The British Army and the Defense Science and Technology Laboratory (Dstl) recently completed a several-week field test in Essex, England, using drones equipped with advanced sensors and artificial intelligence technology to quickly search for landmines and unexploded ordnance over a large area to improve demining efficiency and reduce the risk of casualties.
In the force structure of any major military power, a considerable proportion of the force is invested in the search, clearance and detonation of landmines and various types of munitions. This is closely related to the scale and danger of the relevant threats. Take the Russia-Ukraine conflict as an example. Ukraine has now become the country most seriously polluted by landmines and explosive remnants of battlefields in the world, even surpassing Syria and Afghanistan. Estimates show that as much as 67,000 square miles (approximately 174,000 square kilometers) of Ukraine, accounting for nearly 30% of the country's land area, is covered with anti-personnel mines, "butterfly mines" and other dispersed anti-personnel munitions, and anti-vehicle mines, as well as up to about 5 million unexploded artillery shells, cluster munition submunitions and other dumb bombs.
Together with explosives left over from other conflicts, terrorist attacks, and previous wars, EOD and Explosive Ordnance Disposal forces continue to accumulate experience in actual combat and continue to learn how to deal with such threats through painful lessons. Against this background, the UK is advancing the Project Ground Area Reconnaissance and Assurance (GARA), which aims to provide technical support for the British Army's Future Counter-Explosive Ordnance Capability (FCEOC). The British Army is working with Dstl and 33 Engineer Regiment (Explosive Ordnance Disposal and Search) to explore possible ways to speed up and safeguard high-risk explosive ordnance disposal tasks through new technologies.
As part of the project, units conducted several weeks of testing earlier this year in a simulated complex minefield environment. A wide range of mixed ammunition types and materials were deployed in the test site, including metal casings, low metal content and plastic casing devices. Many targets were obscured or buried. Searching and processing using traditional methods is slow and dangerous.
In the latest round of exercises, participants used a multi-rotor drone platform equipped with high-performance optical, thermal imaging, long-wave infrared, magnetometer and other sensors, combined with computer vision algorithms, to transmit the collected data back to Army personnel at the rear for analysis. The AI tools deployed enable rapid retraining when new information emerges, and can automatically locate and identify landmines and unexploded ordnance in field environments. This is equivalent to adding a "reconnaissance layer" between the human operator and the digital system. The UAV system first searches for, identifies and accurately calibrate threats above the ground, and the threat distribution is known before the soldiers arrive. Taking it a step further, if the operator discovers a new threat type on the screen, images and related data can also be uploaded to update the AI model in real time.
For security reasons, many technical details of this new system are still confidential, but from the overall concept, it is incorporated into a three-tier processing process: first detection and classification, then labeling and prioritization of threats, and finally remote clearance through robots, drones carrying explosive charges, or, as a last resort, manual intervention at close range.
Major Mark Fetters, head of the British Army's future counter-explosive capabilities, said that contemporary battlefields are full of various types of explosives, and the introduction of new equipment will enable explosive ordnance disposal personnel to complete tasks faster and free people from direct exposure to explosion hazards. He added that the project team is also evaluating the scope for continued evolution of such capabilities as other technologies mature. As different types of sensors continue to become lighter, more power-efficient, and smaller, they will be able to be integrated onto smaller unmanned aerial platforms, continuing to increase the overall capabilities available to explosive ordnance disposal and search personnel.
According to reports, the relevant information about this AI drone demining program comes from public materials released by the British Ministry of Defense.