计算机科学与探索2026，Vol.20Issue(6)：1647-1665,19.DOI:10.3778/j.issn.1673-9418.2507094
面向任务关键群智能系统的微服务化弹性协同计算体系
Microservice-Based Elastic Collaborative Computing Architecture for Mission-Critical Swarm Intelligent Systems
摘要
Abstract
Autonomous decision-making,multidimensional collaboration,and robust operation are key supports for swarm intelligent systems such as UAV swarms to accomplish mission objectives with high quality.However,due to the tight logical and physical coupling among code,resources,and platforms,swarm systems based on static computing architec-tures can achieve only limited collaboration at the behavior,function,and subtask levels,and only within limited dimen-sions and depths,which fundamentally constrains the development of advanced swarm intelligence.To address this issue,based on an in-depth analysis of the connotation of deep collaboration in swarm intelligent systems and the new computing requirements arising therefrom,this paper integrates the ideas of service-oriented architecture(SOA),robot operating sys-tem 2(ROS2),container virtualization,and microservice collaboration,and studies and designs a microservice-based elastic collaborative computing architecture(MSECCA)that provides new computing capabilities such as elastic organization,dynamic construction,and capability interoperability.In particular,this paper focuses on the study and design of container-based virtualized microservice encapsulation for heterogeneous resources and tasks,thereby decoupling physical resources,tasks,and platforms,and laying the foundation for remote capability sharing and dynamic migration and construction within the swarm.Furthermore,a runtime logic model for task microservices and a spatiotemporal composite event-driven mechanism are studied and designed,and a multi-agent-based distributed collaborative scheduling mechanism is further proposed.Meanwhile,a swarm messaging mechanism with support for communication consistency constraints is studied and designed,which eliminates the communication coupling among microservices as well as the limitations imposed by heterogeneous communication channels in the swarm on message delivery,and provides capabilities such as open and dynamic access and task microservice migration.The prototype system developed based on this theoretical framework has been validated in a semi-physical simulation environment,and the validation results show that the proposed system has good technical feasibility in the field of collaborative operations of UAV swarms.关键词
群智能系统/机器人操作系统(ROS2)/微服务/协同/弹性体系Key words
swarm intelligent systems/robot operating system 2(ROS2)/microservices/collaboration/elastic architecture分类
信息技术与安全科学引用本文复制引用
张家铭,何周灿,张凯龙,陈树炎,王涵涵,刘欢..面向任务关键群智能系统的微服务化弹性协同计算体系[J].计算机科学与探索,2026,20(6):1647-1665,19.基金项目
国家自然科学基金面上项目(61972318) （61972318）
陕西省科技计划项目(2023-GHZD-47) （2023-GHZD-47）
航空基金项目(20171953016). This work was supported by the General Program of the National Natural Science Foundation of China(61972318),the Shaanxi Provin-cial Science and Technology Planning Program(2023-GHZD-47),and the Aviation Science Foundation(20171953016). （20171953016）
