Optimal offline path planning of a fixed wing unmanned aerial vehicle (UAV) using an evolutionary algorithm

Path planning is the process of generating a path between an initial location and a target location that has optimal performance against specific criteria. This paper addresses the problem of offline path planning as applied to autonomous miniature fixed wing unmanned aerial vehicles (mini-UAVs). The path representation takes into account aircraft dynamics by incorporating the turn rates and velocities of the UAV and follows a waypoint guidance method that is adopted in commercial aviation industry. The aircraft dynamics model allows the computation of fuel use, throttle, and velocity at different time instants throughout the path. A rigorous model validation is carried out prior to using the model for optimal path identification. An evolutionary algorithm is used to optimize the path distance and threat exposure encountered by the UAV for a mission. The optimization algorithm is a stochastic, zero order, elitist method similar in many respects to nondominated sorting genetic algorithm (NSGA-II) but includes explicit diversity maintaining mechanism in both the objective and variable space. A number of case studies are included to highlight the benefits offered by our approach.