Research On UAV Trajectory Planning Algorithms

As the unmanned aerial vehicle (UAV) has broad application prospects in military field and civil field, the UAV technology has been widely concerned and applied in the world. The UAV trajectory planning is the key technology that guarantees UAV to achieve autonomous flight and it also plays a key role in completing the UAV missions. This dissertation concentrated on the research works listed below and achieved some creative results:1. It described the UAV flight management system briefly, with the aim to present the content of flight path planning and its role in management system.2. Safe surface was gotten by digital map preprocessing techniques and then the vertical maneuverable parameters of aircraft include maximum climbing angle (slope of the terrain) and maximum normal overload (curvature of the terrain) could be met. Then a three-dimensional path coinciding with the dynamic constraints of UAV can be got from safe surface and two-dimensional path.3. The work compared the advantages and disadvantages of conventional modeling methods, focusing on analyzing the disadvantages of Voronoi diagram and giving the corresponding enhancement methods. Voronoi diagram would be constructed by using the new feasible point generating method that consider threat radius. Safe corridor was built for every edge of Voronoi diagram and surface features and radar threats were quantized with a group of points that were used to determine the average cost of each edge. Simulation results highlighted that threats can be avoided more effectively.4. Cloud self-adaptive genetic algorithm was proposed in this dissertation with the aim to overcome the shortcomings of GA. i.e., low convergence rate and easily getting a local optimum solution. The probabilities of crossover and mutation were adaptively generated by X-conditional cloud generator. Set evolution generation to a fixed value, search flight path using general decoding method and heuristic decoding method respectively. Calculate the cost in different situations and different number of populations. The results were presented in this dissertation.5. Improved chaotic genetic algorithm was proposed. The characteristics of chaos were simulated and analyzed. In order to improve the average of the distribution of chaotic sequence, the concept of random exponent in determinate range was proposed and the concept was applied to the chaos genetic algorithm.6. Real-time planning method was proposed according to the principle of partial reconstruction of Voronoi diagram. It focused on the situation that a single sudden threat or more threats were detected.