Flight path planning based on niche genetic algorithm

An improved niche genetic-algorithm-based approach to the problem of flight path planning was proposed. The vehicle path as a sequence of speed and heading transitions occurring at discrete times was modeled, and this model specifically contains the vehicle dynamic constraints in the generation of trial solutions. The initial trial solutions were not generated randomly, but generated according to the initial environment. The initial location and goal were connected, and then the initial heading was calculated. The initial population was generated according to the initial velocity and heading. To prevent the population prematurely reaching local minima, fitness sharing method was introduced to the algorithm. Simulation studies show that the proposed algorithm is more quickly than the previous algorithm in finding a near-optimal obstacle-free path in a dynamically changing environment.