New Delhi: Marta Bernardino and Sebastião Mendonça were still in school when the hills near their homes outside Lisbon in Portugal kept going up in flames year after year. Those memories stayed with them. By the time they turned 19, the duo had decided to build something that could replant the charred slopes no human or machine could safely reach.
They developed Trovador, a compact and six-legged reforestation robot that walks like a spider and carries saplings across burnt and unstable terrain. The two students see it as a tool for the many parts of Portugal where fires have carved deep scars into mountainsides.
Portugal has lived with this problem for decades. A 2024 study led by University of Lisbon atmospheric scientist Carlos C. DaCamara documented more than 1.2 million acres (an area equal to 54 percent of the country’s territory) burned between 1980 and 2023. The year 2017 still stands out for its scale of loss, when 32,000 acres of tree cover vanished and wildfires caused three-quarters of that destruction. Much of this happens on steep slopes that volunteers and fire crews simply cannot access.
Bernardino and Mendonça looked at those numbers and saw the same barrier again and again: the terrain. As they explain in their crowdfunding video, “Steep terrain prevents manual planting and heavy machinery from reaching most of Portugal’s burned areas.”
More than 60 percent of the country’s forests sit on rugged inclines, and traditional replanting struggles to keep pace with the constant cycle of fire and soil degradation.
The Robot At A Glance
The duo put together their first prototype in 2023 using recycled parts and a budget of €15. Even this early version managed to plant saplings 28 percent faster than a human crew, and the trees showed a 90 percent survival rate without needing post-planting care.
Smithsonian Magazine highlighted how that result convinced the teenagers to take the concept further, building a sturdier version capable of operating on slopes of up to 45 degrees.
Bernardino said, “We build all-terrain robots that carry baby trees on their backs and plant them autonomously across difficult terrain.”
Trovador’s six-leg mechanical frame spreads its weight evenly, avoiding the soil compaction caused by tractors or other heavy machines. That soft footprint matters because compacted soil loses oxygen and slows water movement – the two things young roots desperately need.
A depth-sensing camera guides the robot as it walks, helping it step around rocks, hollows and debris. Before planting, its onboard AI checks soil pH and moisture. Once conditions look suitable, the robot performs its three-step digging, placing and tamping cycle. Field trials and existing research suggest the method consistently delivers 85-90 percent survival.
Drones have been promoted as another route for reforestation, although they spread thousands of seeds at once with little accuracy. Bernardino explained the limitation, “Drones… scatter seeds with low precision, wasting one of the most scarce natural resources.”
Some pilot studies show survival rates of 0-20 percent. Trovador avoids that scatter by planting rooted saplings one at a time, selecting micro-spots where moisture or shade offer better odds.
The robot can place up to 200 saplings per hour and then upload GPS points, soil readings and battery levels to the cloud.
The two are now working on software that will help Trovador steer away from dry patches and toward more promising pockets of soil without human cues.
Global Attention
Their idea has travelled fast. Bernardino and Mendonça reached the finals of National Geographic’s 2024 Slingshot Challenge, earning a $10,000 grant. They later received a major European sustainability award for robotics. Environmental specialists watching the project see strong potential but also state that the robot’s long-term performance still needs to be tested across more sites.
The experts said the design “offers a practical framework for reforestation in areas that are unsafe or difficult for people to access”, while also pointing out that durability, navigation through dense vegetation and sustained reliability must be demonstrated before the robot is adopted widely.
Cost remains a point of discussion. The two creators plan to keep the robot affordable by offering Trovador as a service rather than selling the machines. Their model allows municipalities, insurers, forestry teams and NGOs to outline a burnt zone in an app, choose native species and receive a quote for delivered saplings. They describe it as “trees-in-the-ground”.
According to their estimates, this will come in cheaper than manual crews and ultimately more economical than drone operations once seed wastage is factored in.
After this summer’s field trials around Lisbon, the team is now refining the minimum viable version for commercial use. Their ambition is to see Trovador working in large and damaged landscapes by 2026, a timeline defined by growing urgency.
Bernardino’s target is reforestation that is “fast, precise, easy to check and ready to scale across the millions of hectares climate models say we must restore this decade”.
About the Author
