Abstract: Aiming at the "rich data but poor information" problem under Ubiquitous computing-based intelligent Manufacturing(U-Manufacturing)environment, a semantic analysis principle to extract information from discrete shop logistics real-time data effectively was proposed. Radio Frequency IDentification(RFID)technology was used to collect real-time logistics data in process route. Complex Event Process(CEP)method was applied to aggregate the simple logistics event(e.g. RFID reading event)into six complex logistics events. Logistics Progress Matrix(LPM)was consequently established to record the occurrence of each complex logistics event at different time, and the rank and elements of LPM were introduced as the main parameters to reflect the execution progress of single production batch. The operational procedure of the above semantic analysis principle was given and its corresponding prototype system was developed. The feasibility of the proposed principle was proved by a practical application in an auto parts manufacturer.

Key words: ubiquitous computing-based intelligent manufacturing, discrete shop, radio frequency identification, real-time logistics data, semantic analysis, complex event process

摘要： 针对U-制造环境下"数据有余、信息不足"的典型问题，研究了从车间物流实时数据中有效提取信息的语义分析原理。采用射频识别技术采集工艺路线物流实时数据；通过复杂事件处理方法将简单事件（即射频识别读事件）聚合为6种常见的物流复杂事件；构建物流进度矩阵完整地记录不同时刻复杂事件的发生，以矩阵的秩和元素作为主要参数反映和描述单批次生产任务的执行进度。给出了所提原理的运行模式，开发了对应的原型系统。通过某汽车电机制造商的应用案例验证了所提原理的可行性。

关键词: U-制造, 离散车间, 射频识别技术, 实时物流数据, 语义分析, 复杂事件处理