基于时序逻辑任务的人机融合异构多智能体协同控制研究进展  

Research progress on the collaborative control of the human-machine fusion heterogeneous multi-agent based on temporal logic tasks

作　　者：张心骜 方浩[1] 赵欣悦 陈仲瑶 柯唯翎 ZHANG Xin'ao;FANG Hao;ZHAO Xinyue;CHEN Zhongyao;KE Weiling(School of Automation,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]北京理工大学自动化学院,北京100081

出　　处：《科技导报》2024年第12期167-177,共11页Science & Technology Review

基　　金：国家重点研发计划项目(2022YFA1004703)。

摘　　要：现有的基于时序逻辑任务的多体协同控制方法,通常采用将时序任务描述为形式化语言后,将其转化为自动机,并与环境模型做乘积,最后在乘积自动机中做图搜索等方式完成任务规划。对现有方法的优缺点进行了比对,从目前常用的结合时序逻辑语言的控制方法出发,梳理了人机融合异构团队控制方法、系统对任务违反程度鲁棒性控制和人机协作任务间的耦合任务分配这3项关键技术的发展脉络,并盘点了TSTL等新兴时序语言描述在人机融合架构中的良好表现。分析了当前该类协同控制研究存在的任务描述难、解耦分配难、在线计算量大的科学技术瓶颈问题。The existing multi-agent collaborative control methods based on temporal logic tasks usually adopt formal language to describe the temporal task,convert it into an automaton,multiplicate it with the environment model,and finally complete task planning by performing graph-search in the product automata.The advantages and disadvantages of existing methods are compared and summarized in this paper;Starting from the commonly used control methods which have combined temporal logic languages,the development of three key technologies,namely man-machine fusion heterogeneous team control method,system's robust control of task violation degree and coupling task assignment between human-machine collaboration tasks,are emphatically sorted out.Meanwhile,the good performance of emerging temporal languages such as TSTL in human-machine fusion architecture is also reviewed.Finally,an in-depth analysis was conducted on the scientific and technological bottlenecks in current research on collaborative control,including difficulty in task description,decoupling allocation,and large online computing load.These bottleneck issues which urgently need to be solved in the future development,have become the main study trend of this field in the future.

关 键 词：人机融合 时序逻辑 异构多智能体 协同控制 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]