UCAV自发育认知导航理论与关键技术研究

Intelligent autonomous navigation is one of the key supporting techniques to face the complex battlefield environment of Unmanned Combat Aerial Vehicles (UCAV) as an unmanned combat flatform that can be far from the control base station. It needs the requirements for UCAV in the aspects of entironment sensing and interactive ability, autonomous learning ability, consciousness ability and autonomous development trait. However, there are several counterchecking problems, such as the research paradigm for UCAV intelligent autonomous navigation having not been established, the ability to percept living surroundings needing to be enhanced, current intergrated navigation lacking the ability of entironment interaction. Therefore, we will introduce cognition concept and autonomous mental development (AMD) theory into UCAV intelligent navigation to simulate an open-ended, life-long, online and enhanced environment interaction of human cognition growth, and make in-depth research on the theory and key techniques of the autonomous developmental cognitive navigation (ADCN). We will mainly break through the key issues of ADCN architecture designing, battlefield awareness in unknown environment, autonomous developmental cognitive map (ADCM) building and high-precision positioning by humanoid SLAM (simulitaneous location and mapping). And finnally, a cognitive navigation theory will be established, some cognitive navigation techniques will be proposed. Our contributions will supply theoretical and technical supports for a new intelligent navigation system for UCAV. The foreseed achievements will also promote UAV navigation to develop to the direction of intelligence and autonomy, and furthermore, provide a technical reference for other unmanned combat flatforms of intelligent autonomous guiding. In sum, this project has an important theoretical significance and also a practical value.

UCAV作为能够远离操控基站的无人作战平台，智能自主导航是其适应复杂作战环境的核心支撑技术之一。为满足UCAV导航对环境感知与交互能力、自主学习能力、意识处理能力以及自主发育特性等方面的需求，解决目前UCAV智能导航研究范式尚未建立、生存环境感知能力亟待增强、组合导航不具备与环境交互能力等问题，本项目将认知概念和自主心智发育理论引入UCAV智能自主导航，模拟人类认知发育的开放、终身、在线强化的环境交互过程，开展自发育认知导航理论与关键技术研究，突破自发育认知导航结构设计、战场未知环境感知、自发育认知图构建、仿人同步定位与地图构建等关键问题，建立UCAV认知导航理论，提出UCAV认知导航方法，为实现UCAV智能自主导航提供理论和技术支撑。本项目的预期研究成果不仅能促进无人机导航向智能化、自主化方向发展，而且能为其它无人作战平台智能自主导引提供技术参考，具有十分重要的理论意义和实用价值。

为满足无人作战飞机（Unmanned Combat Aerial Vehicles，UCAV）导航过程中对环境感知与交互能力、自主学习能力、智能自主导航能力等需求，解决目前UCAV认知导航结构尚未建立、生存环境感知能力亟待增强等问题，本项目将认知理论引入UCAV智能自主导航，开展UCAV自发育认知导航理论与关键技术研究，重点研究了认知导航结构框架、环境特征感知、认知图构建、认知导航定位等内容，建立了UCAV认知导航理论技术体系。项目主要取得了以下研究成果：（1）设计了UCAV认知导航环，为UCAV认知导航的研究提供整体研究思路，为认知导航关键技术的研究奠定了基础；（2）提出了一系列环境特征提取方法，实现了UCAV认知导航过程高鲁棒性特征的有效获取；（3）针对基于位置的认知图和基于目标的认知图，提出了多种自发育认知图构建方法，实现了UCAV导航过程中已探索空间环境的表征；（4）提出了多种用于UCAV自主定位、目标指引（Goal-directed）导航和仿人同步定位与地图创建（Simultaneous Localization and Mapping，SLAM）方法，有效解决了认知导航过程中位置估计、面向目标导航知识生成等问题。针对研究内容，项目共发表学术论文22篇，其中国内外核心期刊19篇，被SCI检索4篇，EI检索12篇。培养博士研究生4名、硕士研究生1名，其中1篇硕士学位论文获得校级优秀学位论文奖。项目组成员参加国内外学术会议4人次。申报国防专利1项。完成了相应计算、仿真和分析程序的开发。本项目成果为提升导航系统智能自主能力提供了理论依据及技术支撑，为推进UCAV导航“无人化”发展具有重要的理论意义和实用价值。