ABSTRACT
There are an estimated 500,000 annual drowning deaths worldwide but there are no broad prevention initiatives besides developing stronger water competencies and traditional lifeguard search and rescue. It can take just three minutes to reach the last stage of drowning which is irreversible and accounts for 7% of all injury-related deaths worldwide. To prevent drowning a rescuer must be able to identify and reach a victim on time. Where swimming takes place outdoors in natural spaces it can be challenging for a lifeguard to identify and reach a drowning victim in time. This paper presents a working prototype which uses gesture recognition technology for search and rescue to prevent drowning. A comprehensive literature review and patent search was completed before design and development of the prototype solution. During prototype trials a lifeguard who simulated drowning behaviour was detected using gesture recognition technology within a geofenced area and a life jacket delivered in under three minutes. The trials were conducted in calm weather with temperature in the range of 25 to 40°C. Future research should include testing in conditions such as heavy rain and strong winds. It should also include testing signal interference and geofencing constraints in wild swimming areas.
