Dynamic Pathfinding for a Swarm Intelligence Based UAV Control Model Using Particle Swarm Optimisation

In recent years unmanned aerial vehicles (UAVs) have become smaller, cheaper, and more efficient, enabling the use of multiple autonomous drones where previously a single, human-operated drone would been used. This likely includes crisis response search rescue missions. These systems will need method navigating unknown dynamic environments. Typically, this require an incremental heuristic algorithm, however, these algorithms increasingly computationally memory intensive as environment size increases. paper used two different Swarm Intelligence (SI) algorithms: Particle Optimisation Reynolds flocking to propose overall system for controlling groups through proposes Pathfinding (PSOP): dynamic, cooperative algorithm; and, Drone Flock Control (DFC): modular model agents, in 3D environments, such that collisions are minimised. Using Unity game engine, real-time application, simulation environment, data collection apparatus were developed performances DFC-controlled drones—navigating with either PSOP algorithm or D* Lite implementation—were compared. The simulations do not consider UAV dynamics. tasked given target position environments varying quantitative on pathfinding performance, computational usability collected. data, advantages PSO-based demonstrated. was shown be successful creation high quality, accurate paths, usable efficient typical when part SI-based control model. study demonstrated capabilities SI approaches means multi-agent simple environment. Future research may look apply DFC model, advanced which considered factors like atmospheric pressure turbulence, real-world UAVs controlled