Drone flocking optimization using NSGA-II and principal component analysis

Individual agents in natural systems like flocks of birds or schools fish display a remarkable ability to coordinate and communicate local groups execute variety tasks efficiently. Emulating such into drone swarms solve problems defense, agriculture, industrial automation, humanitarian relief is an emerging technology. However, flocking aerial robots while maintaining multiple objectives, collision avoidance, high speed etc., still challenge. This paper proposes optimized drones confined environment with conflicting objectives. The considered objectives are avoidance (with each other the wall), speed, correlation, communication (connected disconnected agents). Principal Component Analysis (PCA) applied for dimensionality reduction understanding collective dynamics swarm. control model characterized by 12 parameters which then using multi-objective solver (NSGA-II). obtained results reported compared that CMA-ES algorithm. study particularly useful as proposed optimizer outputs Pareto Front representing different types can be scenarios real world.