Physical Layer Security for Satellite based IoT Edge Services with Deep Reinforcement Learning for Energy Efficiency

IoT is perceived as Internet of Threats by many businesses and consumers. With cyber-attacks becoming more sophisticated day by day, IoT data security, privacy and confidentiality is one of the biggest challenges. Security becomes a greater concern when wireless networks are used for IoT connectivity as they are easily vulnerable to security threats such as eavesdropping, jamming, data modification etc.

Another important issue is to power these IoT edge devices which in some scenarios will be located in extremely remote areas. From the consumer’s perspective, post-rollout maintenance of sensor nodes such as replacing batteries in a short amount of time is extremely undesirable. Hence, there is a need to use machine learning techniques particularly deep reinforcement learning so that the IoT devices/agents can become energy self-sufficient by making optimized decisions based on the learnt policies.

This project will address the these issues jointly by developing novel energy management and data transmissions techniques for energy self-sufficient and secure IoT systems using deep reinforcement learning and a robust PLS approach. Furthermore, we plan to test our designed model on mission critical Smart Grid sensor and monitoring scenario to ensure that the developed solutions are efficient, scalable, and implementable for massive IoT systems.


Project Leader:
Dr Tommy Huynh, La Trobe University

PhD Student:
Duaa Fatima, La Trobe University