When we think about rebuilding after a natural disaster, the first image that comes to mind is of human teams taking risks amidst unstable rubble, but Japan, in collaboration with Swiss researchers, showed that this could be about to change.

At Tsukubase it revealed a giant attached robotic hand, an excavator capable of picking up heavy or fragile debris with equal precision, bringing a new level of safety to rescue operations. The project called CAFE It has already been developed for 5 years and combines brute force with refined intelligence.

The hand was designed with pneumatic actuators that function as muscles moved by air with sensors in the tips of the fingers, it manages to “feel” each object and decide in real time how much force to apply, whether to lift deformed metal blocks or to delicately hold a fragile material.

In the demonstration, the same hand went from a soft grip to a ton firmness without losing control. More than strength, the differential is in the intelligence behind the excavation, the team from the Nara Institute of Science and Technology develops a system that learns digital simulations such as digging, removing obstacles and avoiding collapses.

When transferred to the real world, the robot follows only fixed orders, they adapt in real time by evaluating depth, pressure and stability of the terrain, this is vital in unpredictable scenarios where traditional machines or human workers are at high risk.

One of the problems that this technology promises to solve are so-called natural dams, when landslides block rivers and put entire communities at risk of flooding, instead of relying on workers exposed to danger, the robotic hand could take care of the excavation with AI-guided precision, releasing the flow of water more quickly and safely.

With the capacity to lift up to 3 tons, it is also useful in places inaccessible to conventional heavy equipment. The project has already reached the TRL4 level, proving to work in a controlled environment. Now the goal is to reach 5 by November 2025, testing its efficiency in real disaster conditions.