Remote Controlled Trucks Are Delivering Aid Where It Is Deadliest

Delivering food in a war zone or a flood plain is one of the most dangerous jobs in the world. The United Nations World Food Programme is now preparing trucks that can make those final kilometres without a driver on board. Its AHEAD project, built with the German Aerospace Center, fits amphibious SHERP trucks with cameras, sensors and AI so a human operator can steer them remotely through the most dangerous terrain on Earth. The reason is simple: WFP drivers have died doing this work, and a vehicle that can be driven from safety means nobody has to sit behind the wheel when the route runs through a conflict zone or a flooded river crossing.

The project drew wide attention at the AI for Good summit in Geneva in July 2026, where the hulking white vehicle sat on the show floor like a Martian rover. Bernhard Kowatsch, who heads WFP's innovation division, called the technology a potential game changer that should allow the agency to reach people who otherwise never would have been reachable.

A Truck That Floats

The SHERP is a Ukrainian built all terrain vehicle designed for places roads do not exist. Its giant tires let it float across water, drive through swamps and flooded rivers, and climb over obstacles up to one metre high. WFP already runs these trucks with human drivers in Sudan, South Sudan and Uganda, where seasonal flooding regularly cuts off entire communities for months at a time.

Those deliveries carry real risk. WFP has lost numerous drivers over the years on such routes, which is why the agency asked DLR, Germany's aerospace research agency, to make the vehicles remote controlled. The result is the AHEAD project, short for Autonomous Humanitarian Emergency Aid Devices, a collaboration that has been running since 2019.

Why AI Is Essential

Remote driving sounds straightforward until you consider the conditions. There are no streets and no lane markings. The vehicle must ford rivers, cross soft ground and avoid debris, all while streaming video and sensor data back to an operator who may be kilometres away with a lag on the connection. Armin Wedler, who coordinates the AHEAD project at DLR, told AFP that the team built on decades of remote control research, but stressed that the system would not be able to process everything without AI.

Machine learning helps the vehicle make sense of rough terrain, fuse camera and sensor feeds, and keep the truck stable and on route even when the control link degrades. The concept places a human operator in a shipping container converted into a control room in a safe area, who takes over the vehicle only for the last, most treacherous leg of its journey. Wedler was also clear about limits. The team deliberately chose to keep a human in the loop rather than make the vehicle fully autonomous, because complex humanitarian situations, such as a truck suddenly surrounded by desperately hungry people, are impossible for any autonomous algorithm to handle safely.

From Bavarian Test Sites to Africa

The technology has already been demonstrated in the field. In tests near Munich in late 2022 and again in summer 2024, the team showed remote driving on land. Then in June 2024 came the milestone: for the first time, the vehicle was remotely driven on land and through water, delivering a ration of WFP food to a remote island in southern Germany while its operator sat in a control centre about 500 metres away.

In 2025 the project moved from prototyping into a structured multi year validation phase, backed by new funding from the Bavarian Ministry of Economic Affairs. That roadmap includes large scale civil disaster protection exercises in Germany and humanitarian validation in Africa, with field tests planned in Uganda in 2028. The targets are the scenarios where drivers are most at risk: conflict zones such as Sudan, minefields, and recurring floods in countries like South Sudan and Mozambique.

The Bigger Picture

AHEAD sits inside a broader shift in humanitarian logistics, where agencies are testing AI across the delivery chain. At the same Geneva summit, the UN refugee agency UNHCR showed a legal virtual assistant that helps lawyers representing refugees quickly find the rights available under country specific legal frameworks, something its developers describe as life saving for many refugees. A UN initiative called DISHA, working with partners including Google and McKinsey, uses AI on anonymised mobile phone data to spot mass population movements during disasters and rapid satellite image analysis to assess building damage after events such as the recent earthquakes in Venezuela, so responders go to the right place at the right time.

What sets the truck apart is that it moves the technology from software into physical danger. If the 2028 African field validation succeeds, the same kit could be retrofitted to other vehicles in humanitarian fleets, and the design is being developed so other organisations, not only WFP, can adopt it.

A Deliberately Human Design

There is a quiet lesson in how this project is built. In an era of hype around fully autonomous vehicles, the engineers chose a design where AI handles perception and stability while a person makes the decisions that matter. The goal was never to remove humans from aid delivery. It was to move them out of the blast radius. For the communities on the far side of the flood, the difference between a truck that comes and a truck that cannot come safely is measured in meals, medicine and lives.