面向巡检任务的无人机三维航迹自适应规划算法

An Adaptive Planning Algorithm for 3D Flight Track of Unmanned Aerial Vehicle in Patrol Tasks

关键词:  巡检无人机  航迹规划  爬行动物搜索算法  混沌映射  威布尔飞行算子

Key words:  patrol unmanned aerial vehicle  flight track planning  reptile search algorithm  chaotic map  Weibull flight operator

摘要:

传统方法求解复杂环境下巡检无人机航迹规划问题，容易出现收敛于局部最优航迹、全局搜索精度差等不足。为此，本研究提出一种非线性威布尔飞行爬行动物搜索算法(Non-linear Weibull Flight Reptile Search Algorithm,NWFRSA)，用于巡检无人机航迹自适应规划。通过引入改进无限折叠迭代混沌映射(Iterative Chaotic Map with Infinite Collapses,ICMIC)初始化策略来提升初始航迹解的多样性和质量，利用S型特征函数非线性进化因子实现全局最优航迹的探采均衡，并设计威布尔飞行算子变异策略使算法能跳离局部最优航迹，丰富搜索空间并提高收敛精度。同时，本研究构建了巡检无人机的航迹规划模型，设计了考虑航迹总长度、飞行高度、飞行转角及威胁模型的权重目标函数，将三维空间内的航迹规划转换为约束条件下的多目标优化问题。研究结果表明，在不同复杂程度障碍物与威胁区域分布的飞行环境下，相比于粒子群优化算法(PSO)、蝴蝶优化算法(BOA)和爬行动物搜索算法(RSA)，NWFRSA均能有效减小航迹代价(4.53%-34.47%)，有助于提高巡检任务效率和安全性。

Abstract:

Conventional methods are prone to convergence to the local optimal flight track and poor precision of global searching when solving the flight track planning problem of patrol unmanned aerial vehicle in complex environment.A Non-linear Weibull Flight Reptile Search Algorithm (NWFRSA) is proposed for adaptive planning of patrol unmanned aerial vehicle track.An improved Iterative Chaotic Map with Infinite Collapses (ICMIC) initialization strategy is introduced to improve the diversity and quality of the initial flight track solutions.The non-linear evolutionary factor of S-type feature function is used for the exploration and mining equilibrium of the global optimal flight track.The Weibull flight operator mutation strategy is designed to make the algorithm skip from a local optimal flight track,enrich the search space,and improve the convergence precision.Meanwhile,the flight track planning model of patrol unmanned aerial vehicle is constructed,and the weight objective function of the model considering flight track length,flight altitude,flight angle,and threat is designed.The flight track planning in three-dimensional space is converted into a multi-objective optimization problem under constraints.Experimental results show that in flight environments with different levels of complexity and distribution of obstacle threat areas,compared with Particle Swarm Optimization (PSO),Butterfly Optimization Algorithm (BOA),and Reptile Search Algorithm (RSA),NWFRSA can effectively reduce the cost of track(4.53%-34.47%),improving the efficiency and safety of patrol tasks.