面向无人系统的网络协同理论与技术

The network coordination of unmanned systems is to organize unmanned robotic devices or vehicles into a network and coordinate their actions and contributions towards a common objective of a task. It is a promising way of breaking up the limit of a single unmanned system, cutting the overall cost of implementation and maintenance, and boosting the efficiency of task completion. In a network of unmanned systems, however, the links are strongly "Indeterminate"; the nodes are highly "Independent"; and the resources are significantly "Inhomogeneous". This project is motivated by such "3Is" properties, and investigates problems and scenarios unique to the coordination of networked unmanned systems. Specifically, this project focuses on three aspects of coordinating unmanned systems to develop connection, transmission, and computing related solutions, and proposes a framework for network coordination based on independent cognition, proactive network optimization, and connection control intelligence; a theory for highly reliable three-dimensional data transmission in air-ground networks; and a model for dynamic fusion of uncertain data and real-time computing. This project also develops an application for monitoring, search and rescue in ski resorts based on the network coordination of unmanned systems. This project aims to theoretically and technically contribute to the implementation and application of proactive, efficient, intelligent, and adaptive coordination of networked unmanned systems, through solving the challenges brought by the uniqueness of unmanned-system-involved tasks, the complexity of networking operations, the variety of network architectures for choice, and the discrepancy of radio access technologies.

面向无人系统的网络协同通过网络通信协调多个无人机器或装置，协同一致地完成复杂任务，是突破无人系统单机作业局限、降低无人系统综合作业成本、提高无人系统任务执行效率的有效途径。本项目深入分析无人系统网络协同的核心问题和复杂场景，针对无人系统网络所具有的连接强不确定性、节点高自主性和资源富异构性三大特点，研究无人系统网络协同架构及其基础理论和关键技术，从无人系统网络协同的“连”、“传”、“算”三个方面入手，提出自主认知、主动优化、智能连接的协同架构和组网策略、空天地一体化三维立体网络的高可靠传输理论、不确定信息的动态融合和时效计算模型，并构建面向滑雪场自主巡逻与应急救助应用的无人系统网络协同验证平台，解决无人系统业务类型的特殊性、网络行为的复杂性、组网模式的多样性及接入机制的差异性所带来的基础性挑战，为实现主动、高效、智能、弹性的无人系统网络协同提供理论基础和技术支撑。

面向无人系统的网络协同通过网络通信协调多个无人机器或装置，协同一致地完成复杂任务，是突破无人系统单机作业局限、降低无人系统综合作业成本、提高无人系统任务执行效率的有效途径。本项目深入分析无人系统网络协同的核心问题和复杂场景，针对无人系统网络所具有的连接强不确定性、节点高自主性和资源富异构性三大特点，从无人系统网络协同的“连”、“传”、“算”三个方面入手，重点研究一个体系架构、三大关键技术，并构建一个应用验证平台，为实现主动、高效、智能、弹性的无人系统网络协同提供理论基础和技术支撑。具体地，在体系架构方面，项目组提出了无人系统网络协同新型体系架构，厘清各技术模块间的相互依赖关系，建立协同运作规范和标准，可作为无人系统网络协同的基准架构；在关键技术方面，项目组研究攻克了面向动态异构复杂场景的强自主节点智能连接、面向三维网络的高可靠数据传输、异构不确定信息动态融合与时效计算核心关键技术，全面解决无人系统业务类型的特殊性、网络行为的复杂性、组网模式的多样性及接入机制的差异性所带来的基础性挑战，显著提升复杂协同任务完成的效率和质量；在验证平台方面，项目组基于上述理论成果，设计实现了无人系统网络协同验证平台并进行了应用示范，有效验证了本项目的相关理论研究成果。在五年执行期内，项目组在包括JSAC、TMC、TON、INFOCOM等国内外顶级期刊和会议上发表高质量论文75篇（其中CCF推荐的A/B类论文共42篇），申请国家发明专利19项（其中授权17项），制定ITU国际标准6项，培养博士研究生19人、硕士研究生37人，培养了一批无人网络领域高层次中青年人才，有力推动了无人系统相关领域的技术发展，圆满完成了预期目标。