A Particle Swarm Approach for Flight Path Optimization in a Constrained Environment

Potentials of Particle Swarm Optimization techniques in the field of flight path generation are investigated in this paper. The objective is the development of a particle swarm-based procedure performing off-line two-dimensional flight path optimizations compliant with operational constraints. Assuming a typical surveillance mission, such constraints are defined in terms of “target” and “no-fly” zones, desired flight time on targets, fixed way-points and landing areas. Flight path is described through a discrete number of suitable flight parameters varying within defined ranges, each particle of the swarm being represented by a numerical combination of these parameters. A novel obstacle avoidance model based on the evaluation of a so-called area function has been introduced. All trajectories, made up of circular arcs and straight lines, start from a specified point with fixed velocity vector, ending within a selected landing area. Both single-objective and multi-objective optimization procedures have been formulated and implemented. The former minimize the total flight path length; the latter also try to maximize the trajectory length covered over specified target areas. Sensitivity studies with increasing problem complexity have been performed changing both number and positions of “target” and “no-fly” zones in order to assess the effectiveness as well as the robustness of the proposed approach. Algorithm capability in finding the optimum path through non-circular and concave obstacles has also been tested. Finally computational time monitoring allowed making a preliminary assessment of Particle Swarm Optimization suitability for practical applications.