In a significant advancement for underwater robotics, researchers at the California Institute of Technology have taught a submarine robot to use turbulence as a propulsion mechanism rather than fighting against it.
The research team, led by Professor John Dabiri and former graduate student Peter Gunnarson (now at Brown University), developed a system that allows their CARL-Bot to detect and exploit vortex rings—underwater equivalents of smoke rings—to travel efficiently through water. Their findings were published in the journal PNAS Nexus on May 12, 2025.
"We were brainstorming ways that underwater vehicles could use turbulent water currents for propulsion and wondered if, instead of them being a problem, they could be an advantage for these smaller vehicles," explains Gunnarson, who built the CARL-Bot (Caltech Autonomous Reinforcement Learning roBot) during his time at Caltech.
The robot uses a single onboard accelerometer to sense when it encounters a vortex ring, then executes precise maneuvers to position itself within the vortex's material boundary. Once entrained in the vortex, the robot is propelled across distances without expending additional energy. In laboratory tests using a 16-foot tank, this technique achieved a nearly five-fold reduction in energy consumption compared to conventional propulsion methods.
While CARL-Bot was originally designed with artificial intelligence capabilities for navigation, the researchers discovered a simpler approach for decision-making underwater. The team developed basic commands that help the robot detect a vortex ring's location and position itself to "hop on and catch a ride basically for free," as Gunnarson describes it.
This innovation has significant implications for ocean exploration, where small autonomous underwater vehicles are often limited by battery life and can be easily overpowered by ocean currents. The technology could enable longer-duration missions for environmental monitoring, oceanography research, and underwater infrastructure inspection. Professor Dabiri also hopes to apply these principles to his work with bionic jellyfish, potentially creating hybrid systems that combine biological organisms with electronic controls for efficient ocean exploration.