摘要
智能多源自主导航系统以北斗导航为基石、以惯性导航为支撑,旨在提供高度智能化的定位导航与授时服务,通过实现即插即入、无缝衔接、无感切换、可信完备等功能,以满足复杂多变的场景需求.采用以深度学习、强化学习为代表的智能学习方法,从复杂性与多尺度视角出发,探索智能多源自主导航中的优化与协同理论,建立以精准稳定、可信安全、可泛化、可解释、内嵌动力学机理为关键要素的智能学习方法,推动构建弹性、健壮、动态、可控与可信的精准智能多源自主导航模型与方法体系,从而有力支撑国家综合定位导航与授时体系的创新发展和先进导航技术的规模化应用.
Abstract
The intelligent multi-source autonomous navigation system(IMANS),based on Beidou navigation and supported by inertial navigation,aims to deliver advanced PNT(positioning,navigation,and timing)services.It offers plug-and-play functionality,seamless sensor integration,imperceptible switching,and dependable performance to meet the challenges of complex and dynamic environments.By leveraging intelligent learning techniques,such as deep learning and reinforcement learning,and applying multi-scale analysis and complexity theory,the optimization and collaboration theories in IMANS are explored.This paper then develops intelligent learning frameworks prioritizing precision,stability,reliability,safety,generalizability,interpretability,and embedded system dynamics.These innovations contribute to the creation of resilient,robust,dynamic,controllable,and dependable precise IMANS models and methodology,driving the advancement of the national comprehensive PNT system and enabling the large-scale adoption of advanced navigation technologies.
NETL
NSTL
万方数据