When a group of 100-pound robots was seen moving across the lake behind Florida International University’s (FIU) Green Library, it marked a step forward in marine environmental monitoring. While their movements appeared random and sometimes close to colliding, these robots were actually following AI-driven algorithms designed to maximize the area they could scan.
Paulo Padrão, who recently earned his Ph.D. from FIU’s Knight Foundation School of Computing and Information Sciences, led this research. He explained that traditional monitoring methods such as buoys provide information from only one location, but using AI-powered robots allows for mapping entire bodies of water. “Buoys and other traditional water monitors only tell you what’s happening in one spot,” Padrão explained. “But when we put AI-powered robots in a lake, they can map the entire area.”
Padrão noted that advances in artificial intelligence are enabling robotic systems to become more autonomous. “Instead of pre-programming every movement, our robots can now learn and adapt, making decisions based on the data they’re collecting. We literally just press start and they navigate to different locations based on what they discover,” he said.
A key part of the research involved digital twin technology—virtual simulations that mirror real-world systems. According to Padrão: “A digital twin is like a video game version of a real-world system. It allows us to run experiments of a system on a computer before testing it in real life.” This approach reduces logistical challenges by allowing researchers to conduct tests virtually before deploying physical robots.
The team built a large water tank facility at FIU’s Biscayne Bay Campus with support from the Department of Defense. Cameras track robot positions within the tank, providing accurate data for creating virtual models used for further experimentation and training environments.
AI also improves how marine robots operate by allowing them to interpret their surroundings while navigating. Instead of following fixed routes or relying on constant communication with central computers, each robot independently decides where to go next based on environmental data it collects—such as temperature or dissolved oxygen levels—even if communication is interrupted.
“Traditional systems relied on constant communication with a central computer to stay coordinated, but our AI-powered approach eliminates this dependency,” Padrão said. Under guidance from his advisor Leonardo Bobadilla, the team developed methods enabling robots to communicate locally with nearby units while efficiently covering large areas together—a strategy referred to as forming a “smart swarm.”
Reflecting on his time at FIU, Padrão highlighted collaborations across departments and partnerships with organizations like the Florida Department of Environmental Protection. Their work demonstrated new ways for autonomous tracking of water quality hotspots and received three award nominations at IEEE ICRA 2025.
“It was also energizing to be part of a larger, multidisciplinary team,” he added.
Padrão will continue his career as an assistant professor at Providence College’s Department of Mathematics and Computer Science starting this fall. He plans to pursue further research in robotics and machine learning alongside STEM education initiatives.
