News article

At the heart of the UK’s largest multi-domain AI defence trial

29/05/2025
Back to all news

Picture this. Portland Harbour, a normally quiet coastal area of Dorset, awash with warships, main battle tanks, and a Chinook, conducting a military exercise to collect essential data for automatic target detection and recognition systems. That was the scene during one week in March, in a comprehensive trial we co-ordinated, known as WINTERMUTE 3.

Picture this. Portland Harbour, a normally quiet coastal area of Dorset, awash with warships, main battle tanks, and a Chinook, conducting a military exercise to collect essential data for automatic target detection and recognition systems.

That was the scene during one week in March, in a comprehensive trial known as WINTERMUTE 3. Coordinated by Frazer-Nash, this was the latest in a series of three trials by the Defence Science and Technology Laboratory (Dstl).

The aim of WINTERMUTE 3 was to better understand various targets to advance surveillance technologies, including Automatic Target Recognition (ATR), by collecting the right data with the right meta-data sets. With advances in Machine Learning (ML) expected to revolutionise ATR performance in complex systems, the exercise sought to understand how this form of Artificial Intelligence (AI) could be improved across multiple domains and platforms.

Mass data was gathered through synchronised formations and attack profiles by different military units in numerous scenarios. The multispectral footage and, plus relevant meta-data and contextual information, developed during the exercise will now be exploited in advancing software development across defence. 

Leading the way as a trusted partner

Our organisation of this large-scale exercise was no mean feat with over 15 suppliers, and multiple terabytes of data to manage. Taking less than five5 months to plan and implement, the task’s ultimate aim was to ensure the collected data had high exploitation value for Dstl and could help inform the future development of defence technology.

Having previously supported Dstl with their previous two WINTERMUTE trials, we seamlessly stepped into the lead role to bring partners and assets together for this week-long trial. We were also able to utilise our expertise in defence research and development, and our own Unmanned Air Systems team was there, gathering data to inform our algorithm and systems development.

Our WINTERMUTE Business Manager, James Brodier, had this to add:

“The WINTERMUTE 3 trials are the epitome of an enterprise approach to data gathering. With targets available across land, air and sea, co-ordination was ambitious and challenging, but our collective effort meant Dstl and wider industry now has meaningful data to advance their work. We are proud to have been at the centre of this important exercise that is valuable for both the UK and our allies.”