Publication

Abstract : Swarm robots have the potential to be utilized as a part of various applications due to the obvious advantages they offer, namely their resilience, their adaptability to different environments, and their reduced reliance on humans, particularly for hazardous or laborious tasks such as search and rescue operations (SaR). Several particle swarm optimization (PSO) algorithms have been proposed and developed for controlling a robot swarm to achieve the desired behavior in SaR operations. In this paper, we present an implementation of the previously proposed Robotic Darwinian Particle Swarm Optimization (RDPSO); an exploration algorithm which overcomes the limitation of convergence on multiple targets. A simulation of a SaR mission is presented using Robot Operating System (ROS) and Gazebo, with a visualization in Rviz using intensity maps that mimic real-world scenarios, such as victim identification by voice/sound intensity mapping, localization of fire source by temperature intensity mapping, and identification of sources of radioactive leaks by radiation intensity mapping and experiments were conducted using RPSO and RDPSO algorithms for two scenario missions, that is, scenario with two victims and another with four victims. Experimental result shows that RDPSO has better performance compared to RPSO for multiple target SaR operations.