SDR), digital signal processing (DSP) operations, and machine learning on a single chip, leading to improvements in latency, energy consumption and bandwidth as massively parallel on-chip computing and interconnect resources can be utilised.

Unfortunately, radio frequency machine learning (RFML) systems are not as mature as computer vision systems. In particular, some fundamental problems to be solved include: finding good feature representations, achieving good generalisation in real-world settings, and devising high-speed FPGA implementations.

Additional information

Offering:

The successful candidate will be awarded a scholarship for 3.5 years at the RTP stipend rate (currently $41,753 in 2025) subject to satisfactory academic performance. International applicants will have their tuition fees covered.

Successful candidates:

• Must have a have a Bachelors degree (1st class honours or equivalent) or a Master's degree with a substantial research component.
• Applicants with an interest in Cyclostationary signal processing, Low-precision neural network implementations or Synthetic data generation are strongly encouraged to apply.

How to apply:

To apply, please email Professor Philip Leong the following:

• CV,
• Transcript, and
• Brief explanation on what motivates you to work on this project and why you would be a good choice for this topic. 

Want to find out more?

Opportunity ID

The opportunity ID for this research opportunity is 3622

Other opportunities with Professor Philip Leong