“It was the terrain itself that was the enemy.” That candid assessment from one Portuguese reforestation leader sparked one of the more interesting climate-tech inventions of recent years, a six-legged, AI-powered robot named Trovador-not built by a major robotics firm, but by two 19-year-old engineering students dead set on restoring the forests of their homeland.
Portugal’s wildfire crisis is out of control: between 1980 and 2023, more than 1.2 million acres of forest-54% of the country’s territory-were burned. And many of those fires tore through mountainous and cliffside terrain, where more than 60% of the nation’s green areas lie. Such landscapes are unreachable by tractors and dangerous for human crews, leaving broad swaths of burned land to erode and deteriorate. Traditional replanting methods-volunteer crews, manual planting, or even drone-based aerial seeding-have all proved inadequate. Drone seeding is fast but often casts more than 4,000+ seeds per acre, with survival rates as low as 0-20%, wasting precious native seed stock.
Trovador approaches the problem from the ground up-literally. Its hexapod design keeps the weight evenly distributed, avoiding the soil compaction that heavy machinery causes, disrupting the oxygen and water flow to both roots and microorganisms. Each leg has a specialized function: the front legs have sharp tips that dig into steep ground; the middle legs are fitted with grippers for saplings, while the rear legs tamp the soil. The robot executes a very precise sequence of digging, placing, and tamping that has been validated to yield 85 to 90% survival rates in field trials. Powered by AI, the robot deploys its depth camera to map obstacles, adjusting its path in real time, scaling 45-degree slopes while dodging boulders and dense vegetation.
What sets Trovador apart is its blend of autonomous mobility with the latest in soil intelligence. Embedded sensors at the tip measure pH, humidity, and other parameters before planting, much as the principles behind multisensory soil monitoring systems do in precision agriculture. That way, each sapling is planted into a micro-niche that offers the best survival potential. The data collected, along with GPS coordinates for each plant, their soil moisture, and battery status, are then sent to the cloud for remote monitoring. Future software updates also mean the robot will be able to sidestep those drier patches entirely, focusing its efforts in areas that look most promising.
This is the current, full-scale prototype, shaped like a dog for stability, which can carry and plant up to 200 saplings an hour, leaving little more than a faint footprint. The lightweight, low-impact gait preserves soil pore space vital for water infiltration and root aeration. The planting mechanism includes drills to create holes in compacted or rocky soils, self-adjusting in real time to variable conditions on the ground. The AI of the system can also randomize the planting patterns of native species to support biodiversity and ecosystem resilience.
The origins of this project are as humble as they are inspiring. Marta Bernardino and Sebastião Mendonça built their first prototype in 2023 with recycled parts for just €15. That proof-of-concept was planting saplings 28% quicker than any human, without any post-planting care. Recognition was quick in coming: finalists in National Geographic’s 2024 Slingshot Challenge with a $10,000 grant, and the youngest-ever recipients of Europe’s top Robotics for Sustainability award.
But to scale the technology, more than just engineering prowess is needed-a viable deployment model is required, too. Instead of selling the robot, the team will offer “trees-in-the-ground” as a service. Potential clients-municipalities, insurers, forestry firms, NGOs-can use an app to draw a polygon on a map, select native species and get a quote. Bernardino estimates the service could be up to six times cheaper than manual crews and four times more cost-efficient than drones when seed wastage is factored in.
From an engineering perspective, Trovador represents the confluence of autonomous robotics, AI-enabled environmental sensing, and terrain-adaptive locomotion. It taps into the precision of IoT-enabled in situ soil analysis platforms-where parameters such as NPK nutrients, conductivity, and moisture are measured-into each planting decision. Its obstacle detection and path-planning algorithms run parallel to those utilized in advanced agricultural drones but optimized for ground-based, high-precision interventions in post-disaster landscapes. If it succeeds in its large-scale deployment goal of 2026, Trovador will fundamentally redefine reforestation logistics in rugged wildfire zones around the world. By fusing mechanical endurance and ecological intelligence, it may provide a model for climate-adaptive restoration, one sapling, one slope, and one data point at a time.
