UAV Path Planning for Ice Intelligence Purposes using NLP

Abstract

As the oil and shipping industry are interested in operating in arctic waters, the need for ice intelligence gathering is rising.The thesis describes the implentation and results of a path planning framework for an UAV used for ice intelligence purposes. THe framework produsces paths based on optimization of a non-linear problem, using the IPOPT library in C++.A model for information uncertainty is implemented, and optimal paths based on minimizing the total information uncertainty are compared to optimal paths based on minimizing distance between UAV and target. Both off line and offline path planning is tested with single and multiple targets.It was found that minimizing information uncertainty can work very well for path planning for ice berg surveillance, or for surveillance of a small search grid.Minimizing information uncertainty generally gave better results than minimizing distance between the UAV and given targets.The implementation should be made more robust, and interfaces towards other UAV sysmets has to be made before the path planning platform has any practical use.