AI Robots Can Monitor Diver Stress Using Exhaled Bubbles
University of Minnesota researchers have developed an AI system enabling underwater robots to non-contactedly assess diver stress by analyzing exhaled bubbles, overcoming traditional sensor limitations.
Researchers at the University of Minnesota Twin Cities have pioneered a novel AI system that empowers underwater companion robots to monitor a diver's physiological state in real time. This groundbreaking technology functions by observing the bubbles a diver exhales, marking the first application of robotic vision to estimate a diver's human respiration rate (HRR).
Scuba diving, particularly in challenging environments, inherently exposes individuals to significant physical stress, ranging from fatigue to potentially life-threatening respiratory issues. By tracking the rate and size of bubbles released from a diver's regulator, robots equipped with cameras can now identify early indicators of stress, hyperventilation, or exhaustion.
This non-invasive method addresses a persistent hurdle: conventional underwater medical sensors and wearable devices often falter because thick diving suits impede the necessary physical contact for accurate readings. Furthermore, wireless data transmission through water is notoriously difficult. The objective, according to senior author Junaed Sattar, was to provide divers with a dedicated robotic safety partner capable of detecting physiological stress.
To train the AI, the team employed a unique "fuzzy labeling" system. Given the often murky conditions of underwater footage, researchers manually classified numerous images while using synchronized audio cues—the distinct sounds of regulator exhalations—to precisely teach the robot how to recognize a breath. The system, named HREyes, can categorize breathing as "below-normal" (defined as 20 breaths per minute) and notify the diver, converting visual data into actionable breath counts to gauge duress.
The research team compiled an extensive dataset of audio and visual recordings from diverse locations, including Lake Superior, Square Lake, and the Caribbean Sea, to ensure the autonomous underwater vehicle (AUV) could operate effectively in varying water temperatures and clarity levels. Future plans involve integrating breathing rate data with an analysis of diver movement to create a comprehensive "wellness profile" for enhanced safety during underwater explorations.
This development represents a significant leap in underwater safety and human-robot collaboration. By leveraging AI and computer vision to non-invasively monitor physiological cues like respiration, it overcomes critical limitations of traditional sensors in aquatic environments. This technology could enhance diver safety in exploration, scientific research, and potentially even underwater construction or resource extraction, analogous to the push for in-situ monitoring in extreme terrestrial or space-based operations.
Edited by the news editor with AI from the original report — please refer to the original source.