AI underwater robots track scuba diver stress by watching exhaled bubbles

Bubbles as a distress signal

Researchers have developed a novel way to keep scuba divers safe underwater: artificial intelligence-powered robots that watch exhaled bubbles. The system tracks changes in bubble patterns to detect when a diver is stressed or in trouble.

Divers breathe compressed air, which forms bubbles when exhaled. Under normal conditions, these bubbles follow a predictable rhythm. Stress, panic or the onset of decompression sickness alters breathing rates and patterns, producing a distinct bubble signature the AI can flag.

How the robotic system works

Small autonomous underwater vehicles fitted with cameras and acoustic sensors patrol dive sites. They stream video of exhaled bubbles to an onboard AI model trained to recognise stress-related irregularities. The robot can then alert surface teams or swim closer to the diver for a visual check.

Unlike traditional diver-monitoring methods that rely on handheld sonar or periodic check-ins, this system provides continuous, passive observation. It does not require divers to wear extra sensors or devices.

Real-time safety intervention

The primary goal is early detection of panic or decompression sickness, which can cause disorientation, joint pain and paralysis if untreated. Once the AI identifies a stressed diver, the robot can guide them to ascend slowly or signal for a rescue.

In tests within controlled pool environments, the system correctly identified stress-induced breathing changes in over 90 percent of cases. Researchers are now planning open-water trials to evaluate performance under currents, varying visibility and marine life interference.

Broader implications for underwater work

Beyond recreational diving, the technology could transform safety protocols for commercial divers, underwater welders and military personnel. These workers operate in high-risk conditions where communication is limited and decompression illness is a constant threat.

The robot's autonomy also reduces the need for human safety divers, lowering operational costs and risks. Researchers emphasise that the system is not meant to replace human supervision but to augment it with a tireless, vigilant digital eye.

Next steps include integrating the bubble-analysis AI into existing underwater drone platforms already used for pipeline inspection and marine research. Commercial deployment could begin within three to five years, pending regulatory approvals and further validation.