Path planning for unmanned aerial vehicle based on genetic algorithm & artificial neural network in 3D

Abstract

The planning of path for Unmanned Aerial Vehicle (UAV) is always considered to be a vital task. Path planning for UAV for avoiding the obstacle in its path can be accomplished by finding the solution for an optimization problem. Genetic Algorithm which is a global optimization tool can be of great use to solve the optimization problem for path planning of UAV. Artificial Neural Network (ANN) works well for function fitting quickly and can be used to approximate almost any function. The Genetic Algorithms are good at converging to the globally optimum solution generation by generation. Each generation is expected to be better than its previous generation. Neural Networks work faster than Genetic Algorithms for finding the solution to a given problem but may get converged to local optimum instead of global optimum. In this paper a new method for path planning for UAV to avoid obstacle coming in its path based on the combination of Genetic Algorithms and Artificial Neural Networks has been proposed in which the output generated from the Genetic Algorithms is used to train the network of Artificial Neural Networks. The model for path planning is based on 3D digital map.