Curvature constrained trajectory planning for a UAV through a sequence of points: A perturbation approach

Path planning for a UAV though a sequence of points is an important problem in many application domains. We use Dubins curvature constrained vehicle model [1] for the UAV. We design a scheme to develop approximate solutions to the curvature constrained sequence problem by allowing small perturbations to point locations in lieu of solutions with path length values, close to known lower bounds. The approach develops characterizations of discontinuities in the path length function and performs perturbation at these values. Our approach is straight forward to implement. It is practically useful as in many real world applications of UAVs, small deviations from exact point locations do not affect the mission objectives.