基于TLD模型的UAV三维实时平滑航迹规划  

3-D real-time smooth path planning for UAV based on TLD model

作　　者：李牧东[1,2] 赵辉[1] 黄汉桥[1] 翁兴伟[1] 

机构地区：[1]空军工程大学航空航天工程学院,陕西西安710038 [2]复杂航空系统仿真重点实验室,北京100076

出　　处：《系统工程与电子技术》2017年第1期93-100,共8页Systems Engineering and Electronics

基　　金：航空科学基金(20155196022);国家青年科学基金(71501184)资助课题

摘　　要：针对无人机三维实时航迹规划问题,提出了一种基于三层决策模型的平滑航迹规划方法,以提高其规划效率和可操作性。首先根据无人机三维飞行的特点,并结合自身性能、威胁和障碍以及地形等约束,设计三层决策目标函数和决策变量;其次为了避免不必要的迂回和路径,提高航迹的平滑性,提出了元胞化地图的变长探测方法和启发式优化策略;最后,利用所提出的认知行为优化算法对这一问题进行求解,以提高规划路径的搜索效率。不同场景地形的仿真及与现有经典方法的比较结果表明:该方法能有效实时规避威胁和地形障碍,且具有较高的可执行性,能够快速生成安全、平滑的飞行路径。Aiming at the problem of three-dimensional （3-D） real-time path planning for the unmanned aeri- al vehicle （UAV）, a smooth planning method based on the tri-level decision-making （TLD） model is proposed so as to improve the planning efficiency and touters maneuverability. Firstly, considering the constraints of UAV including self-performance, threats and obstacles, and terrain, TLD objective functions and decision vari- ables are designed according to the characteristics of 3-D flying. Secondly, the cellular map with variable detec- tion length method and the heuristic optimization strategy are proposed to avoid the unnecessary detours and paths as well as improve the smooth performance of the flight path. Finally, the proposed algorithm named cog- nitive behavior optimization algorithm （COA） is introduced to solve this math model in order to improve the search efficiency of planning path. The results obtained by different scenario terrain simulation and comparisons with other classic methods indicate that the proposed approach can real-time avoid threats and terrain obstacles effectively. Meanwhile, it has a high enforceability and can generate the safe and smooth flight path quickly.

关 键 词：无人机 三维实时航迹规划 三层决策规划 认知行为优化算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]