An Ant Colony Optimization using training data applied to UAV way point path planning in wind

Abstract

Path planning for small unmanned air vehicles (UAVs) becomes a difficult problem when accounting for wind. Wind can affect the path quality in a nonlinear manner requiring extended segment lengths for accurate following. A method is presented to find near-optimal paths through stochastic optimization based on a training set. In general the method applies to quickly find a near-optimal solution of a continuous function with function parameters. The training set is composed of optimized solutions for different parameters. By a method similar to Ant Colony Optimization, a probability distribution is created based on the training set to create random paths. In this case the similarity of the desired path to examples in the training set is used to weight the probability distribution. The training data can be created offline using computationally intensive methods and the stochastic optimization can be used to create good paths in a timely manner.