A team of AI driven unmanned ground vehicles navigates obstacles, locates fires, and coordinates suppression tasks in simulated high risk industrial scenarios.
Cyborg Dynamics Engineering, in collaboration with Griffith University, has demonstrated a team of artificial intelligence-powered unmanned ground vehicles capable of autonomously detecting and extinguishing fires. The collaborative robot system is intended to support firefighting operations by enabling remote response in potentially dangerous environments.
According to the research team, the technology could help reduce direct human exposure to hazardous conditions while assisting with navigation, fire detection and coordinated suppression tasks. Potential application areas include mining sites, industrial facilities and other high-risk locations where remotely operated vehicles are already used.
In the soft trials, the system was evaluated in simulated and hybrid simulation-physical environments. A physical unmanned ground vehicle navigated around obstacles while coordinating with up to four simulated robotic teammates to locate and suppress multiple simulated fires. The trial recorded a 99.67% success rate in navigating and extinguishing two fires.
The robots were trained using a multi-agent reinforcement learning framework supported by a structured three-stage curriculum. Training progressed from single-robot navigation tasks to multi-robot obstacle avoidance and then to coordinated firefighting scenarios involving multiple robots and fire sources. This approach enabled the system to perform both low-level navigation and higher-level collaborative task allocation.
Current firefighting unmanned ground vehicles deployed in sectors such as mining are typically remotely controlled by human operators. The research focuses on automating elements of low-level control and enabling coordinated behaviour across multiple agents. By processing data from various onboard sensors, the system is designed to support faster decision-making within the tested scenarios.
Ryan Marple, Co-founder and General Manager at Cyborg Dynamics Engineering says, “The future of these kinds of vehicles – and the focus of this research – is the automation of low-level control and swarming behaviour across multiple agents. Such autonomous swarms can respond to complex situations in a way that just isn’t possible with direct manual control.”
