Researchers at the University of Minnesota Twin Cities have developed an artificial intelligence system that enables autonomous underwater vehicles to estimate a diver’s respiration rate by analyzing exhaled bubbles. The system uses cameras to track the frequency and volume of bubbles released from a diver’s regulator. The resulting data can indicate changes in breathing associated with stress, hyperventilation or exhaustion.
The research, published in The International Journal of Robotics Research, is described by the team as the first use of robotic vision to estimate a diver’s human respiration rate underwater. Monitoring breathing underwater is difficult because wetsuits and drysuits can prevent conventional medical sensors from maintaining contact with the body. Wireless data transmission through water is also limited, making it harder to use connected monitoring devices.
“Our goal was to give divers a dedicated robotic safety partner to provide a second set of ‘eyes’ capable of reading physiological stress underwater,” said Junaed Sattar, an associate professor in the university’s Department of Computer Science and Engineering and senior author of the paper. “This work is a first step towards assessing not just one but a group of divers in the robot’s field of view.”
The researchers trained the AI model using thousands of manually categorized underwater images. Synchronized audio recordings of regulator exhalations helped the team identify when breaths occurred in footage affected by low visibility.
“While monitoring breathing is a standard vital sign on land, doing so underwater presents immense technical challenges,” said Demetrious Kutzke, a doctoral student in the university’s Robotics & Vision Laboratory and lead author of the study.
The team collected audio and visual data in Lake Superior near Duluth, Minnesota, Square Lake near Stillwater, Minnesota, and the Caribbean Sea off Barbados. The range of locations allowed the researchers to test the system in different water temperatures and visibility conditions.
During field trials, a communication system called HREyes classified a diver’s breathing rate as below normal, normal or above normal. The categories were defined as fewer than 14 breaths per minute, 14 to 20 breaths per minute, and more than 20 breaths per minute, respectively. The researchers plan to combine breathing-rate measurements with analysis of diver movement to produce a broader assessment of a diver’s physical condition.
The research team also included Vennela Dupati, an undergraduate student in the university’s computer science, electrical engineering and computer engineering departments. The work received support from the U.S. Department of Defense’s Science, Mathematics, and Research for Transformation Scholarship and the National Science Foundation.
Photo credit: Junaed Sattar
