| 要約: | In the three-dimensional terrain environment, the basic fireworks algorithm is easy to fall into the optimal solution and has the problem of slow convergence when path planning.A selection crossover fireworks algorithm is proposed.The grid method is used to build a three-dimensional terrain environment and set a threat area such that an unmanned vehicle can select appropriate nodes to explore the path.The fitness function is derived in combination with the fuel costs, smooth costs, and threat costs to restrict the generation position of the path nodes and ensure that the planned road path is smooth and far from the threat area.A path search is conducted via the explosion, mutation, mapping, and selection operations of the basic fireworks algorithm.In addition, the roulette selection operation for the path nodes is added such that the nodes farther from the original path have a higher explosion probability to restrict the search direction of the path, thereby speeding up the search speed of the algorithm.Based on this, a selection crossover spark is introduced to enhance the information interaction between fireworks in the population and improve the performance of searching the global optimal solution by crossing the path segments between nodes after wheel selection.The simulation results show that compared with the basic fireworks algorithm, the fitness of the proposed algorithm in the simple and complex terrain environments improves by 6% on average, and the running time shortens by 13.5% on average.In various terrain environments, the unmanned vehicle can successfully evade the threat area using the proposed algorithm and find a smoother path with lower fuel cost within a relatively short time.
|
|---|
