Drones With Claws Perch on Arctic Icebergs

If you have ever tried to fly a consumer drone in sub-zero temperatures, you know the feeling of watching your battery percentage tick down like a countdown clock. You get maybe fifteen minutes of flight time before the hardware starts begging for a charger. In the Arctic, that limitation is a project killer. If you want to study an iceberg, you can’t just hover over it; you need to live on it.

Researchers at the Université de Sherbrooke in Quebec are tackling this by rethinking how robots interact with unstable terrain. They have developed the “Ice Dart,” a 2.65 kg carbon-fiber drone designed to do something humans can barely do with crampons: land on a 60-degree ice slope and stay there. By using a passive, mechanical system inspired by feline claws, this drone can latch onto glaciers and icebergs, effectively turning itself into a long-term sensor station instead of a short-lived aerial camera.

The Energy Problem in the Wild

In the crypto and AI worlds, we talk a lot about “uptime” and “compute costs.” In robotics, the equivalent is energy density. When a drone is in the air, it is burning energy just to fight gravity. That is the least efficient way to collect data. If you can land, you can shut down the motors, stop the RF noise, and let your sensors run for days or weeks on the same charge that would have lasted minutes in the air.

The Sherbrooke team, led by Alexis Lussier Desbiens and Isaac Tunney, understood that the barrier to remote Arctic monitoring wasn’t just the flight software—it was the landing gear. Ice is a nightmare for robots. It is slippery, hard, and constantly changing. Traditional landing legs bounce or slide. The Ice Dart solves this with a suspension system featuring 38 friction disks that soak up the impact of a hard landing, paired with “microspines” that bite into the ice.

Mechanical Intelligence Over Software Bloat

What I find most interesting here as a builder is the reliance on passive mechanics rather than complex sensors. We have a tendency to want to solve every problem with an AI model or a new piece of software. But the Ice Dart uses a mechanical trigger. When the suspension compresses upon landing, the spines are forced into the ice. There are no actuators or motors needed to “clutch” the surface. It happens naturally.

They even accounted for the physics of a slope. The drone has four legs in an X-pattern. The “downhill” legs, which take the brunt of the weight, deploy a larger spine. The “uphill” legs use a smaller, thinner spine that catches even under light pressure. This allowed the team to achieve a 100% success rate in Iceland’s Fjallsjökull glacier region, even with 30 km/h winds pushing the craft around.

What This Means for the Builders

If you are building in the hardware or AI space, there are three takeaways from the Ice Dart that apply to your work, regardless of whether you care about icebergs:

• Efficiency is the ultimate feature: The ability to “perch” allows a small, cheap drone to do the work of a million-dollar satellite or a fuel-heavy helicopter. In any system, finding the “low-energy state” for your process is how you scale.
• Passive systems beat active ones: By using the force of the landing to drive the spikes, the researchers removed a point of failure. If you can solve a problem with physics instead of code, do it.
• Edge cases are the real world: Most drones are tested in sunny California. Testing in Iceland at 0°C proves that the hardware can survive the thermal contraction and moisture that kills most electronics.

The Skeptic’s View: The “Roll Over” Risk

As impressive as a 60-degree landing is, I’m still skeptical about the long-term autonomy. Icebergs are not static platforms. They melt, they shift, and they occasionally flip over entirely. The researchers are already working on an “emergency takeoff” feature for when the ground literally disappears, but that requires a level of constant environmental sensing that might eat back into those energy savings.

Furthermore, while the mechanical grip is solid, the Arctic is a corrosive, brutal environment. Small metal spines and friction disks are prone to icing over or jamming after repeated cycles. The transition from a controlled study in Iceland to a full-scale deployment in the Canadian Arctic this August will be the real stress test.

The Bottom Line

The Ice Dart represents a shift toward “persistent” robotics. We are moving away from robots that just “visit” an environment and toward robots that can inhabit it. For builders, the lesson is clear: don’t just focus on how your product performs when it’s running at full tilt. Focus on how it survives when the power is low and the conditions are at their worst. Landing on a 58-degree sheet of ice is impressive; staying there until the job is done is the real win.

Read the original at IEEE Robotics →