Two College Students Are Building a Robot to Replant Burned Forests
Two College Students Are Building a Robot to Replant Burned Forests Marta Bernardino and Sebastião Mendonça invented Trovador, a six-legged, A.I.-powered robot that can plant trees in hard-to-reach, wildfire-damaged terrain Nineteen-year-olds Marta Bernardino and Sebastião Mendonça are developing a robot capable of reaching and reforesting areas where humans have been unable to. Trovador For 19-year-olds Marta Bernardino and Sebastião Mendonça, the forest was the intimate, untamed backdrop of their childhood. “It was a living playground where we built worlds, a sanctuary where the concepts of ‘importance’ were felt instinctively rather than taught,” says Bernardino. As children growing up near Lisbon, the two always believed that the forest would remain a constant in their lives. But with each year, they watched as fires ravaged the forests not far from their homes, leaving behind scorched gray hillsides. Desperate to revive these forests, the two then-high school students set out to create Trovador—a robot capable of reaching and reforesting areas where humans have been unable to. The state of Portugal’s forests A 2024 study by Carlos C. DaCamara, an atmospheric scientist at the University of Lisbon, revealed that between 1980 to 2023, over 1.2 million acres burned in wildfires across mainland Portugal, equivalent to 54 percent of its territory. In 2017, the country recorded 32,000 acres of tree cover loss, with wildfire accounting for 75 percent of that destruction, the highest in a year to date. Moreover, Portugal is the southern European nation most affected by wildfires, based on the scale of burned areas and the sharp rise in recent wildfires. To begin their project, Bernardino and Mendonça set out to understand the current methods used for reforestation and the reasons behind the forests’ slow recovery. “The initial, passive hope that nature would heal itself was shattered when we learned the soil was too damaged and the fires too frequent for recovery,” Bernardino adds. Though volunteers and community members strived to revive the burned forests, it was physically impossible to reach the most vulnerable parts, which happened to be on steep, treacherous slopes. “The defining moment came,” Bernardino says, “when a project leader articulated the brutal truth: the terrain itself was the enemy, making manual replanting a dangerous and often impossible task.” She continues, “The inspiration was no longer a feeling of loss, but a cleareyed recognition of a flawed system. We saw that existing solutions—from volunteer planting to drone seed-dropping—were failing to meet the scale and complexity of the problem.” Quick facts: The impact of climate change on wildfires Between 2003 and 2023, extreme wildfire activity worldwide increased by 2.2-fold. Wildfire seasons are lengthening too, starting earlier in the spring and lasting longer into the fall. Over 60 percent of forests in Portugal lie on steep, rugged terrain, where planting is unsafe and labor is scarce, Bernardino explains. Tractors can’t handle slopes, and they compact the soil. Using heavy vehicles for reforesting can disturb the oxygen and water supply to plants and soil microorganisms. Such disturbances can cause substantial damage to the soil systems, which in certain cases can be long-lasting and even irreversible, harming the productivity of the forest and the overall functionality of the ecosystem. Drone-based aerial seeding is one viable alternative highly considered today for reforestation. However, the technique has its own challenges. While it’s competent in precision identification of suitable locations for reforestation, the method typically uses thousands of seeds per acre (at least 4,000) for blanket seedings, making it less economical. “Drones, while flexible, scatter seeds with low precision—wasting one of the most scarce natural resources,” Bernardino adds. One pilot project focusing on certain conifer species found their survival rate when dropped from drones fell between 0 and 20 percent. “Since the early 2000s, Portugal has lost over half of its forest cover, triggering erosion, water loss and biodiversity collapse,” Bernardino explains. “This crisis hits rural communities hardest: places like Fundão and Alentejo, where forests provide food, water, income and cultural identity. As ecosystems vanish, so do livelihoods.” And the rapid loss of forest cover isn’t limited to Portugal—it extends around the globe. Recent data from the University of Maryland’s Global Land Analysis & Discovery (GLAD) lab, reported in the World Resources Institute’s “Global Forest Review,” found that an unprecedented 16.6 million acres of primary rainforest was lost in the tropics in 2024. Researchers at the GLAD lab estimate that tropical primary forests vanished at an accelerated pace of 18 soccer fields every minute last year. The loss—largely caused by massive forest fires—is almost double that of 2023. “The problem itself became our blueprint,” recounts Bernardino, “and we dedicated ourselves to creating a solution that embraced all the constraints: steep terrain, high survival rates and autonomy.” A firefighter tackles the flames next to a road as vegetation burns during a wildfire in Vila Real, Portugal, this past August. David Oliveira/Anadolu via Getty Images Designing a solution In 2023, Bernardino and Mendonça set out to create Trovador—a six-legged robot able to walk on rugged slopes and plant trees. Their first €15 ($17) prototype, built from recycled parts, planted 28 percent faster than humans with a 90 percent survival rate. The saplings also thrived without any post-planting care. The two are currently working to improve the efficiency of the robot and hope that their current prototype is able to handle longer operations on steeper terrains. “We build all-terrain robots that carry baby trees on their backs and plant them autonomously across difficult terrain,” says Bernardino. The innovators didn’t expect the wave of interest that followed their initial prototype. As a top finalist for National Geographic’s 2024 Slingshot Challenge, they won a grant of $10,000, and the invention was also featured in the magazine as one of the world’s most promising youth-led climate solutions. “On the tech side, the robotics world took notice, too—we became the youngest ever to receive Europe’s top award for Robotics for Sustainability,” says Bernardino. The hexapod robot is capable of climbing slopes of up to 45 degrees while detecting and simultaneously avoiding any boulders in its way. Trovador is also equipped to carry and plant up to 200 saplings per hour. Unlike a tractor, it barely makes an indent on the ground thanks to its light movement, preserving pore space for air and water in the soil. A depth camera attached to it maps any obstacles and allows it to slightly adapt its trajectory in real time. It also uses artificial intelligence and sensors to analyze the pH and humidity of the soil, after which Trovador will follow a three-step dig-place-tamp sequence to plant rooted saplings instead of seeds. “The sequence is validated to hit up to 85 to 90 percent survival in field trials and literature,” says Bernardino. With built-in sensors, Trovador uploads real-time data like GPS coordinates of each plant, soil humidity and battery life to a cloud, allowing the team to monitor the robot remotely. Moreover, during future soil analysis, the robot will be trained to skip the dry ground and steer planting to micro-niches with better odds. Bringing a viable product to market Miguel Jerónimo, a landscape architect and coordinator of Renature projects at the Group for Studies on Spatial Planning and the Environment, an independent environmental organization in Portugal, is optimistic about the tool. “Trovador appears to be an innovative project with potential, particularly as it was developed by two young students who turned a low-cost prototype into a possible approach to one of Portugal’s environmental challenges,” says Jerónimo. “The concept of a six-legged robot designed to move across steep slopes and dense vegetation offers a practical framework for reforestation in areas that are unsafe or difficult for people to access.” While Jerónimo is hopeful about the success of Trovador, he’s equally apprehensive about the robot’s durability in the actual field. “Moving from an experimental prototype to a reliable field-ready tool will require robust testing to ensure it can handle the rough, humid and heavily vegetated conditions typical of Portuguese forests,” he says. “Operational endurance, mobility in dense vegetation and ease of maintenance are areas that need further exploration before the system can be considered ready for broad use.” Additionally, the price tag on the tool also needs to be taken into account. “Keeping production costs low will be essential,” the landscape architect points out. “The robot must be affordable if it is to become a useful and accessible instrument in large-scale reforestation efforts rather than a one-off innovation.” However, Bernardino and Mendonça already have some ideas on how to make it affordable. Instead of selling the Trovador robot itself, the team plans to first market it as a platform that they operate as a service, selling “trees-in-the-ground.” By 2026, they hope to make the robot robust and user-friendly enough to deploy it in large-scale plantations. “Clients [like] municipalities, insurers, forestry firms or NGOs can open our app, outline a polygon, choose native species and receive a quote,” Bernardino elaborates. “Pricing is expected to be a big step up from the current methods, up to six times cheaper than manual crews and four times more cost‑effective than drones once seed wastage is factored in.” The innovators are narrowing in on a minimum viable product. For the next few months, the Trovador team intends to improve the tool based on feedback they received after field testing it in Lisbon this past summer. Both Bernardino and Mendonça’s hopes and ambitions remain high. With the robot, they aspire to make “reforestation that is fast, precise, audit-ready and scalable to the millions of hectares climate models say we must restore this decade,” says Bernardino. Get the latest stories in your inbox every weekday.
Marta Bernardino and Sebastião Mendonça invented Trovador, a six-legged, A.I.-powered robot that can plant trees in hard-to-reach, wildfire-damaged terrain
Two College Students Are Building a Robot to Replant Burned Forests
Marta Bernardino and Sebastião Mendonça invented Trovador, a six-legged, A.I.-powered robot that can plant trees in hard-to-reach, wildfire-damaged terrain
For 19-year-olds Marta Bernardino and Sebastião Mendonça, the forest was the intimate, untamed backdrop of their childhood. “It was a living playground where we built worlds, a sanctuary where the concepts of ‘importance’ were felt instinctively rather than taught,” says Bernardino. As children growing up near Lisbon, the two always believed that the forest would remain a constant in their lives.
But with each year, they watched as fires ravaged the forests not far from their homes, leaving behind scorched gray hillsides. Desperate to revive these forests, the two then-high school students set out to create Trovador—a robot capable of reaching and reforesting areas where humans have been unable to.
The state of Portugal’s forests
A 2024 study by Carlos C. DaCamara, an atmospheric scientist at the University of Lisbon, revealed that between 1980 to 2023, over 1.2 million acres burned in wildfires across mainland Portugal, equivalent to 54 percent of its territory. In 2017, the country recorded 32,000 acres of tree cover loss, with wildfire accounting for 75 percent of that destruction, the highest in a year to date. Moreover, Portugal is the southern European nation most affected by wildfires, based on the scale of burned areas and the sharp rise in recent wildfires.
To begin their project, Bernardino and Mendonça set out to understand the current methods used for reforestation and the reasons behind the forests’ slow recovery.
“The initial, passive hope that nature would heal itself was shattered when we learned the soil was too damaged and the fires too frequent for recovery,” Bernardino adds. Though volunteers and community members strived to revive the burned forests, it was physically impossible to reach the most vulnerable parts, which happened to be on steep, treacherous slopes.
“The defining moment came,” Bernardino says, “when a project leader articulated the brutal truth: the terrain itself was the enemy, making manual replanting a dangerous and often impossible task.”
She continues, “The inspiration was no longer a feeling of loss, but a cleareyed recognition of a flawed system. We saw that existing solutions—from volunteer planting to drone seed-dropping—were failing to meet the scale and complexity of the problem.”
Quick facts: The impact of climate change on wildfires
- Between 2003 and 2023, extreme wildfire activity worldwide increased by 2.2-fold. Wildfire seasons are lengthening too, starting earlier in the spring and lasting longer into the fall.
Over 60 percent of forests in Portugal lie on steep, rugged terrain, where planting is unsafe and labor is scarce, Bernardino explains. Tractors can’t handle slopes, and they compact the soil. Using heavy vehicles for reforesting can disturb the oxygen and water supply to plants and soil microorganisms. Such disturbances can cause substantial damage to the soil systems, which in certain cases can be long-lasting and even irreversible, harming the productivity of the forest and the overall functionality of the ecosystem.
Drone-based aerial seeding is one viable alternative highly considered today for reforestation. However, the technique has its own challenges. While it’s competent in precision identification of suitable locations for reforestation, the method typically uses thousands of seeds per acre (at least 4,000) for blanket seedings, making it less economical. “Drones, while flexible, scatter seeds with low precision—wasting one of the most scarce natural resources,” Bernardino adds. One pilot project focusing on certain conifer species found their survival rate when dropped from drones fell between 0 and 20 percent.
“Since the early 2000s, Portugal has lost over half of its forest cover, triggering erosion, water loss and biodiversity collapse,” Bernardino explains. “This crisis hits rural communities hardest: places like Fundão and Alentejo, where forests provide food, water, income and cultural identity. As ecosystems vanish, so do livelihoods.”
And the rapid loss of forest cover isn’t limited to Portugal—it extends around the globe. Recent data from the University of Maryland’s Global Land Analysis & Discovery (GLAD) lab, reported in the World Resources Institute’s “Global Forest Review,” found that an unprecedented 16.6 million acres of primary rainforest was lost in the tropics in 2024. Researchers at the GLAD lab estimate that tropical primary forests vanished at an accelerated pace of 18 soccer fields every minute last year. The loss—largely caused by massive forest fires—is almost double that of 2023.
“The problem itself became our blueprint,” recounts Bernardino, “and we dedicated ourselves to creating a solution that embraced all the constraints: steep terrain, high survival rates and autonomy.”
Designing a solution
In 2023, Bernardino and Mendonça set out to create Trovador—a six-legged robot able to walk on rugged slopes and plant trees. Their first €15 ($17) prototype, built from recycled parts, planted 28 percent faster than humans with a 90 percent survival rate. The saplings also thrived without any post-planting care. The two are currently working to improve the efficiency of the robot and hope that their current prototype is able to handle longer operations on steeper terrains.
“We build all-terrain robots that carry baby trees on their backs and plant them autonomously across difficult terrain,” says Bernardino.
The innovators didn’t expect the wave of interest that followed their initial prototype. As a top finalist for National Geographic’s 2024 Slingshot Challenge, they won a grant of $10,000, and the invention was also featured in the magazine as one of the world’s most promising youth-led climate solutions. “On the tech side, the robotics world took notice, too—we became the youngest ever to receive Europe’s top award for Robotics for Sustainability,” says Bernardino.
The hexapod robot is capable of climbing slopes of up to 45 degrees while detecting and simultaneously avoiding any boulders in its way. Trovador is also equipped to carry and plant up to 200 saplings per hour. Unlike a tractor, it barely makes an indent on the ground thanks to its light movement, preserving pore space for air and water in the soil. A depth camera attached to it maps any obstacles and allows it to slightly adapt its trajectory in real time. It also uses artificial intelligence and sensors to analyze the pH and humidity of the soil, after which Trovador will follow a three-step dig-place-tamp sequence to plant rooted saplings instead of seeds. “The sequence is validated to hit up to 85 to 90 percent survival in field trials and literature,” says Bernardino.
With built-in sensors, Trovador uploads real-time data like GPS coordinates of each plant, soil humidity and battery life to a cloud, allowing the team to monitor the robot remotely.
Moreover, during future soil analysis, the robot will be trained to skip the dry ground and steer planting to micro-niches with better odds.
Bringing a viable product to market
Miguel Jerónimo, a landscape architect and coordinator of Renature projects at the Group for Studies on Spatial Planning and the Environment, an independent environmental organization in Portugal, is optimistic about the tool. “Trovador appears to be an innovative project with potential, particularly as it was developed by two young students who turned a low-cost prototype into a possible approach to one of Portugal’s environmental challenges,” says Jerónimo. “The concept of a six-legged robot designed to move across steep slopes and dense vegetation offers a practical framework for reforestation in areas that are unsafe or difficult for people to access.”
While Jerónimo is hopeful about the success of Trovador, he’s equally apprehensive about the robot’s durability in the actual field. “Moving from an experimental prototype to a reliable field-ready tool will require robust testing to ensure it can handle the rough, humid and heavily vegetated conditions typical of Portuguese forests,” he says. “Operational endurance, mobility in dense vegetation and ease of maintenance are areas that need further exploration before the system can be considered ready for broad use.”
Additionally, the price tag on the tool also needs to be taken into account. “Keeping production costs low will be essential,” the landscape architect points out. “The robot must be affordable if it is to become a useful and accessible instrument in large-scale reforestation efforts rather than a one-off innovation.”
However, Bernardino and Mendonça already have some ideas on how to make it affordable. Instead of selling the Trovador robot itself, the team plans to first market it as a platform that they operate as a service, selling “trees-in-the-ground.” By 2026, they hope to make the robot robust and user-friendly enough to deploy it in large-scale plantations.
“Clients [like] municipalities, insurers, forestry firms or NGOs can open our app, outline a polygon, choose native species and receive a quote,” Bernardino elaborates. “Pricing is expected to be a big step up from the current methods, up to six times cheaper than manual crews and four times more cost‑effective than drones once seed wastage is factored in.”
The innovators are narrowing in on a minimum viable product. For the next few months, the Trovador team intends to improve the tool based on feedback they received after field testing it in Lisbon this past summer.
Both Bernardino and Mendonça’s hopes and ambitions remain high. With the robot, they aspire to make “reforestation that is fast, precise, audit-ready and scalable to the millions of hectares climate models say we must restore this decade,” says Bernardino.