SAR Box: Developing AI to enable unmanned search and rescue

点击翻译按钮获取中文摘要

Saving lives at sea is time critical, North Sea temperatures average 10.5°C where unprotected casualties have less than 30 minutes before swim failure and death.More people than ever are working offshore due to the continued development of offshore wind (OW), oil and gas (O&G) decommissioning and commercial shipping. OW projects are moving further from shore where wind speeds and wave heights increase. Average distance to shore of farms under construction in 2019 was 59km, up from 35km in 2018\. This will increase over the next decade, forcing operators to rethink onsite SAR provisionOSS are developing the world's first remotely operated Search and Rescue (SAR) service to keep offshore workers safe, keep rescuers out of harm's way and support the ever-growing blue economy.The service requires a network of semi-autonomous Unmanned Rescue Vessels (URV's), ready to rapidly deploy in case of an emergency. Reducing emergency response times to save lives at sea whilst reducing the requirement put people at unnecessary risk.The system will use SARbox (sensors and machine learning / AI developed in this project) to identify and actively track casualties drifting in the water enabling the URV to navigate to them and position for recovery. Thereby taking the search out of search and rescue.The overall objective is to enable unmanned SAR. This requires the development of a sensor array combined with AI/Machine learning to identify and track casualties in the water, supporting the navigation of the URV to perform casualty recovery. This will form the SARbox system and this project will take SARbox from TRL2-5\.