Research And Implementation Of Route Planning Based On Genetic Algorithm

Since there existing critical radar threats to UAV in the high altitude airspace, it is of great importance to carve out an airway which can succeed in avoiding these radar threats in its flight area, but only when the airway meets all the constraints of the UAV including performance limitations and requested tasks, it can be used.(1) Aimed at the complex flight environment and coupling constraints of drone, the writer has done a lot of research on the application of genetic algorithms on airway planning by analyzing and comparing various algorithms, and presents a lot of improvement and innovation in this paper by studying on simple genetic algorithms. Firstly, this paper presents a dynamic memory system by using the actual genetic codes when airway nodes map the genetic algorithms chromosome, this memory system helps to greatly increase the efficiency of the algorithm by reducing the difficulty of decode and decreasing the search time for genetic algorithms. Secondly, this paper puts forward an initialization method to prevent airway nodes overlap(chromogene overlap), and presents a cost function to assess the airway on the base of the theory that Fitness is the only indicator for genetic evolution to decide what to choose. In a result, the assessment will be improved with the more precise value of fitness by analyzing the weighting factors of cost function under specific environment and tasks through AHP(Analytic Hierarchy Process). Finally, due to the weakness that simple genetic algorithm easily fall into local optimum, the writer introduces self-adapted genetic factors which can improve the simple genetic algorithm and help it be used better in searching optimum route for unmanned aerial vehicle in the high altitude airspace.Design Hybrid genetic algorithm of using simulated annealing algorithm to improve the current algorithm. Compared with the current algorithm, the Hybrid genetic algorithm can reduce the search time and improve search precision.(2) All the factors including the constantly changing threat, drone’s changing performance limitation and perform tasks are obstacles for the research in this paper, but the writer resolve the problem.by define a new begin node and design parallel genetic algorithm of using multithreading technology to re-plan the route.(3) In order to simulate the actual flight environment, this paper presents an intelligent manual-operated interface platform to simulate the 3D airspace environment of drone. The different 3D airspace models of the changing radar threats can be seen by setting up and changing the parameters of the radar threats with mouse or keyboard on the platform. Finally, achieve route planning by packing the algorithm with using Dynamic Link Library.Depend the experiment results,using Genetic algorithms for high-altitude and high-speed UAV route planning can effectively avoid radar threats, the Hybrid genetic algorithm can reduce the search time and improve search precision, It is succeeds to verify that Genetic Algorithm is applicable to route planning. Parallel genetic algorithm can effectively improve planning efficiency and meet the partial amend request. The designed interface can be used as a planning platform.