RFID assisted Flexible Manufacturing System H.M.D.B.Herath University of Moratuwa, Sri Lanka dbhdinesh@gmail.com R.S.V.Piyasena University of Moratuwa Sri Lanka svindika14@gmail.com P.P.G.C.Prasanna University of Moratuwa, Sri Lanka prasannachanuka.uom@gmail.com Y.W.R.Amarasinghe University of Moratuwa, Sri Lanka ranamajp@gmail.com Dzung Dao Griffith School of Engineering Member, Queensland Micro & Nano Technology Centre, Griffith University, Australia d.dao@griffith.edu.au George Mann Faculty of Engineering and Applied Science, Memorial University of Newfoundland, Canada gmann@mun.ca Abstract A novel Flexible Manufacturing System (FMS) with customer-order based production is introduced using web-services technology and a real-time updatable inventory database to enhance operational efficiency of the manufacturing environment. The proposed FMS consists of intelligent control systems integrated with RFID technology, a smart conveyor system, robot arms and sorting mechanisms and a real-time updatable inventory database with application software. This study demonstrates the significance and benefits of a customer-order based production using smart conveyor system with the integration of RFID technology for product identification and handling, specifically in the manufacturing industry. Keywords- Radio Frequency Identification (RFID), Smart Conveyor System (SCS), Flexible Manufacturing System (FMS) I. INTRODUCTION With the increasing of global population, manufacturing industries seek for methods to increase the productivity to tally with the ever rising demand. Pioneers in the manufacturing industry invest billions in developing automated systems to replace the traditional manufacturing systems which are not capable of meeting the global demand. Due to the busy life styles of modern society, people seek for much easier methods to fulfill their requirements. Therefore the online purchasing of items is available for a large variety of products, which has created more business opportunities and satisfied customers. The customer-order based production could be effectively introduced to the manufacturing industry. By the adoption of the customer-order based production systems, not only the enduser but also the manufacturer is benefited. The automated transportation of raw materials, finished/unfinished products etc. between two manufacturing cells is a common process on a factory floor. Different methods are used to achieve this basic function such as conveyor systems, automated guided ve Unlike traditional conveyor systems, a smart conveyor has the ability to identify its loadings and adjust speed and direction according to the real-time field data in the manufacturing environment. Smart conveyors equipped with sensors, wireless devices and intelligent logics, are able to achieve one or more functions: sense (detect), identify, interact, decide and act. Therefore a smart conveyor has an intelligent ability for selfcontrolling. An advanced product identification method such as Radio Frequency identification (RFID) technology could be effectively used to advance the sensing method of smart conveyor systems and to overcome the difficulties faced in the existing optical systems. This will facilitate automatic data capturing and processing of real-time data, excluding the error prone manual activities [1]. Although many research projects have been carried out related to RFID technology and its applications in wireless communication, only a few have been carried out related to product handling in manufacturing environments. With the rapid development of RFID technology, adapting RFID as a supply chain tool promised more efficient supply chain through fast tracking of items. However RFID could be effectively used within a manufacturing environment and provide better results with less implementation costs since the difference between a sale or no-sale simply depends on the maintaining of the right inventory mix. Previous surveys indicate that a major challenge the industries face is training and educating their employees in RFID technology [2]. This study is specially focused on the development of a flexible manufacturing system which adopts customer-order based production, RFID technology for product sensing, automated product handling and real-time updatable databases as in Fig.1. Figure 1. Flexible Manufacturing System 978-1-4673-5221-5/13/$31.00 2013 IEEE 487
In the integration of RFID technology to manufacturing industries, web technologies could be used effectively to manage the services to enable them to be easily registered, published, searched and invoked by users or the third systems. -key -based ordering and distribution systems to satisfy customer requirements. RFID technology could be effectively adapted for such industries for efficient sorting and real-time product traceability which benefits the industry itself as well as the customers. Walmartthe largest retailer in the world adapted RFID technology in a pilot project in 2004 and continues to use RFID as a supply chain tool [3], [4]. James Jungbae Roh et al. [5] proposed a classification of RFID for the type of adaption and the association of the benefits with the scale and scope of adaption. RFID technology has a wide range of applications in logistics [6], airlines industry [7], construction [8] etc. Recent studies have been carried out in the analysis of safety and benefits in the Agri-food supply chain [9]. Automatic identification and traceability have been identified as key factors for efficient manufacturing shop floors. Some concept cases which used RFID technology to capture real-time manufacturing data could be seen in production line [10] and work-in-progress (WIP) inventories management [11]. Wong et al. [12] adopt an authentication method to apparel products using RFID passive tags and lightweight cryptography. Yingfeng Zhang et al. [13] adopt an agent based gateway system for real-time ubiquitous manufacturing. In recent past several studies have been carried out [14], [15] to discover how RFID technology could be used to enhance various aspects of the healthcare industry. II. SYSTEM DEVELOPMENT A. Manufacturing process analysis Prior to the design and development of the FMS, the process of the manufacturing system should be analyzed. We considered manufacturing industries with production lines where products/components need to be sorted according to the specifications/requirements of the manufacturer. Also manufacturing industries where sorting of products is required in relevance to the customer/area of distribution, were taken into consideration. RFID technology could be adapted to such systems (ex: assembly lines) effectively to improve efficiency in the sorting and product handling processes. Therefore a survey was carried out in relevance to electronic component manufacturing industries, food industries and apparel industries in assessing and understanding the basic processes, production assembly tasks, operational environments, current drawbacks of the systems etc. B. Requirement analysis A thorough requirement analysis should be undertaken before implementing RFID technology for a manufacturing system. The requirements of the stake holders would vary with the type of the industry. From the survey carried out, the generalized requirements of the stake-holders were identified as in Table 1. TABLE I. Stake-holder Customer Managers Production line operators REQUIREMENTS OF THE STAKE-HOLDERS Requirements to view the product categories, the availability of the products, prices etc. current sales, the current stock levels, availability of components, real-time traceability of products in the shop floor, to provide sales and inventory reports view the availability of stocks, view the customer orders, ease of sorting, ease of product identification and handling, real-time traceability of products in the shop floor Stock keeper real-time visibility of the overview of stock levels, to view customer requirements, reminder and alert to maintain the stocks above the slack levels, maintain supplier records The major drawbacks of the existing methods were identified after analyzing the manufacturing industries and the production assembly tasks. Drawbacks due to the existing product identification methods (Barcodes/IR) such as unreliability, close proximity range and line of sight requirement, etc. Manual inspection to double-check for errors Time consuming manual data entry to update stock records Production time loss due to inefficient product handling methods Co-ordination difficulties within the shop-floor due to stand alone semi-automated systems After analyzing the general requirements of the stakeholders, a RFID assisted flexible manufacturing system is proposed for customer based production assembly tasks with the integration of a Smart Conveyor System and a robot hand for product handling. The proposed system comprises of RFID sensing, smart conveyors, robot hand and real-time updatable inventory database. Figure 2. Architectural Framework 978-1-4673-5221-5/13/$31.00 2013 IEEE 488
C. Architectural Framework A system architecture framework is developed once the overall system and sub systems are identified. This provides a better understanding of the overall system configuration. The detailed design of the system is carried out with the functionalities and communications between components, after defining the framework. 1) Sensing & product identification layer- The first layer is a data capturing front-end system with IR sensors and RFID transponders and transducers. IR sensors detect the presence and positions of the products within the manufacturing floor. Products are tagged with RFID transponders which consist of unique identification numbers. These identification numbers correspond to the product type and information about the product such as its price and manufacturing details. RFID readers are placed according to the manufacturing floor plan. The embedded middleware system in the RFID reader assists in the filtering, aggregation, and routing of RFID data. 2) Processing and decision making layer- The captured data are classified and stored on databases which are specifically responsible for storing the manufacturing date and time as well as monitoring and controlling the process of sorting of products respectively. According to the identified product type and details, decisions are made to sort the products and move to relevant locations. This layer is responsible for the speed & position controlling signals are sent to the robot arm and the main conveyor. 3) Work-flow layer- The work-flow layer is used to coordinate, manage, and integrate the processes within the manufacturing system. The integration and coordination of the smart conveyor system, RFID and other sensing systems and robot arm is the main function of this layer, which includes guiding the products into relevant conveyors using the pick and place robot arm and controlling and synchronizing of conveyor speeds to match with the assembly processes. 4) Application layer- The application layer provides the necessary application interfaces for data exchange between the databases. It consists of the main server, inventory databases and graphical user interfaces. The main server integrates all the communication interfaces and is responsible for the connectivity of the smart conveyors and the robot arm. 5) System functionality- The FMS introduced consists of two main parts, hardware and software. The hardware includes a web and application server, a sensing and product identification system using RFID technology and Infra-Red, smart conveyor system and a robot hand. In terms of software, a web based application software, Relational Database Management System (RDBSM) and embedded control software for the controlling of smart hardware are introduced. A system overview is shown in Fig. 3. In the proposed FMS, the customer is able to place the order for the desired quantity of products via the web based application software. Then through the main server, application software checks for the availability of components required for production using the inventory database. The integration of all the communication interfaces, connectivity of the smart conveyors and the robot arm are the responsibilities of the main server. Once the order is placed through the web-based application, the components referring to the specific products requested are moved to the main conveyor from the inventory to be sorted for assembly lines. Each subcomponent of the inventory is tagged with a passive RFID transponder which enables identifying of Figure 3. Overall System 978-1-4673-5221-5/13/$31.00 2013 IEEE 489
the objects and sorting. As the RFID reader identifies the objects on the main conveyor, the inventory database is automatically updated to display the current amount of subcomponents available in the stocks. The main function of the robot arm is the sorting of subcomponents and placing them on the sub conveyors. The RFID tagged components are identified from the unique identification number and sorted accordingly. III. DESIGN AND DEVELOPMENT OF PROTOTYPE To validate the applicability of the proposed flexible manufacturing system, a prototype is developed. The prototype developed consists of four main sub systems: Smart conveyor system, RFID identification system, robot hand, real-time updatable inventory database and web based clients. We considered a hypothetical manufacturing system with assembly lines and developed a web interface where the manufacturer has the capability of defining the number of products, subcomponents and their requirement for each product. A. Smart conveyor system The prototype system consists of one main conveyor and two sub conveyors with the main conveyor directly linked to the RFID system. The RFID reader identifies the product type from the unique code embedded in the tag and sends signals to the control system through the main server to guide the product to the relevant conveyor line using the robot hand. Simultaneously through the control system, speed of the main conveyor is adjusted to synchronize with the robot hand for the precise pick and placing of products. Product locations are identified along the conveyor belts using IR sensors to optimize the controlling of the conveyor speeds. The main conveyor adopts, where individual speed and direction control zones are available along the conveyor length. This enables the conveyor to adjust its speed and/or direction using the feedback gained from external factors. If a certain product item is sensed from the IR sensor but does not get identified by the RFID reader, the independent zone control concept could be effectively used to remove the item from the conveyor belt. B. Radio Frequency Identification A RFID system consists of three main components; RFID transponders (tags), Transducer (reader) and RFID middleware as shown in Fig. 4. Low frequency range RFID reader and passive tags were used for the prototype system. Low frequency range was preferred since the prototype requires close proximity ranges for product identification. As the whole prototype system spans in a small area, a high frequency reader would cause many problems in the product identification process such as detecting products which have been already detected, detection of products which have not been placed on the main conveyor etc. Passive tags were preferred over active tags since the prototype system does not require re-writing of data into the tags. Since each and every product/component is tagged with a RFID tag which contains a unique product code, the cost effective solution would be passive tags. The adapted RFID transponders (tags) are compatible with EM4100 protocol, where each transponder carries 64 bits of Read Only memory (ROM). For the product identification, a Low Frequency transceiver (reader) is used with an operating frequency of 125 khz which communicates with the main server through a USB interface. The reader supports EM4100 and EM4001 protocols which are common data formats for RFID transponders. It has a reading distance of 5-8cm and reads the first ten digits of the RFID transponder with a communication speed of 106Kbits/s. A java client application is developed for the communication between the RFID reader and the main computer through the Human Interface Device (HID). C. Robot Hand The basic function of the robot hand is to pick components from the main conveyor and place on the sub conveyors according to the received signals from the RFID reader. Since the prototype was developed to validate the proposed system, a scaled down pick and place robot with 3-DOF and an end effecter was designed. The designed 3-DOF robot hand comprises of two linked-arms for the horizontal movement and a telescopic mechanism for the vertical movement. The robot manipulator movements were obtained using two servo motors and a DC motor while the end effecter used a mini servo for the gripping. For main components of the robot, Perspex was used. The Free Body Diagram of the robot hand and a 2D projected view of all the reachable locations of the end effecter are shown in Fig.5. D. Real-time updatable inventory database and web-based Clients A Relational Database Management System (RDBMS) is introduced in the prototype system for the visibility of real-time updatable inventory. Through the web based interface, customers are given the opportunity to view the real-time 95mm 210mm Figure 4.RFID system Figure 5. (a) Free Body Diagram (b) 2D-view of robot workspace 978-1-4673-5221-5/13/$31.00 2013 IEEE 490
status of the inventory stocks. This enables the customers to place orders depending on the availability of the stocks. The database server consists of specific databases such as product information, RFID tag information, availability and date/time corresponding to the products released from inventory. Also this sub-system provides a function to assist controlling and monitoring of the inventory stocks. If the stocks are below a pre-defined level, a warning is displayed in the user interface provided to the administration. This helps to prevent stocks from running out and maintain safety stocks. For the communication between the RFID reader and the main computer a Java client application is developed to write data through the Human Interface Device (HID) to the Relational Database Management System (RDBMS). Once the order is placed via web-client, the php server is notified through http POST message protocol. The business logic is implemented in a php application. This application processes the http requests and the desired business logic is invoked and the database system is updated accordingly with the current status of the inventory stocks. The client application can request this data through the php server and display them. When the RFID reader reads a value, the embedded system activates the control logic for the controlling of the robot arm and the conveyor belts and sends a message to the php server. This message is again processed by the php application and updates the database system. The communication methodology within the main system is shown in Fig 6. Two separate web interfaces are developed for the administrator and the customer. The administrator or the person Figure 6. Communication methodology Figure 7. RFID system handling the inventory is provided with a web interface, where the admin has the privileges to update the inventory database. The web interface is developed in such a manner that the admin has the following features. Define the available product types Define and enter the amount of sub-components required for each product type Update the inventory by adding sub-components with their corresponding RFID category The real-time updatable inventory database provides a function to assist controlling and monitoring of the inventory stocks. If the stocks are below a pre-defined level, a warning is displayed in the user interface provided to the administration. This helps to prevent stocks from running out and maintain safety stocks. The customer is provided with an interface where he/she is ry. Therefore the customer is able to view the maximum number of product available for purchasing from each of the product category. This maximum number of products available from each type is calculated using the information entered by the admin, which depicts the number of components required for manufacturing of a certain product type and the availability of the sub-components in the inventory. orders through the web-interface provided for the admin. Then the components required for the products ordered are placed on the conveyor. Admin is also provided with an interface which shows the amount of components in the inventory after an order is placed. But the inventory database is only updated once the components are identified by the RFID reader on the main conveyor belt of the smart conveyor system. (The inventory database is updated automatically.) IV. RESULTS AND DISCUSSION Adopting Radio Frequency Identification technology for a manufacturing industry promotes Flexible Manufacturing System (FMS). This enables a paperless working environment and real-time traceability of products within the shop-floor which improves the operational efficiency and reliability. In this paper a FMS is introduced with a smart conveyor system which adopts RFID technology for product identification and handling. In the system introduced, RFID technology assists a smart conveyor system, a 3-DOF robot hand and a real time updatable database with web based applications. The main contributions of the proposed RFID assisted FMS are identified as follows; Real-time updatable inventory with stock records An accurate real-time updatable inventory is an important factor in managing a Just-in-time (JIT) production module. The RDBSM introduced in this prototype system provides up-todate stock information on the components available in the inventory. This provides real-time visibility of the overview of 978-1-4673-5221-5/13/$31.00 2013 IEEE 491
stock levels which in turns reduces the possibility of running out-of-stock. On-demand manufacturing process Since the production is carried out according to the customer requirement, the manufacturer has fewer risks in the sales process. Tracking the real-time consumption The automated data capture using RFID and the real-time updatable inventory, enables tracking of real-time consumption and removes the human factor in track-and-trace operations. This improves the efficiency of the manufacturing environment. Overcome drawbacks in existing product identification methods Many difficulties faced in barcode systems and other object identifying methods such as the need for proper alignment, closer proximity reading etc. are overcome by the implementation of RFID technology. Furthermore, large data could be embedded in to the tag and could be placed inside the product to avoid damages. Demand forecasting Using the statistics available in the system, the administration is able to forecast the demand for future production. The main challenges in implementing RFID technology are identified as follows; Cost challenges- a major challenge faced by the industries when implementing RFID technology is the justification of the investment. When adapting RFID technology, the major costs includes hardware, middleware, application software, consultancy charges, employee training etc. Lack of expertise- In Sri Lankan context the industries are not eager in adapting RFID technology, mainly due to the lack of expertise in RFID technology. Training and educating the employees in RFID technology is another challenge faced by the local industry. In-adequate technical support- Since the locally available RFID systems (both hardware and software) are from foreign vendors, limited technical support is available. Recycling- Electronic waste due to discarded tags Though RFID technology, conveyor systems, robot hands and database management systems are already being used in the industry as individual systems, the overall system proposed in this paper which synchronizes the RFID technology, the SCS, the robot arm and the inventory system with web based applications is a novel concept of advanced manufacturing technology (AMT). One of the main contributions from this project is the smart conveyor system linked with the robot arm, where the robot arm adjusts its movements using the feedback gained from the smart introduced through the intelligent conveyor system equipped with RFID technology and optical sensors. This enables controlling of particular areas of the same conveyor along its length. Large retail companies have successfully adopted RFID technology in the supply chain using GSM devices. Similarly, for a manufacturing environment, handheld smart devices and GSM devices which can communicate with mobiles, etc. could be introduced as further improvements. Real-time WIP inventory monitoring and production exceptions alert systems could be developed as well. In this paper passive RFID transponders were used for object identification. Although the overall system could be developed at a higher level by the adoption of active RFID tags, the implementation cost would play a major role when used in a larger scale. Figure 8. Customer Interface Figure 9. Prototype system V. REFERENCES [1] H. Jun, J. Shin, Y. Kim, D. Kiritsis and P. Xirouchakis, "A framework for RFID applications in product lifecycle management," International Journal of Computer Integrated Manufacturing, vol. 22, no. 7, p. 595 615, 2009. [2] J.Morrison, "Help Wanted," RFID Journal, pp. 13-20, March-April 2005. [3] M. Roberti, "Wall-Mart begins RFID rollout," RFID Journal, 2004. [4] M. Roberti, "Wal-Mart Relaunches EPC RFID Effort, Starting with Men's Jeans and Basics," RFID Journal, 2010. [5] J. J. Roh, A. Kunnathur and M. Tarafdar, "Classification of RFID adoption: An expected benefits approach," Information & Management, vol. 46, no. 6, pp. 357-363, August 2009. [6] E. Nagai, T. Cheng and K. Lai, "Mobile Commerce integraed with RFID technology in a container depot.," Decision Support Systems, vol. 43, pp. 62-76, 2007. [7] D. Wyld, M. Jones and J. Totten, "Where is my suitcase? RFID and airline customer service," Marketing Intelligence & Planning, vol. 23, no. 4, pp. 382-394, 2005. 978-1-4673-5221-5/13/$31.00 2013 IEEE 492
[8] J. Song, C. Hass, C. Caldas, E. Ergen and B. Akinci, "Automating the task of tracking the delivery and receipt of fabricated pipe spools in industrial projects," Automation in Construction, vol. 15, no. 2, pp. 166-177, 2006. [9] M. Zhang and P. Li, "RFID Application Strategy in Agri-Food Supply Chain Based on Safety and Benefit Analysis," Physics Procedia, vol. 25, pp. 636-642, 2012. [10] G. Huang, Y. Zhang and S. Newman, "RFID-enabled real-time wireless manufacturing for a daptive assembly planning and control," International Journal of Intelligent Manufacturing, vol. 19, no. 6, pp. 701-713, 2008. [11] G. Huang, Y. Zhang and P. Jiang, "RFID-based wireless manufacturing for real-time management of job shop WIP inventories," International Journal of Advanced Manufacturing Technology, vol. 36, pp. 752-764, 2008. [12] H. Wong, C. Hui and C. Chan, "Cryptography and authenticationon RFID passive tags for apparel products," Computers in Industry, vol. 57, no. 4, pp. 342-349, 2006. [13] Y. Zhang, Q. George, T. Qu, O. Ho and S. Sun, "Agent-based smart objects management system for real-time ubiquitous manufacturing," Robotics and Computer-Integrated Manufacturing, vol. 27, pp. 538-549, July 2010. [14] W. Yao, C. Chu and Z. Li, "Leveraging complex event processing for smart hospitals using RFID," Journal of Netwok and Computer Applications, vol. 34, no. 3, pp. 799-810, May 2011. [15] Y. Meiller, S. Bureau, W. Zhou and S. Piramuthu, "Adaptive knowledge-based system for health care applications with RFIDgenerated information," Decision Support Systems, vol. 51, no. 1, pp. 198-207, April 2011. 978-1-4673-5221-5/13/$31.00 2013 IEEE 493