The 'marsupial' robots building a 3D replica of Europe's forests

Deep in a Swiss pine forest, the GPS is completely dead. A ground-based robotic walker navigates the undergrowth entirely blind to satellites, relying on its own sensors to read the terrain before launching a drone directly from its back into the canopy.

This mechanical "marsupial" pairing is the hardware driving DigiForest, a multinational push to build exact 3D replicas of European woodlands. The European Union needs to know exactly how much carbon these forests are sequestering to hit its climate targets, and rough estimates based on acreage no longer cut it. By feeding high-fidelity maps into AI models, the project aims to replace manual guesswork with hard, automated data.

Retiring the tape measure

Calculating the biomass of a forest has traditionally required human workers to literally pull tape measures around trunks. The walker-and-drone teams are designed to remove that bottleneck entirely.

Operating across managed test sites in Finland, Switzerland, and the United Kingdom, the robots autonomously extract specific tree traits. They log trunk diameter, calculate height, and identify species without human intervention. By building a granular digital twin of the landscape, the system tracks soil health and biodiversity markers across vast areas.

Crucially, these models function as an early warning network. If a specific patch of woodland is struggling, the digital replica flags the decline long before the damage becomes visible to a forester standing a mile away.

Surgical timber extraction

This mapping data is designed to push the timber industry towards a "continuous cover" model. Instead of clear-cutting large swaths of land, foresters use the digital twins to pinpoint exactly where to thin the woods.

It is a surgical approach to logging. By extracting specific trees for long-lived wood products, the wider forest remains intact and functional as an active carbon sink. It also preserves the older, complex habitats where woodland biodiversity actually lives.

Vines and price tags

A machine that walks flawlessly among neatly managed European pines faces a stark reality check elsewhere. The algorithms currently driving the DigiForest robots still struggle with extreme environments.

Engineers are trying to adapt the navigation protocols for the chaotic, multi-layered foliage of tropical rainforests and the deep snow of boreal zones. A robot that maps a Swiss valley perfectly is effectively useless if it immediately gets snagged on a jungle vine or stalls in a drift.

Then there is the financial hurdle. The current models are bespoke, highly engineered research prototypes. If this technology is going to make any real dent in global conservation, the hardware must be simplified for mass production, bringing the price down for smaller timber holdings and developing nations.

Sources

• DigiForest Project