The buddy system, where a pair of divers look out for one another, is used by the diving community to mitigate danger. They inspect each other’s breathing apparatus, monitor remaining air supplies, health status, and can provide emergency support during a dive. Due to buddy unavailability however, some divers dive solo, forgoing the safety aspects of the buddy system. We propose a dedicated dive-buddy robot as a solution to this problem. The robot, an autonomous underwater vehicle, could operate as an assistant, controlled by the diver using hand gesture-based communication; a communication method commonly used amongst divers. To capture the gestures, we have developed a smart dive glove integrated with 5 dielectric elastomer strain sensors. The capacitance of each sensor was measured with on-board electronics, translated into a command using machine learning and transmitted underwater using acoustics. Due to travel restrictions relating to the Covid-19 pandemic, a demonstration with the diver and vehicle in the same pool was not possible. Therefore, here we present a demonstration with the diver performing gestures in a pool in Auckland, New Zealand, sending commands to the robot in a pool in Zagreb, Croatia. The commands were sent through acoustics to a computer in Auckland, over cellular internet to a computer in Zagreb, which then relayed instructions to the robot using acoustics. The robot was sent four commands and successfully completed all manoeuvres. The performance of the communication with regards to time delays is assessed and future improvements are discussed.
|