基于进化计算的无人机集群协同作战规划关键技术研究

Unmanned Combat Aerial Vehicle (UCAV) is a new concept of combat mode, with a high degree of intelligence fault-tolerance and combat effectiveness. The intelligent decision-making and control is the key to UAV cluster collaborative operations. To this end, based on the latest theory and technologies of system engineering, simulation modelling and evolutionary computation, this project aims to solve several key problems in the UCAV cluster cooperative operation. That is, i) the structure of UCAV cluster cooperative combat system, ii) multi-objective optimization model of UCAV 3D static/online path planning; UCAV cluster task pre-allocation and redistribution mechanism; and iii) the design of (dynamic) evolutionary multi-objective algorithms for problem solving. Finally, an UCAV cluster cooperative simulation system will be developed to validate the proposed model and algorithms. The research results can enrich the application of evolutionary algorithms (EAs) to large scale dynamic multi-objective optimization, especially for the UCAV cluster collaborative operations. Also, the research can provide theoretical basis and technical support for UCAV cluster collaborative operations, which will help to enhance the operational efficiency of UCAV cluster, and provide solid support for the intelligent joint operation in future.

无人作战飞机(UCAV）集群协同作战是一种全新概念的作战模式，具备智能化程度高、容错性高、作战效能显著等特点，智能决策与控制是实现UCAV 集群协同作战的关键，为此，课题以系统工程、仿真建模、进化计算等学科的最新理论和技术为基础，提出研究基于进化计算的UCAV集群协同作战规划关键技术，包括基于多agent理论的UCAV集群协同作战体系结构；基于多目标优化的UCAV三维航迹静态和在线动态规划建模；UCAV集群大规模攻防任务预分配和突发机制下的任务重分配多目标优化建模；面向航迹规划和任务分配的（动态）智能多目标优化算法设计。最后开发UCAV集群协同作战原型仿真系统用于验证模型及算法。研究成果从理论上丰富了进化计算在大规模动态多目标优化，特别是UCAV协同作战规划领域的应用，在实践上有助于提升UCAV集群作战效能，为打赢未来智能化联合作战提供有力的支撑。

无人机集群协同是未来联合作战的重要作战样式，项目以系统工程、仿真建模、进化计算等理论和方法，深入研究了无人机集群协同作战的系列关键技术，具体提出了基于多agent理论的集中-分布混合式UCAV集群协同控制体系架构；构建了考虑费用、威胁等多个指标的无人机集群三维航迹多目标静态规划模型及在线动态规划模型；构建了平衡代价和效能的无人机集群大规模任务预分配模型及动态重分配模型；分别结合航迹静态规划、任务预分配模型的特征，提出了面向大规模多目标优化问题的智能算法，结合在线航迹规划及动态任务重分配的特征，提出了基于轨迹预测的动态智能优化算法；最后开发了无人机集群协同作战仿真系统验证了模型和算法的性能。项目执行过程中发表高水平论文27篇，其中IEEE汇刊9篇，累计谷歌引用5000余次，出版专著1部，第一发明人受理授权专利7项，项目负责人受邀担任智能优化国际知名期刊IEEE Trans on Evolutionary Computation（IF=11.5）、Swarm and Evolutionary Computation（IF=7.6）等副主编，自动化学报（英文版）、系统仿真学报首届青年编委，荣获中国仿真学会自然科学一等奖（排名1），湖南省青年科技奖，入选军队高层次科技创新人才程，并进一步获得了国家优秀青年科学基金资助。