“`html
Compact self-sufficient underwater vessels, akin to the drones of the sea, could prove invaluable for exploring the ocean’s depths and observing its evolving conditions. However, such aquatic mini-machines can be swiftly overwhelmed by tumultuous oceanic currents.
Caltech researchers spearheaded by John Dabiri (PhD ’05), the Centennial Professor of Aeronautics and Mechanical Engineering, have been exploiting the inherent capability of jellyfish to navigate and explore the ocean, equipping them with technology and prosthetic “hats” through which the creatures can transport minor payloads on their aquatic journeys and relay their discoveries back to the surface. These bio-engineered jellyfish must deal with the rise and fall of the currents they encounter, but since they lack brains, they do not strategize on how best to reach a destination, and once deployed, they cannot be operated remotely.
“We recognize that enhanced jellyfish can serve as excellent ocean explorers, yet they lack a brain,” Dabiri states. “Therefore, one of the aspects we’ve been focusing on is defining what that brain would entail if we were to equip these systems with decision-making capabilities underwater.”
Now Dabiri and his former graduate student Peter Gunnarson (PhD ’24), currently affiliated with Brown University, have discovered a method to streamline that decision-making process, enabling a robot, or potentially an augmented jellyfish, to utilize the turbulent vortices generated by ocean currents instead of opposing them. The researchers have recently shared their results in the journal PNAS Nexus.
In this endeavor, Gunnarson revisited an old companion in the lab: CARL-Bot (Caltech Autonomous Reinforcement Learning roBot). Years ago, Gunnarson constructed the CARL-Bot as part of his research aimed at integrating artificial intelligence into the navigation approaches of such bots. However, Gunnarson has recently devised a more straightforward method than AI to enable such a system to make decisions underwater.
“We were brainstorming ideas about how underwater vehicles could harness turbulent water currents for propulsion and pondered whether, instead of being a hindrance, they could serve as an advantage for these smaller vehicles,” Gunnarson remarks.
Gunnarson aimed to comprehend precisely how a current maneuvers a robot. He affixed a thruster to the side of a 16-foot tank in Dabiri’s lab located at the Guggenheim Aeronautical Laboratory on Caltech’s campus to consistently generate what are termed vortex rings—essentially the underwater counterparts to smoke rings. Vortex rings effectively represent the types of disturbances an underwater explorer would face in the ocean’s chaotic fluid dynamics.
Gunnarson started utilizing the CARL-Bot’s singular onboard accelerometer to gauge its movements and how it was influenced by vortex rings. He observed that occasionally, the robot would become ensnared in a vortex ring and be propelled across the tank. He and his colleagues began to speculate if this effect could be manipulated purposefully.
To investigate this, the team devised straightforward commands to assist CARL in identifying a vortex ring’s relative position and then adjusting its positioning to, in Gunnarson’s words, “hop on and enjoy a ride effectively for free across the tank.” Alternatively, the bot can opt to avoid a vortex ring it does not wish to be affected by.
Dabiri notes that this procedure incorporates elements of biomimicry, drawing inspiration from natural phenomena. Gliding birds, for instance, often leverage strong winds to conserve energy instead of battling against them. Research has also indicated that fish may permit themselves to be carried by the ocean’s swirling currents to conserve energy. However, in both natural instances, the systems utilize relatively advanced sensory input and a brain to achieve this.
“What Peter has uncovered is that essentially with just one sensor, this single accelerometer, and fairly uncomplicated control laws, we can attain similar advantages in terms of harnessing the energy in the environment to transition from point A to point B,” Dabiri explains.
Looking ahead, Dabiri aspires to integrate this research with his hybrid jellyfish. “By employing the jellyfish, we can utilize an onboard accelerometer to measure how this system is being affected,” he says. “Ideally, we can showcase a comparable ability to capitalize on environmental flows to navigate more efficiently through the water.”
The PNAS Nexus article is titled “Surfing vortex rings for energy-efficient propulsion.” The research received backing from the National Science Foundation.
“`