A Digital Twin Approach to Advancing River Emergency Response Systems in Smart Cities

Abstract

Drowning incidents and floods in river cities have become a significant public safety concern worldwide. These incidents result in numerous deaths and injuries, highlighting the urgent need for effective monitoring and rescue systems. Traditional safety detection systems for bodies of water are primarily designed for controlled indoor environments, such as indoor pools, where conditions are stable and predictable; relying on sensors and wearable devices, which are not practical for the varied and challenging conditions of outdoor environments (e.g., distance, wider monitoring areas, and environmental factors such as waves). In response to this challenge, we propose a river emergency response system based on a digital twin model, supported by a human detection model, a water level prediction model, and related algorithms. This piloted system employs a single overhead camera as the primary hardware sensor for continuous real-time safety monitoring. We focus on the Chattahoochee River in the Columbus-Phoenix City area, where drowning incidents have surged in recent years. This system aims to improve rescue response time by generating multi-level of danger alerts based on varying real-time conditions.