CADR-RAS Webinar: Evaluation of Decision Techniques for Context Aware Collaborative Active Protection System
- Date: Tue, 24 Jun 2025, 2:00 pm - 3:00 pm
- Location: Online event
- Cost: Free
- More information: Register
- Contact: Gregory Davis CADR-RAS Centre Manager
- Email: cadr_ras@adelaide.edu.au
- Associate Professor Claudia Szabo School of Computer Sciences, The University of Adelaide
- Eyoel Teffera Systems Engineer, Defence Science Technology Group (DSTG)
Land based active protection systems (APS) have been demonstrated to be effective in improving the self-protection of vehicles. The concept of extending the bubble of protection beyond a single vehicle to multiple vehicles is expected to enhance the survivability of the team of vehicles.
An important aspect of realising this next generation APS concept is the networked collaborative decision making to aid in context awareness and threat mitigation actions. This project is focused on extending this self-protection capability by harnessing recent developments in the AI/ML space to allow vehicles to collaborate and protect shared assets. Our previous work has performed an extensive analysis of various decision making techniques and their performance on a sample scenario. We defined quantitative and qualitative metrics to evaluate algorithm effectiveness, and identified challenges, trade-offs and limitations of existing approaches as they are applied in contested and dynamic environments. Our analysis highlights that multi-agent systems, reinforcement learning, and multi-agent reinforcement learning in particular are good candidates to be applied within APS working in contested and dynamic environments, with the need for extended evaluations across a variety of scenarios. In this project, we aim to integrate and evaluate multi-agent reinforcement learning approaches and evaluate their performance and effectiveness in a sparsely connected environment.
We look forward to having you attend the event!