The MoD Grand Challenge, where robots hunt for potential threats in simulated urban warfare, has been won by team led by Hampshire-based Stellar Services. 

“It is not just the robots, it’s the autonomous threat detection software that decides what sort of threat and attaches a confidence level,” project leader Dr Julia Richardson told EW.



Called Saturn, the system has three robots: A high level unmanned aerial vehicle (UAV), a low-level UAV and a ground vehicle.
“You launch the high level UAV which photographs the area using on-board optical and thermal sensors,” said Richardson. “The data is transmitted to a computer, a laptop at the moment, which converts it to a map.”

The low level UAV, which has an optical camera, is then hand launched into environment. At the same time the ground vehicle with optical, thermal and radar sensors moves in.

Together these accumulate more data, guided with minimal manual intervention by the laptop.

“If a window is detected by analysis of the optical images, and the thermal images show it is hot, and the radar images show metal, it could be a sniper,” said Richardson.

Threat detection algorithms merge the data, identify threat type and attach confidence level. “The software produces a situation awareness plan and does mission planning to increase confidence level,” said Richardson. “It works out a route for the ground vehicle avoiding obstacles to the potential threat.”

Manual intervention is possible. “The operator can always double click on a threat and see the original source images,” said Richardson.

Along with Stellar Services: Blue Bear Systems, Cranfield University, SELEX Sensors and Airborne Systems, TRW Conekt, and Marshall Specialist Vehicles formed the winning team.

The Grand Challenge took place over two weeks at the FIBUA (fighting in built up areas) army training centre at Copehill Down in Wiltshire.

The winning team received the RJ Mitchell Trophy, named after the designer of the second world war Spitfire, and a chance of further funding.