【摘要】 以有人/无人系统协同作战为背景，为实现指挥员与多智能体系统的高效协同作战，在多智能体系统完成高效准确的态势理解基础上，开展战场任务理解研究，并给出多智能体系统任务理解的定义。基于四类典型的作战任务，构建多智能体系统的任务理解框架。同时，针对其中的两类典型作战任务，采用基于极限学习机的任务理解算法实现对指挥员指令快速准确的理解。通过仿真实验，验证了在该任务理解框架下，多智能体系统能够有效理解指挥员的作战指令，更高效地完成作战任务，实现指挥员和多智能体系统的协同作战。更多还原

【Abstract】 In order to realize the efficient cooperative operation between the commander and multi-agent system,the battlefield mission understanding research is carried out based on the efficient and accurate situation understanding of multi-agent system.The definition of mission understanding of multi-agent system is given.In view of four typical combat missions,the mission understanding framework of multi-agent system is constructed.At the same time,according to two kinds of typical combat missions,extreme learning machine is used as the mission understanding algorithm to realize the rapid and accurate understanding of commander’s instructions.The simulation results show that the multi-agent system can effectively understand the commander’s operation instructions and can complete the operation mission more efficiently.Therefore the cooperative operation between the commander and the multi-agent system is realized.更多还原