Competitive Self-Organizing Neural Network Based UAV Path Planning

Abstract

Path planning is the fundamental aspect of applications for autonomous Unmanned Aerial Vehicles (UAVs) system. It allows UAV to find an optimal path relevant to some specific missions within limited time, especially in large sized scenarios. In this paper, a novel competitive self-organizing neural network algorithm is proposed to improve the search ability and speed up the convergence of traditional algorithms. More specifically, in the initialization phase, a new opposition-based learning is adopted to generate better neurons. Next, a secondary competitive layer is added above the hidden layer, thus enhancing accuracy of the algorithm. Simulations validate the proposed algorithm outperforms some intelligence algorithms in terms of optimization ability.