Simulated Evaluation of Navigation System for Multi-quadrotor Coordination in Search and Rescue

Many methods have been developed which assist in the localization of victims trapped in disaster-prone areas. Disaster management after the immediate onset of such sudden occurrences intimates readiness in technology, availability, accessibility, perception, training, evaluation, and deploy ability. This can be attained through evaluation and comparison of different techniques supplanting each other, essentially covering each aspect of the Search and Rescue operation. Developments by academia and industry have led to deep learning advancements like the use of Convolutional Neural Networks resulting in an increasing dependence of first responders on UAV technology fitted with state-of-the-art machines working with real-time information from various sensors. We have, in this paper, proposed a technique to implement a simulation involving detection of life in the immediate occurrence of disasters with the assistance of a deep learning model simultaneously deploying multi-quadrotor coordination among the vehicles with the use of an appropriate region-partitioning method to speed up the process even further. Moreover, other non-conventional techniques have also been discussed.