ARTICLE IN PRESS

Int. J. Production Economics 112 (2008) 630645 www.elsevier.com/locate/ijpe

Development of an RFID-based sushi management system: The case of a conveyor-belt sushi restaurant
E.W.T. Ngai, F.F.C. Suk, S.Y.Y. Lo
Department of Management and Marketing, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong Received 28 February 2006; accepted 2 February 2007 Available online 16 June 2007

Abstract We describe the design and development of a radio-frequency identication (RFID)-based sushi management (RFSM) system in a conveyor-belt sushi restaurant to enhance operational efciency. The system is designed to help a conveyor-belt sushi restaurant to achieve better inventory control, responsive replenishment, and food safety control, as well as to improve its quality of service. This study demonstrates the signicance and benets of using RFID technology specically in the food industry. The lessons learned from this effort help to support and further the academic and practitioner literature, especially in the area of RFID systems development. Finally, we address the limitations of this study and several areas of future development. r 2007 Elsevier B.V. All rights reserved.
Keywords: Radio-frequency identication (RFID); Sushi management system; Food safety; Food industry

1. Introduction Radio-frequency identication (RFID) technology has been widely used in many industries, including the airline industry (Wyld et al., 2005; OConnor, 2006), cattle industry (Mennecke and Townsend, 2005), construction (Jaseiskis and Ei-Misalami, 2003; Song et al., 2006), logistics (Ngai et al., 2007), healthcare (Collins, 2005), and manufacturing (Swedberg, 2006). In this study, we focus on the use of RFID in the food industry. We illustrate the signicance and benets of an RFIDbased sushi management (RFSM) system in a

Corresponding author. Tel.: +852 3400 3584; fax: +852 2765 0611. E-mail address: msfred@inet.polyu.edu.hk (F.F.C. Suk).

conveyor-belt sushi restaurant to enhance operational efciency and food safety. Diners are conscious of food safety, as are food operators. Poor food management can result in a higher potential for parasitic or bacterial contaminants to develop, or even for food poisoning to occur. Food safety monitoring and control is particularly crucial for a Japanese cuisine like sushi. Sushi is a food made from vinegared rice combined with various toppings or llings such as seafood, vegetables, mushrooms, eggs, and meat. This kind of food is renowned for its fresh taste; eating it raw is common. If sushi with raw toppings is exposed for too long at room temperature, its food content is likely to become contaminated and its quality impaired. It is important for sushi restaurants to have an effective food control process to assure that the sushi is safe for consumption. Thus, in this

0925-5273/$ - see front matter r 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpe.2007.05.011

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 631

paper, we make use of the RFID technology with the aim of examining ways to strengthen food safety control in the food industry, particularly in sushi restaurants. In addition, we demonstrate the use of RFID technology to improve inventory control and responsive replenishment in the industry. In a sushi restaurant, the display of sushi on a conveyor belt is becoming more popular. The sushi restaurant features an elliptical counter around which diners can sit and take the food. A conveyor belt revolves around the perimeter of the counter, displaying a colourful variety of sushi, usually two or three pieces per plate. As the sushi plates pass by, diners simply reach out and select their favourites. This provides a cheap and convenient way of eating sushi. However, traditional conveyor-belt sushi restaurants are facing several operational and management challenges: (i) They are time-consuming and there is a potential for human error in billing Traditionally, sushi restaurants have been used to adopting a plate-colour pricing scheme. In this scheme, the price category of a variety of sushi food is signied and coded with plates of differing colours. When diners nish their meals, their bills are calculated according to the price category of each plate colour multiplied by the number of plates of each colour they have taken. This kind of billing calculation is done manually. It is a time-consuming process and human errors sometimes occur in totalling the bills, causing a loss to either a diner or to the restaurant. (ii) There is a potential food hazard Freshness is an important food quality for sushi. To preserve this quality, each sushi item should be stamped with a label to indicate the time when it was prepared. After the expiry date, sushi is supposed to be not suitable for eating because it might be bacterially contaminated and subsequently pose health risks to a diner. A sushi chef is thus used to removing the expired food from the rotating belt regularly and manually. However, it is not easy for the chef to identify and monitor the expired sushi on the conveyor belt. (iii) Sushi stock control on the belt can be ineffective A typical sushi restaurant normally offers more than 30 kinds of sushi on its menu. Diners may frequently not be able to nd all kinds of sushi on the conveyor belt. Though diners might

attempt to specially order a particular type of sushi, they may sometimes be told that the item is out of stock or sold out. This not only frustrates the appetite of diners, but also impacts on the quality of service. Sushi inventory monitoring and control on the belt is crucial to attract new and repeat diners. Both understocking and overstocking have implications for the efciency of the inventory management. Currently, it seems that no system is available to effectively manage the sushi stock inventory on the belt. In addition, it is hard for sushi restaurants to analyse and identify the popularity of the sushi sold. The current practice of sushi restaurants for learning about the popularity of sushi items is based on the chefs experience or perceptions. The data visibility of the stock inventory is poor. From these operational and management challenges, research questions arise. Can information technology be used to help manage the challenges? What are the benets of using information technology? How can the billing process be streamlined and its efciency and accuracy enhanced? Can the removal of the expired food be done precisely and automatically, in order to strengthen food safety? How can the food inventory control be improved in order to achieve a more responsive replenishment? With these questions in mind, we propose using RFID technology as a solution for a conveyor-belt sushi restaurant. Sushi food is well known and popular worldwide, not only in Japan. Many sushi restaurants are operating in the US, Europe, and Asian countries. The market size of sushi restaurants has been growing rapidly in most Asian countries. RFID technology can be effectively applied in sushi restaurants to help improve their food safety, inventory control, service quality, operational efciency, and data visibility. The purpose of this study was to examine how the adoption of RFID technology could help to improve the operational efciency of a conveyor-belt sushi restaurant. We develop a framework and present a case study to illustrate an RFSM system for a sushi restaurant. Specically, the objectives of building the proposed system were to:

simplify the billing process by using a personal digital assistant (PDA) device for counting food plates and calculating the billing amount to enhance billing efciency and accuracy;

ARTICLE IN PRESS
632 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

 

provide tracking information about food on the belt to enable the more effective removal of the expired items in order to strengthen the control of food safety; improve the food inventory control on the belt by responsive replenishment; and provide visible operations statistics for inventory management and marketing promotion.

The rest of the paper is organized as follows. In Section 2, we present a brief overview of previous RFID applications. Then in Section 3, we describe our proposed framework of an RFSM system. Following this, we address the limitations of the system and propose future improvement work. Finally, a conclusion is drawn. 2. Previous RFID applications in the food industry Kumar and Budin (2006) indicated that the US Food and Drug Administration reported a total of 1307 processed food product recalls between 1999 and 2003. In Japan, a massive recall was also reported from two Japanese food makers in 2002 (Jiji Press, 2003). Demand for higher food hygiene and safety is becoming a growing concern (Kerry et al., 2006). RFID technology has already been adopted for use in the food supply chain and to improve the traceability of food (Hutter, 2004). In food supply chain management, a food operator can use RFID technology to record and provide information on all stages from supplier, transportation, and production, to storage and distribution of an individual food item. Food traceability refers to the ability to trace and follow a food, feed, food-producing animal, or ingredients, throughout the production and distribution process (Regattieri et al., 2007). With effect from January 2005, the European Food Law requires food operators to identify their suppliers and customers, and to provide this information to the authorities concerned (EEC, 2002). Use of RFID technology has been reported to be a solution to enable such tracing capability in the supply chain of the food industry (Kumar and Budin, 2006; McMeekin et al., 2006). The RFID tags attached to individual food products allow suppliers to trace every movement of their food products throughout the production, storage, transportation, and point-of-sales processes. Karkkainen (2003) conducted a trial of the use of RFID tagging for short-shelf-life goods in 1998, demonstrating

that the adoption of RFID technology can achieve the benets of improving replenishment productivity and reducing stock loss in the supply chain of short-shelf-life products. Jones et al. (2005) reported that RFID technology can offer a wide range of benets to food retailing in the UK throughout the supply chain, including tighter management and control of the supply chain, reductions in shrinkage, reduced labour costs, and improved customer service, facilitating compliance with traceability protocols and food safety regulations. Kumar and Budin (2006) also concluded that RFID plays a pivotal role in the prevention of food product recalls. Specically, for the meat-processing industry, Mousavi et al. (2002) proposed a solution using RFID technology to track meat products and provide information about them which is attached to them throughout the production process until they become retail packs. They indicated that RFID technology offers a number of potential benets to the meat production, distribution, and retail chain, including improving food quality control, safety, traceability, inventory management, and laboursaving costs. In addition, in the agricultural food industry, Thysen (2000) indicated that IT had become important to the industry as a means of achieving better food control and safety, as well as higher precision in the use of chemicals and in the care of farm animals. Wang et al. (2006) presented an overview of the development of wireless sensor technologies and standards for wireless communications in agriculture and the food industry. These kinds of wireless sensor technologies are used for environmental monitoring, precision agriculture, M2M-based machines and process control, building and facility automation, and RFID-based traceability. They predicted that, given the advantages of RFID technology over the traditional barcode method, the development and deployment of RFID in food traceability would increase signicantly in the near future. In addition, Jedermann et al. (2006) proposed a system using a combination of RFID, sensor networks, and software agents to trace fruit transports, demonstrating an effective use of RFID technology in fruit logistics. In the early phase of RFID technology, its limitations related to its high cost and the unlikelihood of a pay-off of the investment (Burnell, 1999; Riso, 2001). However, according to Jones et al. (2005), the price of an RFID tag was about $1 in 2000, had fallen to $0.250.35 by early 2004, and is

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 633

expected to drop to around $0.05 as RFID technology becomes more widely adopted. Furthermore, the capability of RFID technology has been criticized as being too similar to that of the barcode. Burnell (1999) inferred that most of the functionality needed had already been achieved by barcode technology. Karkkainen (2003) pointed out the limitations of barcode data collection, including the occasional necessity to read barcodes manually and poor barcode readability in some environments. Despite the potential limitations perceived early on in the development of RFID technology, several recent studies have indicated that investing in RFID technology is promising and an excellent long-term capital investment (Karkkainen, 2003; Kumar and Budin, 2006; Regattieri et al., 2007). 3. Architectural framework of an RFSM system In this section, we demonstrate the architectural framework of an RFSM system. First, we give an architectural overview of the system in terms of ve cross-sectional layers of the system presented. Then, we describe six stages in the design and development of the proposed system, including the business process analysis, the requirements analysis and the

RF site survey, system architecture, system design, implementation, and testing and evaluation. 3.1. Architectural overview We designed and developed an architectural framework of an RFSM system using RFID technology. Fig. 1 shows an overview of the architectural framework of the system. It comprises ve layers in total. (i) Data-capturing front-end system The rst layer is a data-capturing front-end system. It has three components: a transponder, a reader, and an antenna. The transponder consists of RFID tags located on the sushi plates. Each tag contains information about the sushi, such as its price, type, date and time of production, calories, and so on. A reader is equipped with an antenna and is used to read or write information onto the RFID tag. When the sushi plates are on the belt and pass by the readers, tag information is retrieved to determine the freshness of the food items. The current time is compared with the sushi production time stored in the tag. If the time

5. Application Layer

Sushi Management System

ERP

POS

Other Applications

4. Work-Flow Layer

Sushi Quality Management (SQM) Work-Flow Engine

3.Processing Modules

Track &Trace Module

Data Monitoring and Alerting Module

2.Data Capturing Layer

Data Capturing and Transformation Framework

1.Data Capturing Front-end

RFID reader with antenna and tags Read point 1 Read point 2 Read pointx Read pointy

RFID Handheld Reader and Tags

Fig. 1. Architecture of the RFSM system.

ARTICLE IN PRESS
634 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

(ii)

(iii)

(iv)

(v)

difference is greater than a specied interval, the sushi is regarded as having passed its sell-by time, and will then be removed automatically. In addition, during the process of billing the diner, an RFID hand-held reader is used to conduct a quick scan of his or her plates of sushi. A small portable printer can be attached to the device for receipt printing. Data-capturing and ltering layer This is a middleware system that assists with the ltering, aggregation, and routing of RFID data. The data include the timestamp, and the reader and antenna identity. Processing modules The captured data are classied and stored on databases which are specically responsible for storing the billing information, and monitoring and controlling the freshness of the sushi, respectively. Some of the captured data are used to directly drive the workow engines in the back-ofce systems. A sushi quality-management workow engine is developed in the system. Workow layer Workow functionality is located at Layer 4 which is used to coordinate, manage, and integrate the processes and data ows within the new sushi management system and the existing system applications like ERP and POS. The middleware has built-in business rules, which are specied in an ifthen format. It allows the rule engine to reason about the ow of information above and below this layer. For instance, a business rule for monitoring the expiry time of sushi looks like if a sushi dish is due to reach its expiry time in 1 min, then an alert will be issued to the chefs computer and the sushi will be displayed in red on the chefs computer screen. According to the needs of the chef, this rule can be changed by adjusting the time at which the alert regarding the expiry time is issued. Applications layer Above the workow layer, there is an API layer which provides the necessary application interfaces for data exchange between new and existing applications. User applications are in the upper application layer which provides the graphical user interfaces for the users to use the applications; that is, an enterprise resource planning (ERP) system, point-of-sales systems, and so on.

3.2. Architectural framework design and development We used a six-stage structural development method to design and develop the architectural framework of the RFSM system: (1) business process analysis, (2) requirement analysis, (3) system architecture, (4) system design, (5) system implementation, and (6) system testing and evaluation. We describe each stage below. 3.2.1. Stage 1: business process analysis A business process refers to the way a particular business is conducted. It relates to the business rules, strategies, and unique ways in which an organization coordinates work, information, and knowledge, as well as to how management coordinates work (Laudon and Laudon, 1998). RFID is an emerging and innovative technology. RFID can be an enabler in effecting changes and achieving productivity gains using techniques that re-engineer the business process. It can help to fundamentally redesign the business processes of sushi restaurants in order to optimize their benets and their operational efciency. As there is little known literature on the operations and management of a conveyor-belt sushi restaurant, we have adopted various approaches to acquire the businessprocess knowledge of sushi restaurants. Herbst and Karagiannis (2000) indicated that business-process knowledge could typically be acquired by conducting interviews or by using a questionnaire survey. We therefore conducted interviews with staff working in a restaurant, with the aim of observing their operations at the same time. In this study, we visited three sushi restaurants and interviewed 10 staff holding a variety of positions, ranging from manager, chef, waiter, and stock-taker to cashier. These interviews not only enabled us to understand the current processes, operations environment, and weaknesses at the managerial level, but also provided signicant insights into the weak areas at the operational levels. In the typical ow of a conveyor-belt sushi restaurant, a stock-taker makes orders of raw materials from various suppliers by phone or fax. Non-systematic stock taking and distribution of the materials to the restaurant oor are observed. The sushi chefs make the new sushi based on their rough estimation of the number of food items still currently available on the belt. After preparing the new sushi, they place it on plates of the colour

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 635

appropriate to its price and give it a paper label to identify it. The new sushi is then placed onto the conveyor belt and is ready for eating. Having nished their meals, the customers call for a waiter to manually count the number of coloured plates and calculate their bills. Normally, a cashier calculates and audits all the billing items to conrm the diners bill. After our process analysis, we identied a number of core business processes that could be redesigned, in particular, the replenishment process, sushi-tracking process, and billing process. We describe how to improve the efciency of these processes in the requirement analysis stage. 3.2.2. Stage 2: requirements analysis and RF site survey Having acquired the business process knowledge and understood the operational environments of sushi restaurants, we began to analyze their business needs and redesign business processes. In addition, we conducted an RF site survey to help to prepare the facilities for the deployment of an RFID system, including pilot testing of RFID tags, readers, and the antenna. To predict the site performance of our RFID solution, we studied and analyzed potential interferences in the restaurant environments like wireless networks, short-range radios, and mobile phones. Having had discussions with the stakeholders concerned, we analyzed and summarized their key functional requirements for an RFSM system; these are shown in Table 1. After analysis of the business requirements of the stakeholders, an RFID-based conveyor-belt sushi system was proposed. The stock-taker could place an order of raw materials from the suppliers via the system. It simplied and shortened the time required for the replenishment process. The stock-keeper could also use the system to systematically record order information from the suppliers and monitor the available stocks in the shop oor and the refrigerator. The restaurant manager could use the system to view statistics of the current inventory, sales records, staff information, and so on. The chefs could replenish and prepare the right quantities of different kinds of sushi in the light of the quantity already available on the belt as shown on the display screen of the system. In addition, the system provided a reminder and alert to the chefs and helped to automatically remove those food items that had reached their expiry time from the belt. Diners could see which sushi dishes were

Table 1 Stakeholders requirements Stakeholders Customer/ diner Requirements for RFSM

 To provide a visual view of an entire

Shop manager

 To maintain staff records  Maintain sushi menu records  Report different kinds of sushi consumed
and available quantity

 Provide sales and inventory reports

 Record the inventory in and out situations  Provide a reminder for responsive
replenishment

 Maintain supplier records

 Update sushi information onto RFID tag

 Provide a reminder and alert of the expired

 Remove the expired sushi automatically  View the quantity of different kinds of sushi
on the belt Waiter

 Simplify the billing procedure  Provide an automatic tool for bill

Cashier

 Streamline the bill payment process

available on the belt from the display screen so that they were able to tell if their delicacies were on the belt, instead of having to keep mental track of the sushi dishes available on the belt. After they had nished their meals, a waiter could bring a PDA device to them to perform a quick scan of the number of coloured plates, in order to make out their bills. A receipt with the transaction number, sushi names, and their prices and quantities consumed could be printed out. A cashier could simply input the transaction number for the purposes of settlement of the bill, without being required to re-calculate all the items on it. 3.2.3. Stage 3: system architecture In this stage, we performed an architectural development of the system. According to Nunamaker

ARTICLE IN PRESS
636 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

et al. (1990), a system architecture provides a road map for the system-building process by placing components into perspective, dening their functionalities, and demonstrating how they will interact with one another. It is an initial process in the system development process. We rstly identied the subsystems and established a framework for subsystem control and communication. Then, we dened the functionalities of each component and described the operations and communication interfaces between one component and another. A system overview is shown in Fig. 2. It depicts an RFSM system consisting of ve key components: (i) user interface, (ii) database server, (iii) LCD screen display, (iv) plates embedded with an RFID tag, and (v) the RFID reader. All of them are linked up with the Intranet or Internet through a wired or wireless network. We describe each of the key components below. (i) User interface: Computers that could access the Intranet were provided to various stakeholders: one computer each was prepared for the manager, the cashier, the stock-keeper, and

the chef, and two were for controlling the LCD display screen. A portal PDA equipped with an RFID reader and a small portable printer was provided to waiters via a wireless network. (ii) Database server: The database server consisted of various specic databases like HR, sushi information, inventory control, supplier information, transaction management, and RFID information. The databases were accessed through the open database connectivity (ODBC) gateway to execute a query. (iii) LCD screen display: LCD screens were located in the front area of a restaurant to allow diners to know what types of sushi were available on the belt. LCD screens were also placed facing the chefs so that they could see the quantities available on the belt and decide on the supply of new food items. (iv) Plates embedded with RFID tag: A tiny RFID tag was embedded into plates carrying sushi. The tag information could be either read or written to by an RFID reader in order to retrieve and update the food information.

RFID embedded sushi plate

Sushi Information Display

RFID Reader Sushi Maker Update Sushi Information

Administration

Transaction

Cashier

Database Small portable printer RFID Reader Wireless Network

Billing Staff

Conveyor Belt

Stock Keeper Inventory Control

RFID Reader

Fig. 2. Overall architecture of the RFSM system.

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 637

(v) RFID reader: RFID readers were placed at xed positions on the conveyor belt to examine the availability and freshness of the sushi. Another RFID reader was also located on the table for chefs to update the tag information on the sushi-carrying plate. An RFID reader was attached to the portal PDA for billing purposes. We further determined the hardware and software requirements of the system. The recommended quantity and processing power of the required equipment depended on the size of the sushi restaurant. In general, the system was composed of the following components: Hardware requirements

  

Tomcat application server RDBMS database Browser with Flash Player

       

RFID tag RFID reader Web and application server Database server LCD screen PDA Wireless and Ethernet network Software requirements Apache web server

3.2.4. Stage 4: system design System design is an important aspect of system development. It involves an understanding of the domain being studied, the application of various alternatives, and the synthesis and evaluation of a proposed solution (Nunamaker et al., 1990). In this stage, a systematic and high-level system design and analysis of the prototype system was performed. The design of the entire system followed a three-tier structure, namely the clientserverserver model. In such a system, the rst tier is the thin client browser, through which users can access applications via web services. In the middle tier, user requests are validated, controlled, and redirected for further processing. The nal tier encompasses the relevant business components where user transactions are processed and stored. We divided the system into subsystems and components based on functionality. A detailed subsystem design was conducted by means of a high-level modelling language. Typically, we used a Unied Modelling Language

Customer Maintain Staff Record HR Database

Manager Sushi Maker Menu Management Sushi Database Display System

Transaction Cashier

Statistical Record Inventory Database

Maintain sushi production record

Sushi-on-the-belt Database

Generate the receipt Waiter/Waitress

Transaction Database

Stock-in-and-out Record System

Maintain supplier Information Stock-Keeper

Supplier Database

Fig. 3. Use case diagram of the RFSM system.

ARTICLE IN PRESS
638 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

(UML) to model and walkthrough use cases of the system. The deliverables of this stage would act as a blueprint for the system implementation in the next stage. Fig. 3 shows a case diagram of the proposed RFSM system. We describe each of the subsystem functionalities below. (i) Human resources management subsystem This subsystem enables a restaurant manager to maintain staff record information, including personal information, training arrangement, leave application, and performance and appraisal evaluation. (ii) Stock in-and-out record subsystem This subsystem helps keep track of the inventory of all raw materials used in a restaurant. It provides a built-in redyellowgreen trafc light signalling function to assist monitoring and control of the stock. The stock above the threshold level is shown in green; yellow indicates that the stock on the shop oor is nearly used up, and that redistribution from the refrigerator is necessary. Red shows that the total stock of a particular raw material has dropped below the threshold level and that a new order should be placed with the suppliers. Supplier information is stored in the system to facilitate the placing of a new order via the system. (iii) Press-to-make subsystem This subsystem enables chefs to update sushicarrying plate tag information after making new sushi. Before the chefs put the plate on the belt, its timestamp is updated via the system so that effective tracking and monitoring of the expired food can be achieved. In addition, stock information on the available sushi on the belt is updated and shown on the LCD display screen to let the diners know that new sushi is available on the belt. (iv) Billing subsystem This subsystem helps a waiter to simplify the billing process. The waiter carries a PDA device with an RFID reader and a small portable printer. Using the PDA device, a quick scan of the coloured sushi plates is performed. A diners bill is then calculated automatically by the system. A receipt with the billing number, sushi names, and their prices and quantities consumed can be printed using the portable printer. In addition, the billing data

are sent through the wireless network to the system databases. A cashier can use the billing data to accurately settle a diners bill. 3.2.5. Stage 5: system implementation Implementing a system is an engineering concept (Scott-Morton, 1984). Prototyping is a quick way of demonstrating a solution to a problem. It also helps to uncover any potential inherent problems. In this implementation stage, we developed the RFSM system according to our architectural and system design. HTML, JSP and Javascript, and Macromedia Flash programming languages were selected for the development of web-based interfaces, as they were portal and compatible with most web browsers. In addition, we used Java and Java servlets for developing modules in the middle tier such as session tracking and ow redirecting, because most of the current application servers like Tomcat widely support these kinds of Java components. Finally, a structured query language (SQL) was used for writing statements and queries in the rational database management (RDBMS) system. The open database connectivity (ODBC) protocol was selected to communicate between the database server and the application server. Both SQL and ODBC are compatible with a variety of RDBMS databases. 3.2.6. Stage 6: system testing and evaluation Having developed the system prototype, we tested and evaluated it thoroughly in this phase to see if it worked correctly and met the requirements. We performed a set of formal tests on the system, including module testing, integration testing, functionality testing, and performance testing, to insure the prototype was running free from bugs and errors. In addition, we further examined the prototype by inviting domain experts to evaluate it. Gasching et al. (1983) indicated that domain expert evaluation is effective to help determine the accuracy of the embedded knowledge in the prototype. We employed a user questionnaire to document responses to system testing. In order to obtain comprehensive feedback, 43 university students studying for parttime degrees in hotel and catering management as well as tourism management (most of them worked in food and catering-related industries in the daytime) were invited to participate in the evaluation. At the evaluation session, the system prototype was demonstrated and feedback was solicited from

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 639

the students through discussion. The results of the prototype evaluation were satisfactory and supported the contention that the RFSM performed its functions as expected. Five sushi chefs and managers were also invited to evaluate the prototype system. A formal questionnaire containing both closed and open-ended questions was designed which consisted of two sections: (i) effectiveness of the prototype system and (ii) usability of the prototype system. The potential users were asked to use a 5-point scale (1 strongly disagree, 3 undecided, and 5 strongly agree) to rate the two main aspects of the prototype system: its effectiveness and its usability. The results of the questionnaire analysis show that the potential users and sushi experts rated the system highly on the above two aspects with a mean score of least 3.5 on a 5-point scale. Based on the results of the evaluation, the prototype is seen to be a promising system for improving the operational efciency of a conveyor-belt sushi restaurant. They were positive about the potential of the system, indicating in particular that it was helpful in speeding up the calculation of the bill and improving the quality control of the sushi. 4. RFID benets and challenges: lessons from the case study The benets and challenges of the development of the RFSM prototype system were examined from different perspectives: technological, methodological, managerial, and applications. Some insights obtained from involvement in developing the RFSM prototype system are summarized in Table 2. The benets of the RFSM system in a sushi restaurant can be described as follows: (i) Keeping the stock records of the raw materials for making sushi electronically This provides greater visibility of the up-to-date stock information on the raw materials for making sushi. This system enables a real-time overview of stock levels and reduces the possibility of becoming out-of-stock. Inventory accuracy is important to help improve visibility and insure the right materials are available, and to better manage the just-in-time production model. (ii) Tracking the real-time consumption and availability of sushi on the belt Better visibility of the inventory and automated data capture can remove the human factor in

Table 2 RFID benets and challenges RFID benets Technological perspective  Proven technology adopted in food industry  Responsive multiple tag real-time read/write capabilities  Long read range without specied sight requirement  Normal functioning even in a restaurant with dimmed lighting or harsh environment Methodological perspective Improving billing process as well as its efciency and accuracy Managerial perspective  Strengthening food quality and safety control  Enabling responsive replenishment  Improving dining experience and service satisfaction  Reducing out-of-stock situations Application perspective  Tracking the availability of sushi on the belt  Providing useful foodcontent information to diners RFID challenges

 High implementation cost  Incompatibility with

 Non-standardization of
RFID technology

Potential re-engineering and impacts on a rms internal business processes

 Insufciency of strong  
expertise in RFID technology Resistance to technological change Expenses for staff retraining

 Inadequate in-depth
technical and application supports

track-and-trace operations with the sushi on the conveyor belt. This allows the sushi chefs or the manager to instantly query the real-time consumption and availability of food items on the belt. Based on this information, the chefs can decide the quantity and type of sushi they need to make to insure sufcient supply of a variety of food on the belt. (iii) Enabling responsive replenishment Information generated by the system can be analyzed to enable demand forecasting of types of sushi. It can assist chefs or managers to perform effective strategic planning in their daily operations.

ARTICLE IN PRESS
640 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

(iv) Removing the expired sushi from the belt automatically to strengthen food safety control The traceability function of the timestamp on the sushi dish can facilitate access to information concerning sushis freshness; that is, the length of time each sushi item has been on the conveyor belt. It can be assured that expired food is automatically removed by the system and food safety control can thus be strengthened. (v) Simplifying the billing procedure and improving its accuracy Traditionally, staff are used to spending a portion of their time manually counting plates for billing purposes. This process is time consuming and error prone. This system automates the repetitive and tedious billing procedures. It not only improves the efciency and accuracy of the billing procedures, but also frees up skills resources to maximize the value of the human capital in a sushi restaurant. (vi) Providing more transparent sushi information to diners like prices, availability, calorie content, and so on RFID tags can carry useful information along with the food, including calorie content, prices, timestamp, nutritional characteristics, and serving suggestions, all the way to an endconsumer. This food information facilitates the selection by diners of the right sushi for their consumption. It also introduces diners to a high-technology dining experience. Some challenges in the implementation of an RFID-based sushi management system are described below. (i) RFID expertise for deployment A recent survey by the Computing Technology Industry Association revealed that 80 per cent of the responding companies said that there were not sufcient numbers of skilled RFID professionals. Two-thirds of them said training their employees in RFID technology and educating them about it was one of the biggest challenges they faced in order to succeed in the RFID market (Morrison, 2005). We believe that the problem is more serious in Asia; like in Hong Kong, where there is a lack of strong expertise in RFID technology deployment. Depending on the scope of the project, the knowledge and expertise base typically covers

RFID readers and installation, RFID tags, antenna design, RFID standards, local regulations, RFI environmental hazards, middleware, database and web management, and so on. (ii) Management commitment Gaining management commitment is one of the biggest challenges in implementing RFID. Return on investment (ROI) is an important consideration for management in assessing RFID investment before RFID project commitment. We have to show management that the competitive advantage and the bottom-line tangible and intangible benets that implementing RFID will bring are to its advantage. (iii) Cost challenges The biggest challenge companies face with RFID is the high cost of implementation. One of the challenges a company faces with the introduction of RFID technology is whether the business really needs the technology and how to justify the investment in its implementation. Costbenet analysis is critical to the successful adoption of an RFID project. At present, the costs of RFID adoption comprise the major investment in hardware, application software, middleware, and tags, and the cost of integrating the RFID-based system with the legacy systems, of consultancy fees, and of employee training. (iv) Technical support for adoption At present, both RFID hardware and software in Asian markets come mainly from the US. RFID solutions originate from foreign vendors with the support of local distributors who provide limited technical support as most of their staff are marketing oriented. We have found that most RFID hardware providers in Asian markets lack technical support staff. This is due to the fact that RFID is still an emerging technology in most Asian countries. 5. Lessons learned Some lessons learned from this study can serve as a guideline that future implementers of RFID systems can adapt to their situations. We describe some of the lessons learned below. (i) RF site testing Onsite testing is one of the crucial steps in assessing the readiness of a facility and possible electromagnetic interference (EMI) with RF

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 641

solutions. In our study, before the development of the RFSM system, we conducted a survey to check the preparations on the premises for the deployment of an RFID system, including pilot testing the RFID tags, the readers, and the antenna. In addition, we also carried out an analysis of potential EMI in the restaurant environments like wireless networks, shortrange radios, and mobile phones in order to better predict the site performance of our RFID solution. (ii) Top management support Top management support is one of the key recurring factors affecting the success of implementing an RFID system. Apart from setting the right strategic direction, support from top management can insure that sufcient resources have been allotted for the development of the RFID system. Furthermore, implementation of the new system is more likely to bring about change like the abandonment of outdated processes. Resistance to change is another important issue to confront. Continuous top management support can also play a pivotal role as an agent of change to create an environment conducive to streamlining the RFID system development. (iii) Competitive IT weapon The proper use of IT is a signicant enabler in creating a competitive edge and enhancing customer satisfaction in the food industry. During the evaluation of our system prototype, food industry practitioners provided positive feedback about the potential use of RFID technology in their industry, commenting in particular that it was benecial for improving food quality control and enabling responsive replenishment. RFID technology used inside a restaurant can also help to create a fresh dining experience and thus enhance diners satisfaction. In the contemporary competitive food industry, the deployment of RFID technology can be a weapon for market success. 6. Conclusion Applying RFID in a conveyor-belt sushi restaurant is an innovative idea that provides diners with a high-technology dining experience. It can contribute to real-time sushi quality information and increased efciency in food management. This case study has described the development of an RFSM system that

can be used to improve operations management in a sushi restaurant, enhancing its efciency and billing accuracy. Though our paper focuses on the development of the RFSM system, some of the context may also be applicable to RFID systems in general. This study is an effort to bring ideas of embodiment into play in the context of RFID. Through this case study, we have demonstrated that RFID technology can be effectively applied in sushi restaurants to help improve their food safety, inventory control, service quality, operational efciency, and data visibility. The understanding of the development of RFSM projects gained from this study could be used as a foundation for conducting future studies involving more RFID application projects in different areas and different organizations. 7. Limitations and future research To end this paper, we discuss the limitations of this research and the extension of future research. First, the inherent limitation of a single case should be noted. Given the single-case study, the external generalizability of the ndings is limited. Future research can address this limitation by examining more RFSM systems in sushi restaurants; a larger sample representing many different RFID projects would also be benecial. Second, in the implementation of the RFSM system, every sushi dish has to have an RFID tag embedded for tracking purposes. RFID tags have been criticized as being high cost and for unavoidably pushing up the system deployment costs. However, the RFID tag used in a restaurant will be in a closed loop system, suggesting that the cost of tags will largely be a one-time investment (as opposed to slap and ship applications in supply chain where most of the tags cannot be re-used). This expenditure will be much less than many other RFID applications in which the RFID tags often get lost during use. Third, the RFSM prototype system can only be applied to simple sushi, which is made up of one raw material topping only. As competition in the food industry is keen, a retailer needs to invent sushi foods with new combinations of raw materials in their toppings in order to retain repeat diners and attract newcomers. Nevertheless, the lessons learned from this case are useful to the development of RFID systems in the food industry. Future research into the areas described below could be conducted to improve the capabilities of sushi restaurants.

ARTICLE IN PRESS
642 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

(i) Many sushi restaurants have already provided more sushi with different combinations of toppings, like octopus, tuna, shrimp, tofu, egg, and cabbage leaves, to suit a variety of diners tastes. By tracking the topping selection statistics in the system, diners topping preferences can be identied and prioritized; thus more sushi with different combinations of favourite toppings can be mixed and created to enhance diners appetites. (ii) A customer membership database can be set up for storing diners personal information, visit frequencies, visit dates and times, eating preferences, and so on. Based on this information, when diners revisit the same restaurant, a better dining experience can be achieved by offering them personalized serving suggestions when they order their preferred sushi. Through this system, the frequent returnees can be tracked and hence effective loyalty campaigns can be launched to retain and reward those returnees. (iii) One of the most important parameters of quality control of food is the temperature. It is feasible to consider applying RFID to control and monitor the temperature of the refrigerator that stores the sushi raw materials to further strengthen food safety control. RFID tags with embedded temperature sensors can achieve a real-time monitoring of the quality of the food. (iv) Nowadays, people are more conscious of food safety. Traceability has become a buzzword with regard to food, particularly following a number of food safety incidents. An end-to-end traceability of all stages in the supply chain is desirable, right from the source where goods and products are manufactured, through handling, storage, transportation, and delivery of a food product from suppliers to consumers, and all the way to inventory control and retail. Until now, RFID applications in the sushi industry have been mostly proprietary and closed loop within the boundaries of a companys operations. Our architectural framework can be further extended to the integrity and traceability of the sushi food supply chain, in order to meet food regulations in certain countries, such as in Europe, that are concerned with strengthening food safety control and monitoring.

Acknowledgements The authors are grateful for the constructive comments of the referees on an earlier version of this paper. This project was supported in part by The Hong Kong Polytechnic University under a Grant number G-YE10. Appendix A A.1. An illustrative example of the system prototype Taking a look at application examples of the system prototype is the best way to understand the RFSM system. The illustrations in this section are intended to provide a better understanding of the system. We demonstrate the primary capabilities of the system and show how it can help to support the management of a sushi conveyor-belt restaurant. Annotations are added to give an insight into the operation of an RFSM system. Fig. A1 shows the food stock of a sushi restaurant. This screen is designed using a red yellowgreen trafc light colour scheme which provides available food stock information to the manager or chef for monitoring the sushi stock. Green indicates that the inventory in the shop oor is above the threshold level, yellow indicates that the inventory is below the threshold level, and red indicates that the inventory is far below the threshold level. The screen can be regularly and automatically refreshed for effective tracking of the food inventory level, in order to achieve responsive replenishment. Fig. A2 shows a particular piece of sushi information. It provides useful food information like prices, calorie, and raw materials to help diners to select their favourite sushi. Fig. A3 shows a screen of a sushi information display board, which demonstrates the real-time sushi location on the conveyor belt. In the centre of the screen, there is a scrolling board, showing the currently available sushi provided in the shop, with its image, price, calorie content, and the number of plates on the belt. The food information is displayed when the plates pass the RFID reader at the table corners. The plates are highlighted in red where the sushi has passed its expiry time. They are removed automatically when passing through the corners of the belt. After the meal, the waiter can use a personal digital assistant (PDA), as shown in Fig. A4, to scan

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 643

Fig. A1. Sushi stock monitoring.

Fig. A2. Sushi-making information.

ARTICLE IN PRESS
644 E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645

food, such as name and price, is shown after scanning. A receipt can be printed for billing.

References
Burnell, J., 1999. Users will overcome many obstacles and implement RFID, study predicts. Automatic ID News 15 (5), 26. Collins, J., 2005. Health care sees safety in RFID. RFID Journal November 23. /http://www.rdjournal.com/article/articleview/ 2005/S. EEC, 2002. Regulation (EC) No. 178/2002 of the European parliament and of the Council of 28 January 2002 laying down the general principles and requirements of food laws, establishing the European Food Safety Authority and laying down procedures in matters of food safety. Ofcial Journal of the European Communities, 1124. Gasching, J., Klahr, P., Pople, H., Shortliffe, E., Terry, A., 1983. Evaluation of expert systems: Issues and case studies. In: Hayes-Roth, F., Waterman, D.A., Lenat, D.B. (Eds.), Building Expert Systems. Addison-Wesley, Reading, MA, pp. 241280. Herbst, J., Karagiannis, D., 2000. Integrating machine learning and workow management to support acquisition and adaptation of workow models. International Journal of Intelligence Systems in Accounting, Finance & Management 9, 6792. Hutter, L., 2004. Food safety. Chain magazine. /www. foodchain-magazine.comS. Jaseiskis, E.J., Ei-Misalami, T., 2003. Implementing radio frequency identication in the construction process. Journal of Construction Engineering & Management 129 (6), 680688. Jedermann, R., Behrens, C., Westphal, D., Lang, W., 2006. Applying autonomous sensor systems in logisticsCombining sensor networks, RFIDs and software agents, sensors and actuators. Sensors and Actuators A: Physical 132 (1), 370375. Jiji Press English News Service, 2003. Wired feed. Tokyo, June 3, p. 1. Jones, P., Clarke-Hill, C., Comfort, D., 2005. Radio frequency identication and food retailing in the UK. British Food Journal 107 (6), 356360. Karkkainen, M., 2003. Increasing efciency in the supply chain for short life goods using RFID tagging. International Journal of Retails & Distribution Management 31 (10), 529536. Kerry, J.P., OGrady, M.N., Hogan, S.A., 2006. Past, current and potential utilization of active and intelligent packaging systems for meat and muscle-based products: A review. Meat Science 74, 113130. Kumar, S., Budin, E.M., 2006. Prevention and management of product recalls in the processed food industry: A case study based on an exporters perspective. Technovation 26, 739750. Laudon, K.C., Laudon, J.P., 1998. Management Information Systems: New Approaches to Organization and Technology. Prentice-Hall, Englewood Cliffs, NJ. McMeekin, T.A., Baranyi, J., Bowman, J., Dalgaard, P., Kirk, M., Ross, T., Schmid, S., Zwietering, M.H., 2006. Information

Fig. A3. Real-time display of sushi position on the conveyor belt.

Fig. A4. PDA device for billing.

the sushi plates (embedding an RFID tag in or under the plate) with the RFID reader. On the device screens, information about the consumed

ARTICLE IN PRESS
E.W.T. Ngai et al. / Int. J. Production Economics 112 (2008) 630645 systems in food safety management. International Journal of Food Microbiology 112, 181194. Mennecke, B., Townsend, A., 2005. Radio frequency identication tagging as a mechanism of creating a viable producers brand in the cattle industry. MATRIC Research Paper 05-MRP 8, MATRIC Website. /http://www.matric.iastate. eduS. Morrison, J., 2005. Help wanted. RFID Journal March/April, 1320. Mousavi, A., Sarhavi, M., Lenk, A., Fawcett, S., 2002. Tracking and traceability in the meat processing industry: A solution. British Food Journal 104 (1), 719. Ngai, E.W.T., Cheng, T.C.E., Au, S., Lai, K.H., 2007. Mobile commerce integrated with RFID technology in a container depot. Decision Support Systems 43, 6276. Nunamaker, J.F., Chen, M., Purdin, D.M., 1990. Systems development in information system research. Journal of Management Information Systems 7, 89106. OConnor, M.C., 2006. Boeing selects chipmaker for parts tags. RFID Journal April, 4 /http://www.rdjournal.com/article/ articleview/2236/S. Regattieri, A., Gamberi, M., Manzini, R., 2007. Traceability of food products: General framework and experimental evidence. Journal of Food Engineering 81 (2), 347356. 645 Riso, F., 2001. Intelligence chip technologyAn RFID update. Paper presented at the Intelligent Supply Chain, CIES The Food Business Forum, 45 October, Hotel Okura, Amsterdam. Scott-Morton, M.S., 1984. The state-of-the-art of research in management support systems. In: McFarlan, F.W. (Ed.), The Information Systems Research Challenge. Harvard Business School Press, Boston, MA, pp. 1341. Song, J., Hass, C.T., Caldas, C., Ergen, E., Akinci, B., 2006. Automating the task of tracking the delivery and receipt of fabricated pipe spools in industrial projects. Automation in Construction 15 (2), 166177. Swedberg, C., 2006. Crown saves manufacturing costs via RFID. RFID Journal November 10. /http://www.rdjournal.com/ article/articleprint/2810/-1/1S. Thysen, I., 2000. Agriculture in the information system. Journal of Agricultural Engineering Research 76 (3), 297303. Wang, N., Zhang, N., Wang, M., 2006. Wireless sensors in agriculture and food industryRecent development and future perspective. Computers and Electronics in Agriculture 50, 114. Wyld, D.C., Jones, M.A., Totten, J.W., 2005. Where is my suitcase? RFID and airline customer service. Marketing Intelligence & Planning 23 (4), 382394.