Due to the rapid development in IoT technologies, real-time manufacturing data capturing attracts increasing attention from both industrial field and academia. RFID is the key technology of IoT, and it has been widely studied to capture the manufacturing data. To tackle scheduling inefficiency resulting from paper-based identification and manual data collections, Zhong et al. presented an RFID-enabled real-time manufacturing execution system, which is fulfilled by systematically configuring RFID devices on the shop floor to track and trace manufacturing objects and collect real-time production data [17]. Yang et al. presented an RFID-enabled indoor locating method for a real-time manufacturing execution system using online sequential extreme learning machine [18]. Lee et al. proposed an RFID-based resource assignment system for garment manufacturing [19]. Liu et al. designed a new production management system by integrating with an RFID-enabled real-time data capturing system for Loncin motorcycle assembly line [20]. The RFID-based remote monitoring system is proposed to provide a transparent and visible information flow for supply chain and enterprise internal resource management [21].

The sensor management is essential in providing reliable manufacturing data. However, the current connect number, sampling rate, and signal types of sensors are generally restricted by the device, a unified sensor management model plays an important role in manufacturing information sensing. Chi et al. proposed a reconfigurable smart sensor interface for industrial wireless sensor network in IoT environment, in which complex programmable logic device was used as the central manager [29]. Mayer et al. presented a service composition system that enables the goal-driven configuration of smart environments for end users by combining semantic metadata and reasoning with a visual modeling tool [30]. Boonma and Suzuki proposed a biologically inspired middleware architecture for self-managing wireless sensor networks [31]. Jung et al. presented a new architecture of Hadoop-based distributed sensor node management system for distributed sensor node management using Hadoop map reduce framework and distributed file system [32]. Wang et al. proposed a Hadoop-based approach for web service management in telecommunication and Internet domains, the basic idea is to adopt two components of Hadoop, that is, HBase, and MapReduce, to manage web services [33]. Kulvatunyou et al. analyzed the OWL-based semantic mediation approaches to enhance manufacturing service capability models [34]. Fröhlich and Wanner presented a run-time support environment for wireless sensor network applications based on the operating system. It allowed the applications to configure their communication channel according to their needs and acquire sensor data through a unified application program interface (API) [35]. Hu et al. proposed a real-time discrete event–based monitoring system for RFID-enabled shop-floor monitoring in manufacturing enterprises. The system used rigorous mathematical techniques for event construction, state prediction, and disturbance detection that are appropriate for big-data environments of current intricate manufacturing systems [36]. Fang et al. presented an agent-based GOS to manage the RFID sensors for the ubiquitous company [16]. Considering the requirements of integration between production activities and enterprise information systems, an integration framework for RFID middleware based on business process rule and data stream technologies are presented by Li and Li [37].

Manufacturing information captured by multiple sensors are often with different format, thus they need to be processed to obtain information in standard schema, and how to share the information with different EISs are also an important question. To address this demand, Anaya et al. described a unified enterprise modeling language, which aimed to backup integrated use of enterprise information models expressed using different languages [38]. Pantazoglou et al. proposed Proteus, a generic query model for the discovery of operations offered by heterogeneous services [39]. Al-Fuqaha et al. designed a rule-based intelligent gateway that can bridge the gap between existing IoT protocols to manage the efficient integration of parallel IoT services [40]. In order to obtain the formal description of manufacturing capability, Luo et al. studied a modeling and description method of multidimensional information for manufacturing capability in CM system [41]. Zhang et al. presented a services encapsulation and virtualization access model for manufacturing machine by combing the IoT techniques and cloud computing [42], and a task-driven manufacturing cloud service proactive discovery and optimal configuration method [43]. Russom et al. discussed the implementation of a configurable laboratory information management system for use in cellular process development and manufacturing [44]. To help engineers use digitization practically and efficiently, Kojima et al. proposed a method based on manufacturing case data that had a direct relation to manufacturing operations, where the data were represented in XML schema [45]. Pang et al. discussed the data-source interoperability service for heterogeneous information integration in ubiquitous enterprises [46]. Chungoora et al. studied the interoperable manufacturing knowledge systems concept, which used an expressive ontological method that derived the improved configuration of product life cycle management systems for manufacturing knowledge sharing [47]. Wu et al. developed a method to improve the integration and sharing of product knowledge during the development phase, the study established product development knowledge integration ontology [48]. Dhokia et al. developed a data model and a prototype information sharing platform for DEMAT machine tools, and the platform provided a unique method to monitor the life cycle status of machines [49]. Qiu et al. proposed an IoT-enabled Supply Hub in Industrial Park for improving the effectiveness and efficiency of sharing physical assets and services [50].

This module aims to construct a smart environment by extending the advanced IoT technologies to traditional manufacturing field. First, the information requirements of the manufacturing system are analyzed. Second, the sensors types (e.g., RFID, barcode, smart meters, and so on) are selected according to the information requirements. As a result, the sensors are deployed in the shop-floor frontline, and the RMMI can be timely sensed by the up-level applications. More details will be presented in Section 3.4.

As shown in the middle of Fig. 3.1, multiple sensors manager is used to manage the sensors, so that the sensors can work properly. First, the sensors are registered into the multisensor manager, where the parameters of the sensors are set. Second, the sensor drivers are installed to ensure the sensors operate reliably. Third, the binding models between the resources and sensors are established. At last, the operating status of sensors is monitored by this module, so that the sensor exceptions can be eliminated as soon as possible. More details will be presented in Section 3.5.

Huge amounts of real-time manufacturing data are captured by the sensors. However, most of the data are duplicated, meaningless, and unreliable, and only limited quantities of manufacturing events are taken care by the up-level users. The multisource manufacturing information processing and sharing module is used to meet the gaps between the raw sensor data and EISs requirements. Three kinds of contents are discussed in this module, namely data preprocessing, information encapsulation, and information sharing mechanism. Data preprocessing is used to aggregate the discrete data into resource level event. Information encapsulation is used to encapsulate into a standard information template which is based on B2MML schema. Information sharing mechanism is used to transfer the qualified data into the related users. More details will be discussed in Section 3.6.

Since many manufacturing resources are involved in the shop floor and their status and capacity data are changing continuously, a vast number of manufacturing data are created. Thus, it is important to select the key monitor points of manufacturing information. As shown in Fig. 3.2, five kinds of key information are interested by the managers, that is, machine, process quality, manufacturing objects, worker, and environment, and they are described as follows.

As discussed previously, the information requirements in manufacturing field involves many things in productive processes, such as materials, WIPs, personnel, tools, equipment, vehicles, and so on. The real-time condition monitor of the manufacturing things has large impact on the performance of the entire production system. According to our recent investigation of several collaborative manufacturing enterprises, many of them use the manual system for data collection. Thus, the collected data are often laggard in time, prone to errors, and tedious. It is a daunting task to trace and track WIP items in a large manufacturing company. Besides, manual identification sheets are frequently damaged, lost, or misplaced, and shop-floor operators are busy with operations that are supposed to add values to products. As a result, the captured information can not accurately and promptly reflect the real-time status. In order to capture the real-time information of manufacturing resources, we choose some typical sensor and deploy them on suitable objects. The deployment information is shown in Table 3.1.

After the multiple sensors are deployed and configured into the traditional manufacturing field, the real-time data of manufacturing resources will be captured continually. However, only a few data reflects the useful manufacturing information. To fulfill this gap, a two-level event model is used to process the raw sensor data into resource status data. Manufacturing data preprocessing helps users to obtain more meaningful and actionable resources information from the large amount of raw sensor events.

Since the output of each sensor is isolated manufacturing data, it is difficult to share them with heterogeneous EISs. A business to manufacturing markup language (B2MML)-based RMMI template is proposed to encapsulate the discrete manufacturing data. After the encapsulation, the manufacturing information can be stored under a standard uniform, which can be accessed easily by the different managers. For that the manufacturing information includes information related to equipment, material, personnel, equipment, environment, and WIP, each of them are defined as an element in the template. Besides, the processing method is added in the template as an element to describe how the objects are processed. The B2MML-based real-time and multisource manufacturing information template is given in Fig. 3.4.

Manufacturing information sharing module is responsible for sharing the RMMI with third-party applications. As seen in Fig. 3.5, two communication methods are used to share the manufacturing data, namely “Push model” and “Get model.” Both of the two models ask the users to register in this system first. Then, the real-time and multisource manufacturing data can be published to the users by the different information communication methods. For Push model, the user needs to submit some basic information of himself, including name, staff ID, telephone number, department, position, the kind of subscribed information, etc. When RMMISS captures the subscribed information, it will send a message which contains standardized real-time information to the relevant staffs through Wi-Fi, GSM, and other communication network. In the Get model, system will send captured information to users at the predefined time points T.

Based on the aforementioned architecture of RMMISS, the hardware device and its main components are designed as shown in Fig. 3.6. Fig. 3.6A presents the concept design, it mainly consists of monitor, industrial computer, heterogeneous interface hub, wireless module, GSM, and multiple sensors. As the name suggests, the monitor is used to display the RMMI. The industrial computer is responsible for connecting and communicating with all kinds of smart objects via wired or wireless methods. The heterogeneous interface hub is used manage the multiple sensors through wired way. The wireless module and GSM are used to connect the multiple sensors in a wireless way. For definiteness and without loss of generality, we only choose two types of sensors, namely, RFID and digital caliper, in the physical realization, as shown in Fig. 3.6B.

In terms of software system, the industrial computer acts as an integrated platform where the relevant drivers of sensing devices and software applications are installed to make the connected sensor devices function normally. To manage the smart sensors in a PnP fashion, the essential procedures are to register the sensors. Thus, a prototype system is designed to establish a bridge for registering and managing the smart sensors. When a new sensor connects to RMMISS, the system are called to register and manage the sensors, the main steps are described as follows: