Supervisor

Dr Saeed Rehman
Rehman, Saeed (Dr)
saeed.rehman@flinders.edu.au
View Flinders profile

Project description

Wireless communication systems are not immune to the development in AI. Wireless systems and technologies are becoming increasingly complex due to the demand for data and different applications from ground-based to satellite communication. The number of parameters characterising a system increases substantially; hence, writing typical traditional algorithms that consider all the parameters becomes an onerous task. Therefore, AI techniques that essentially learn from the operations and apply those algorithms based on learning are emerging these days. As demonstrated for vision, malware and network anomaly classification, AI algorithms are susceptible to adversarial attacks. Adversarial attacks mislead the AI classification/detection model to classify the adversarial sample as a different class than the original class of the sample. Adversarial samples can be generated by injecting extreme small perturbations to the input, which is indistinguishable from a normal input. This research theme aims to analyse adversarial attacks on AI-enabled wireless communication. A fully trainable AI-enabled radio can sense the operating environment and autonomously generate secure, high throughput and resilient tactical waveform to support mission-critical communications. The following research questions are related to this theme and can be separate projects  
  1. How will adversarial attacks compromise an AI-enabled wireless communication system?
  2. How to design a reliable AI-enabled waveform in the presence of an AI adversary?
  3. How to evaluate the existing AI communication system for adversarial attacks
  4. How to develop an adversarial Testbed and demonstrate a proof of concept using software-defined Radio

Assumed knowledge

Machine learning and Wireless communication concepts

Register interest in this project


Note: You need to register interest in projects from different supervisors (not a number of projects with the one supervisor).
You must also contact each supervisor directly to discuss both the project details and your suitability to undertake the project.