毕业生知识库

近年来，无人机技术取得了飞速发展，无论在军事还是民用领域都取得了越来越广泛的应用。军事上，无人机应用逐步呈现出集群化、智能化和自组织化等特点，由此，自然产生了对于无人机集群协同智能决策技术研究的需要。无人机集群协同智能决策技术是决定具有自组织智能协同的无人机集群能否投入实战应用的关键技术问题，包括“智能通信组网”、“智能指挥控制”、“智能态势感知与评估”、“智能目标分配决策”、“智能航迹规划决策”、“智能飞行控制决策”、“智能武器分配决策”、“智能毁伤评估”等一系列领域技术的研究。

本文通过选取“智能态势感知与评估”、“智能协同目标分配”、“智能协同航迹规划”三方面智能决策技术，以群智能算法为主要研究对象，对无人机集群协同智能决策技术进行了简要研究：

首先针对现有以目标威胁值排序的主流态势评估方法不适用于同构无人机集群空中对抗任务的问题，引入一种以打击命中率和反击命中率为核心的新型态势表示形式，研究了一种结合模糊集理论与多属性分析、基于打击收益损失比矩阵的态势评估模型；

其次，在态势评估的基础上，基于对无人机集群协同目标分配的建模，针对群智能算法应用于无人机集群协同目标分配常发生的易过早陷入局部极值而无法跳出的算法“早熟”问题，以混合粒子群算法为研究对象，提出了基于引入“堵塞检测”机制与“强制打散”操作的自适应打散混合粒子群算法，研究了基于该算法的无人机集群协同目标分配；

而后，假设已知无人机集群协同目标分配最终分配结果的前提下，完成了对于单机在线航迹规划的建模，针对现有航迹规划算法无法适用于透明态势动态任务环境下对于移动任务目标在线航迹规划的问题，以狼群算法为研究对象，提出了一种基于改进狼群算法并结合滚动时域优化方法与二维等步长航迹生成方法的单机在线航迹规划算法，研究了该算法在所建模单机在线航迹规划上的仿真应用，并与原狼群算法进行了优化性能的差异比较；

最后，将上述对于单机在线航迹规划的研究拓展延伸到无人机集群协同在线航迹规划中来，在单机在线航迹规划问题建模的基础上引入协同性约束和要求重新构造无人机集群协同在线航迹规划模型，针对透明态势动态任务环境下对于移动任务目标的无人机集群协同在线航迹规划问题，以萤火虫算法为研究对象，提出了一种基于改进萤火虫算法并结合滚动时域优化方法与二维等步长航迹生成方法的无人机集群协同在线航迹规划算法，验证了该算法可以用于文中所建模的无人机集群协同在线航迹规划问题的优化求解，并与原萤火虫算法进行了优化性能的差异比较。

In recent years, UAV technology has achieved rapid development and has been more and more widely used in both military and civilian fields. In the military, UAV applications gradually present characteristics such as clustering, intelligence and self-organization, which naturally gives rise to the need for research on UAV swarm collaborative intelligent decision-making technology. The UAV swarm cooperative intelligent decision-making technology is a key technical issue that determines whether the UAV swarm with self-organized intelligent cooperation can be put into actual combat application, including "intelligent communication networking", "intelligent command and control", "intelligent Situation Awareness and Assessment", "Intelligent Target Assignment Decision", "Intelligent Track Planning Decision", "Intelligent Flight Control Decision", "Intelligent Weapon Assignment Decision", "Intelligent Damage Assessment", and a series of other fields.

Focusing on swarm intelligence algorithms, this paper briefly investigates the UAV swarm collaborative intelligent decision making technology by selecting three aspects of intelligent decision making techniques, namely "intelligent situational awareness and assessment", "intelligent collaborative target assignment" and "intelligent collaborative path planning":

Firstly, to address the problem that the existing mainstream situational assessment methods ranked by target threat values are not applicable to homogeneous UAV swarm air combat missions, a new form of situational representation centered on strike hit ratio and counterattack hit ratio is introduced, and a situational assessment model combining fuzzy set theory and multi-attribute analysis and based on the strike gain-loss ratio matrix is studied;

Secondly, based on the situation assessment and the swarm cooperative target allocation, this research proposes an improved hybrid particle swarm algorithm, which uses the “blockage detection” mechanism and “forced scattering” operation.  The proposed algorithm can address partly the problem of “premature”, which often occurs when the swarm intelligence algorithm is prone to local extremes for the application of UAV swarm cooperative target allocation;

Then, assuming that the final assignment result of UAV swarm cooperative target assignment is already known, the modeling of single aircraft online path planning is completed, and for the problem that the existing path planning algorithm cannot be applied to the online path planning of moving mission targets in the transparent dynamic mission environment, take the Wolf Pack algorithm as the research object, a single aircraft online path planning algorithm based on the improved Wolf Pack algorithm and combining the rolling time domain optimization method and the two-dimensional equal step length trajectory generation method is proposed. The simulation application of this algorithm on the modeled online path planning of a single aircraft is studied, and the difference of optimization performance is compared with the original Wolf Pack algorithm;

Finally, the above-mentioned research on single aircraft online path planning is extended to UAV swarm collaborative online path planning, and a collaborative UAV swarm collaborative online path planning model is reconstructed based on the modeling of single aircraft online path planning problem by introducing synergy constraints and requirements. Aiming to solve the UAV swarm collaborative online path planning of moving mission targets in the transparent dynamic mission environment, take the Firefly algorithm as the research object, an algorithm is proposed based on the improved Firefly algorithm and combines the rolling time domain optimization method and the two-dimensional equal step length trajectory generation method. It is verified that the proposed algorithm can be used to optimize the UAV swarm collaborative online path planning problem modeled in this previous, and the difference of optimization performance is compared with the original Firefly algorithm.