Rescue-bots: a proposed multi-robot architecture for rescue missions

Abstract

With the increasing number of natural disasters and wars that led to the death of innocents all around the world, an innovative and efficient rescue system is a must for helping the survivors after those tragic events specially when traditional rescue systems in such circumstances become significantly inefficient. As a team of five students, we cooperated to design a fully autonomous rescue system that has a greater efficiency in detecting survivors and aiding them without putting the lives of rescue teams at risk. Our rescue system consists of a drone, a control center and a ground vehicle. Firstly, the drone will take off searching for any sign of a possible survivor. When a survivor is detected, the drone will send the location to the control center which in turn does the decision making and sends the required commands to the ground vehicle. The ground vehicle will then go to the survivor’s location providing the survivor with the necessary aid. A fully detailed high-level and low-level design have been done discussing our approach, algorithm and design components of the rescue system. Then, we showed the design implementation of the vehicle, drone, pre-trained model and the drone-vehicle communication based on the high-level and low-level design provided. Several tests have been done to the design regarding the drone’s searching paths, the drone-vehicle communication and the detection algorithm. After doing the tests, we evaluated the performance of the system, drew conclusions and improved any areas that required refinement. The rescue system successfully managed to detect a human figure by the drone with a very high accuracy and managed to guide a vehicle to its location. Received: 03 June 2025 Accepted: 18 November 2025 Published: 11 December 2025