An Industry 4.0 geolocation system for last mile ground disasters survivor detection: Tests and results

This research introduces a novel geolocation system for survivor detection after disasters, based on the Ground Disaster Information Management System (GDIMS) guidelines. Our geolocation system is a rapid deployment, low-power, and wide-area network solution that uses tiny machine-learning-based voice activity detection to obtain timely information in challenging disaster-struck environments. The system is designed to signal survivor existence on an interactive map and further guide search and rescue efforts during the aftermath of a disaster. The impact and relevance of this technology lie in its ability to continuously monitor disasters despite disruptions from remote, steep, and/or debris disaster surroundings, nontrivial locations of possible victims, and failed transportation and communication infrastructure. This research draws from lessons learned from past events, such as the 2017 and 2010 landslides in Mocoa and Gramalote, Colombia. Results from numerical tests conducted in Gramalote, Colombia, show the technological maturity of the system and its capabilities for possible real-world implementation.