Quadrotors distribution controlling for swarm centroid detection using a modified artificial neural network self-organizing map (Modified ANN-SOM)

Abstract

The swarm applications on quadrotor is an interesting topic to be explored because it leads to the nanotechnology. Its application requires the special method for distribution, coordination, formation, and communication among quadrotors. This research is about distribution control of quadrotor swarm in the centroid point. The centroid point position is random. So, controlling for quadrotor that can trace the centroid position with optimal distribution, is needed. In this research, avoiding collisions among the quadrotors is required. In this paper, Modified ANN-self organizing map method is proposed. Modifications are made with the development of neurons which is a part of the map optimization feature by using Particle Swarm Optimization (PSO) method. ANN-SOM is used to identify the initial movement of quadrotor members. When it has been identified, the movement of quadrotors swarm to the centroid point will be done using the PSO method. To avoid the collision at the centroid point, the minimum value of the distance of each quadrotor member is determined by the PSO method. Based on the simulation result, the proposed method is able to control the quadrotor swarm in centroid point tracing optimally with minimum variance and able to avoid collisions among quadrotor swarm members.