基于智能Agent的小型无人机自主飞行任务管理器的设计  

Intelligent Agent-Based Design and Implementation of Autonomous Task Manager for Small UAVs

作　　者：谭雁英[1] 赵荣椿[2] 祝小平[1] 王兴海[1] 

机构地区：[1]西北工业大学无人机研究所 [2]西北工业大学计算机学院,陕西西安710072

出　　处：《西北工业大学学报》2006年第6期754-758,共5页Journal of Northwestern Polytechnical University

摘　　要：自主飞行是先进无人机的重要特征。基于智能A gent技术,采用分层的混合型A gent结构,进行了小型无人机自主飞行任务管理器的设计与实现。该设计方案在吸收各领域专家及操纵手的专门知识和丰富经验的基础上,建立了一套基于认知型A gent结构的顶层自主任务管理推理机制和策略,以及基于反应型A gent结构的任务规划执行底层的规则集。对基于该任务管理器的全自主飞行、半自主飞行模式进行了几十架次的飞行试验验证。试验结果充分表明了其推理机制和策略的有效性以及对应急事件响应的快速性。Autonomy is important for advanced uninhabited air vehicles （UAVs）. On the basis of the intelligent agent technique, we apply hierarchy and hybrid agent structure to designing and implementing the software of an autonomous task manager for small UAVs. In the full paper, we explain the design and implementation in detail. In this abstract, we just add some pertinent remarks to listing the three topics of explanation： （1） general layout of intelligent agent-based autonomous task manager; （2） design of intelligent agent-based autonomous task manager; （3） implementation of software; the subtopics of topic 1 are the selection of the structure type of intelligent agent （subtopic 1.1） and the specific structure of the autonomous task manager（subtopic 1.2）; under subtopic 1.2, Fig. 1 in the full paper shows the schematic of the hierarchical structure of the autonomous task manager; the hierarchical structure consists of the execution level, the navigation management level and the top level for autonomous task management ; t.he subtopics of topic 2 are the design of the execution level （subtopic 2. 1）, the design of the navigation management level（subtopic 2.2） and the design of the top level（subtopic 2.3）; under topic 2, for all three levels, we establish rule-based strategies based on the acquired special knowledge and rich experience of experts and operators. Both the fully-autonomous and semi-autonomous flight modes are successfully demonstrated in flight tests. The several tens of flight tests show preliminarily the effectiveness of the designed strategies and the rapid response of the autonomous task manager to contingency flight events.

关 键 词：无人机 自主飞行任务管理器 智能AGENT 混合型结构 分层 任务再规划 策略 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]