UNIVERSITY OF VAASA FACULTY OF TECHNOLOGY DEPARTMENT OF PRODUCTION Kyösti Alanen POTENTIAL OF RFID TECHNOLOGY IN LOGISTICS - CASE METSO PAPER - Master’s Thesis in Industrial Management VAASA 2008 2 CONTENTS TIIVISTELMÄ ABSTRACT 1. INTRODUCTION 7 1.1 Preamble 7 1.2 Purpose, goals and definition of study 8 1.3 Metso Paper Service 9 1.4 Presentation of the study 10 2. THE BASIS FOR LOGISTICS AND IDENTIFICATION 11 2.1 Logistics 11 2.2 Identification 12 3. INTRODUCTION TO AUTOMATIC IDENTIFICATION TECHNOLOGIES 14 3.1 Bar code 15 3.2 RFID 16 3.3 Biometrics procedures 16 3.4 Optical character recognition 17 3.5 Smart cards 17 4. BAR CODE SYSTEM 19 4.1 Bar code symbol 19 4.1.1 Linear bar code symbol 20 4.1.2 2D bar code symbol 21 4.2 Industrial bar code symbols 22 4.3 Bar code reader 23 4.3.1 Contact readers 24 4.3.2 Non-contact readers 24 4.3.3 Conveyor bar code readers 25 4.3.4 Vision-based reading 26 4.4 Printing of bar code symbol 26 4.5 Standards 28 5. RFID 29 5.1 Tag 29 5.1.1 Tag categories 30 5.1.2 Smart labels 32 5.1.3 Standards 32 5.1.4 Information storage capacity 35 5.1.5 Tag protocol 36 5.2 Reader 38 5.2.1 Layout for readers and antennas 39 5.2.2 Reader protocol 40 5.3 RFID Printer for smart labels 42 5.4 Middleware 42 3 6. HOW TO DETERMINE INVESTMENT ADVISABILITY 44 6.1 SWOT analysis 44 6.2 Pay-back time calculation 45 7. PROCESS DESCRIPTIONS 47 7.1 Warehouse operations 49 7.1.1 Characteristic 49 7.1.2 Problems 52 7.1.3 SWOT 53 7.2 Direct delivery processes 54 7.2.1 Sizer consumables 55 7.2.1.1 Characteristics 56 7.2.1.2 Problems 57 7.2.2 Doctor Blades 57 7.2.2.1 Characteristic 58 7.2.2.2 Problems 59 7.2.3 SWOT for direct deliveries 60 7.3 Consignment stock process for Doctor Blades 61 7.3.1 Characteristics 62 7.3.2 Problems 64 7.3.3 SWOT 65 7.4 Receiving process of spare part package 66 7.4.1 Characteristic 67 7.4.2 Problems 68 7.4.3 SWOT 69 7.5 Return and repair processes 70 7.5.1 Characteristics of returns 71 7.5.2 Characteristics of repairs 72 7.5.3 Problems 73 7.5.4 SWOT for repairs and returns 74 7.6 Roll coatings 75 7.6.1 Characteristics 76 7.6.2 Problems 77 7.6.3 SWOT 77 7.7 Roll workshop operations 79 7.7.1 Characteristics 79 7.7.2 Problems 81 7.7.3 SWOT 81 7.8 Summary 82 8. EVALUATING SELECTED PROCESS IN MORE DETAIL 85 8.1 Modelling operational RFID-managed consignment stock 85 8.2 Modelling technical architecture for RFID-managed consignment stocks. 88 8.3 Pay-back time and net savings 91 8.3.1 Costs 91 8.3.2 Benefits 94 8.3.3 Pay-back time and total net savings 95 4 9. CONCLUSION 97 9.1 Summary of findings 97 9.2 Suggestions for further studies 99 10. SOURCES 100 11. INTERVIEWEES 103 12. APPENDICES 104 5 VAASAN YLIOPISTO Teknillinen tiedekunta Tekijä: Kyösti Alanen Tutkielman nimi: Potential of RFID Technology in Logistics - Case Metso Paper - Ohjaajan Nimi: Josu Takala Tutkinto: Kauppatieteiden maisteri Laitos: Tuotannon laitos Oppiaine: Tuotantotalous Opintojen aloitusvuosi: 2005 Tutkielman valmistumisvuosi: 2008 Sivumäärä: 117 TIIVISTELMÄ: Viime vuosina palveluliiketoiminnan merkitys ydinosaamista tukevana prosessina on kasvanut merkittävästi. Metso Paperin ydinosaamista on paperikoneiden valmistus mutta Service liiketoimintaan panostetaan vahvasti ja siltä odotetaan kasvua. Tällä osa-alueella logistiikalla ja etenkin materiaalin tunnistamisella on iso vaikutus prosessien tehokkuuteen. Viivakoodi on yleisesti ollut hallitseva automaattisen tunnistamisen menetelmä, mutta sillä on omat rajoituksensa. RFID:llä nämä rajoitukset voidaan voittaa. Standardisoinnin sekä teknisen kehityksen ansiosta se on nopeasti noussut vaihtoehtoiseksi menetelmäksi tehostaa logistiikkaa. Siksi Metso Paper Service on nähnyt RFID tutkimuksen tarpeelliseksi. Tämän Pro Gradu tutkimuksen tarkoitus on selvittää, voidaanko RFID:llä tehostaa Metso Paper Servicen nimettyjä prosesseja. Tavoitteena on tunnistaa ne prosessit, joissa RFID:llä voitaisiin saavuttaa liiketoiminnallisia parannuksia ja kustannussäästöjä nykyiseen toimintamalliin verrattuna. Tutkimus on toteutettu haastattelemalla avainhenkilöitä teemakysymyksillä kahdeksasta toimeksiantajan nimeämästä prosessista. Haastattelun ja tilastollisen aineiston perusteella, nykyiset toimintamallit ja materiaalin tunnistamiseen liittyvät oleelliset asiat on kuvattu prosessikarttoineen. Lisäksi teoriaosa esittelee RFID- tekniikan pääosin tasolla, joka käyttäjän on hyvä tietää. Tältä pohjalta on analysoitu, pystytäänkö RFID:n avulla tehostamaan prosessin toimintaa. Tutkimuksessa havaittiin, että useampi prosessi kehittyisi jollakin tavalla RFID:stä, mutta toteutettavuus ja saavutettavan hyödyn määrä vaihtelevat. Kuitenkin yksi prosessi muodostui muita selvästi sopivammaksi. Tutkimuksen viimeisessä osassa on selvitetty RFID:n tarkemmat toiminnalliset sekä taloudelliset vaikutukset tähän prosessiin, niin tarkasti kuin se etukäteen on mahdollista. Lopputuloksena päädyttiin suosittelemaan RFID pilottiprojektia suomalaisten asiakkaiden kanssa. AVAINSANAT: RFID, automaattinen tunnistaminen, prosessin tehostuminen, taloudellinen hyöty 6 UNIVERSITY OF VAASA Faculty of technology Author: Kyösti Alanen Topic of the Master’s Thesis: Potential of RFID in Logistics - Case Metso Paper - Instructor: Josu Takala Degree: Master of Science in Economics and Business Administration Department: Department of Production Major subject: Industrial Management Year of Entering the University: 2005 Year of Completing the Master’s Thesis: 2008 Pages: 117 ABSTRACT: In recent years, the importance of service business as a supporting process of core know-how has increased significantly. The core know-how of Metso Paper lies with the manufacture of paper machines, but plenty of effort has been paid on Service business and a great build-up is expected. Logistics and especially material identification affect to a large extent the overall efficiency of individual processes. In general, bar codes have been the dominant method of automatic identification, but they have their limitations. Thanks to standardisation and technical improvements, RFID has rapidly become an alternative way to improve logistics. Thus, Metso Paper Service has deemed research into RFID worthwhile. The purpose of this Master’s Thesis is to determine whether RFID could improve given operations of Metso Paper Service. The goal is to identify application areas, where significant improvements and cost savings might be gained by introducing RFID, compared to current ways of operation. The study has been conducted by interviewing key persons of eight processes. The client has chosen the processes, and interviews were conducted with theme- questioners. Based on interviews and statistics, essential aspects of current operating and material identification methods were modelled with process descriptions. In addition, the theory section introduces RFID in the level that is beneficial to users’ point of view. Based on these, it is analysed whether RFID could improve the processes. The study found out that RFID could somewhat improve several processes but feasibility and gained improvements vary. However, one application was found out to have the greatest potential. The last part of this study clarifies detailed operational and financial issues, as far as that can be achieved in advance. As a result, the study ends up recommending a RFID pilot project among Finnish customers. KEYWORDS: RFID, automatic identification, process improvements, financial benefits 7 1. INTRODUCTION 1.1 Preamble In an industrial environment, automatic identification procedures have been around many years. They exist to provide information about items and related things and have become very popular. They have been used to accelerate processes and to reduce time-consuming or routine work among purchase and distribution logistics, manufacturing and material flow systems. They have resulted in more accurate inventories and more efficient material handling, because identification is not relying of human beings as much as in the past. Some considerable time ago bar code systems started a revolution in logistic identification systems and nowadays they can be found in almost every product. Although they may be extremely cheap and bar code compliant devices are easy to obtain, they are proving to be inadequate in an increasing number of cases. The reason relates to their comprehensive limitations, such as the very short read distance and low storage capacity. Nevertheless, bar codes remain a very useful method of conducting identification in many applications. Technically optimal way to carry out extensive automatic identification is based on smart card technology. In such a system data is stored in a silicon chip that is to be attached on a card. For instance, credit cards are based on that technology. In general, smart cards are impractical for logistic purposes, although they can store lots of data but identification is based on mechanical contact. Thus, non-contact ways for identification between object and readers were needed and developed. (Finkenzeller, 2003: 1) In recent years the most talked-about procedure in the field of automatic identification has been RFID. It stands for radio frequency identification. The RFID is no longer a state-of-the-art procedure, but rapidly developing information 8 technology has made it more attractive to an extensive amount of potential users. The development of microprocessors and silicon chips has helped RFID to overcome some technical challenges and pushes its cost downwards whenever a new generation of chips has been launched. The RFID provides capability to attach an electronic identity to a physical object, which effectively extends Internet into the physical world. For logistics this can lead to faster order automation, tighter process control, precise up-to-date inventories and real-time locations. In a wider scale, business partners are able to share information on the goods through a supply chain in a way not yet conceivable a few years ago. (Glower, 2006: 5) 1.2 Purpose, goals and definition of study The study has been made on the assignment of Metso Paper Service. The purpose is to figure out whether RFID could improve given operations of Service. The goal is to identify application areas, where significant improvements and cost savings might be gained by employing RFID, compared to current ways of operation. Metso Paper Service falls into many processes, where material identification is an important part of everyday operation. Depending on the character of the process, needs and ways for material identification varies but the common feature is that RFID is not yet either used or considered carefully anywhere. Thus, this study has been considered reasonable. The definition of the study is to focus on the usability of RFID in the processes of Metso Paper Service and not on the technical details of RFID no more than to a degree necessary to understand about how RFID can actually be used. Another constraint is to focus on logistics and not maintenance operations, while identifying potential RFID application areas. In addition, possible pilot projects or implementation works are outside the scope of the study, as well. 9 1.3 Metso Paper Service Metso Paper is a global provider and market leader in pulping, paper and board production, as well as power generation technologies. Its product portfolio serves customers throughout their processes, from pulp making to the wrapping of finished rolls. The company has its own operations and production in 28 countries and its products and services are sold by more than 20 sales units and 40 service centres in different parts of the world, as well as the logistics centres in Finland, the USA and China. Approximately one third of the global paper production is performed on production systems supplied by Metso Paper. The largest market areas are Europe, Asia and North America. (Metso Paper, A) Metso Paper Service is part of Metso Paper’s Paper and Board business line. It covers three sectors such as traditional equipment service, maintenance and product support services. The maintenance service stands for Metso Paper’s partial or full responsibility of customer maintenance operations. The product support service means that Metso can support process and product development in co-operation with the customer. However, both of them are outside the scope of this study, since the study focuses on equipment service processes. The equipment services consist of three expert sectors, which are Field, Roll and Spare part services. In addition, Metso Paper Service provides automation and field system services and upgrade solutions tailored to customers’ needs but these are outside the scope the study. Consequently, the focus is on the processes in expert sectors, which are clarified as follows: • The field service includes maintaining and enhancing the performance of fibre processing and paper making lines. • The roll workshop stands for mechanical roll service, replacement of rolls and roll covers. • The spare part service includes daily spare, spare part packages and consumables. (Metso Paper, B, 4) 10 1.4 Presentation of the study The study consists of introduction, theory and empirical research sections. The first chapter is an introduction that deals with the purpose, goals and definition of the study. Chapters 2 – 6 form the theory section, giving some basic information on logistics and automatic identification methods. Chapters 4 and 5 take deeper insight into bar codes and RFID through literature review. The viewpoint is end-user oriented and deep technical details are mainly omitted. These techniques are the most famous ones to conduct automatic identification in logistics. The purpose of the study is to determine whether RFID could improve applications in Service. However, it was also seen reasonable to introduce alternative bar codes in order to understand differences in their capabilities. Consequently, it attempts to clarify why Metso Paper Service is particularly interested in the possibilities of RFID. Finally, the theory section ends with chapter 6, which introduces two methods that can be used to determine investments advisability. These are SWOT analysis and pay-back calculation. Both of them were used during the empirical research stage. The chapters 7 – 8 deal with empirical research. The chapter 7 takes a look into processes, in order to understand their main operational issues. The processes are studied one by one ending up to SWOT analysis, which is used to evaluate RFID suitability in that application. Finally, the chapter ends with a summary indicating the most potential processes for RFID. The most potential one or ones will be further investigated in chapter 8, which clarifies how RFID-managed processes run operationally and what kinds of devices are needed. The chapter ends with calculations that indicate pay-back time and net savings in a given period of time. Finally, the findings are summarised in chapter 9 and suggestions made for further RFID studies. 11 2. THE BASIS FOR LOGISTICS AND IDENTIFICATION 2.1 Logistics Logistics is the applied science of planning and implementing the acquisition and use of resources. However, if you ask people to define logistics each one might give a different answer. Generally, the definition depends on the concept the definer has of its application context. Thus, logistics can be defined in many ways depending on a person's business and role in life. For example, the automotive industry defines logistics as the entire process from the source of raw materials to the manufacturing process that results in cars for purchase. (Jones, 2006: 21) Dictionary of transport and logistics vocabulary determines logistics as follows: Logistics is time-related positioning of resources to meet user requirement. (Lowe, 2002: 147) To clarify the concept of logistics further, it can be considered to consist of a standard set of actions. Those are definition of need, identification of limits, determination of terminal objectives, measurability and assessment. Thus, all logistics actions are based on meeting a completely predetermined need. Simultaneously each resource has limitations, ranging from minimum to maximum acceptability of the situation. The stated set of terminal objectives guides all logistics activities in their application. By establishing measurable criteria for processes, the process holder is able to guide the progress towards terminal objectives. Measurements can be tangible or intangible. In case of a tangible characteristic, such as number of pieces, everybody can easily determine, if the criteria have been met. But intangible characteristics are usually understandable only for process holders. Continuous assessment needs to run on the background to determine the success or failure of 12 any single activity. It is conducted by comparing the progress against the given measurable criteria. In practise, assessment before starting the activity may simply be checking that all needed things are in place. During the activity assessment controls near-term success of short-term needs that eventually lead to overall success. At the end of activity assessment it is ensured that all terminal objectives have been met. (Jones, 2006: 21-23) 2.2 Identification Identification means classifying, counting and organising objects. These operations are the very essentials in logistics environment such as manufacturing, distribution and various stages of supply chain operating from the scale of individual consumer to global trade. In the past industrial identification was done visually just by observing the characteristics of objects. When identical objects have to be identified distinguishing markings have been added. Further accurate and efficient means are needed to recognise those markings, in order to identify the objects. Therefore, an identification system consists of identifying markings and readers of those markings. The first readers were human beings but by the time technical innovations resulted in cameras and laser devices they started to be used as readers. Simultaneously basic written markings have evolved into commonly used bar codes that can be found from almost every package and item. (Committee on Radio Frequency Identification Technologies, 2004: 3) Typically identification in the supply chain involves tracking and controlling critical information in real-time. Consequently, it enables reacting to changing circumstances faster and gain real-time competitive advantages. For instance, effective material identification delivers transparency in supply chains and results in: • Quicker receiving, as unnecessary manual steps are removed. 13 • Less time is spent in solving logistics issues. • Early notification of subcontractor delivery issues enables corrective actions to be taken early. • Exact status of inventory is available in real-time, resulting in fewer buffers. • Faster and easier inventory management and efficient logistics process. • Decreases loss of goods and assets. • No unexpected product shortages. • More manageable product liability. • Increased customer satisfaction as service levels improve. ( Trackway, 2008) 14 3. INTRODUCTION TO AUTOMATIC IDENTIFICATION TECHNOLOGIES When human beings are involved in the identification of objects, keeping track of changes in locations and inputting this information into a database, the process is time consuming. In addition, the process is vulnerable to human errors. (Muller, 2002: 89). Thus, automatic identification procedures (Auto-ID) exist to provide timely and accurate information on goods and products in transit. A long time ago bar code systems started a revolution in that field but nowadays there are also several another identification systems available, such as Radio Frequency Identification (RFID). Currently, bar codes and RFID are the most widely used identification procedures in logistics. Meanwhile, others not so significant procedures in terms of logistics are Biometrics, Optical Character recognition and smart card procedures. Picture 1 summarises the most important auto-ID procedures. (Finkenzeller, 2004: 1-2) There are some features that must be taken into account when evaluating automatic identification technologies. One of them is error rate. It refers to the probability that a given number of scanning occasions include an error. Then the expected number of errors can be calculated by multiplying error rate by the number of characters scanned. Another one is the first read rate. It refers to the probability that an attempt to read a character is successful on the first attempt. The next important feature to be considered is scanning distance between the reader and the object and can a moving object be scanned. Last but not least, it must be considered whether the technology permit modification of the recorded data or not. That ability will offer the greatest flexibility. In addition, in a busy environment the time that a single scanning takes may play important role to an operating speed. (Palmer, 2001: 3-4) 15 Picture 1. Summary of the most important auto-ID procedures. 3.1 Bar code The bar code is a binary code comprised of a bunch of bars and gaps arranged in a sequential order in some predetermined way. This design can be interpreted numerically and alphanumerically, in order to identify the object. Bar codes are read by optical laser scanning based on the different reflection of a laser beam from the black bars and white gaps. There are several different types of bar code symbols currently in use, each of them were developed for the purpose of some particular application. As a consequence, the same physical design can represent different meanings in different types of bar codes. (Finkenzeller, 2004: 2-3) The reading distance is relatively short, ranging from the near contact only up to about a couple of meters. Reading always requires a clear line of sight to the undamaged barcode to ensure correct interpretation of it. The data security is high and normal error rate can even be less than one error in 1 million characters. Conventionally, the first read rate is better than 80 % and might even be close to 100 %. The biggest limitation of the bar code system is that codes can be written only once and any additional information cannot be added later on. Partly due that reason the price of enabling a bar code system is cheap. (Palmer, 2001: 9) 16 3.2 RFID Radio Frequency Identification (RFID) is a procedure to identify objects by using radio waves. Therefore, there must be an identification element called a tag, which holds the identification data, attached to the object. The tag responds to the radio waves. Successful identification does not require a direct line of sight between the tag and the radio waves transmitter called a reader. In addition, the scanning distance can vary from near contact always up to a long way off, depending on the coupled design of the reader and the tag antenna. The error rate is very low and conventionally tags can be read on the first attempt. These characters enable quite a wide range of applications for RFID and simultaneously make it more efficient and accurate than other identification technologies. Tags are an essential part of RFID systems. The simplest version of tags is a passive tag. It does not have its own power source and is entirely dependent on getting power from the reader. A passive tag does not provide much space for data and can not be rewritten; usually only identification date is encoded into it. Adding even a simple sensor or power source into a tag can increase its utility radically. This makes a tag active. More information can be encoded initially and even additional information can be encoded later on. The cost of RFID tags vary from a fraction of US dollars with passive tags up to several hundreds with active tags. This has been delaying RFID revolution. In recent years prices have gone down, boosting the utilisation of this technology. (Committee on Radio Frequency Identification Technologies, 2004: 3-5) 3.3 Biometrics procedures Biometrics procedures in identification systems mean the ways that living beings are possible to identify. That is done by comparing unmistakable and individual physical characteristics. One way to conduct identification is to take a fingerprint not only from the finger itself, but also from the object that individual in question 17 has touched. Then the system checks the database in order to find a match. This is a commonly used procedure to permit entrance and also to identify criminals. Another application is voice identification. In that case the user talks to the computer, which convert the voice into a digital signal. The software evaluates the signal against the database. If it matches, a reaction can be initiated. (Finkenzeller, 2004: 4) 3.4 Optical character recognition Optical Character Recognition (OCR) was developed for the purpose that characters could be read both in the normal way by people and automatically by a machine. This system provides high density of information but failed to become a popular application, because of its compliance and expensive components, in comparison with other identification products. An additional negative aspect is that the reading rate is slower and error rates higher, if compared to bar code system as an example. In some scale, this procedure is used in service and administrative fields. (Finkenzeller, 2004: 3-4) 3.5 Smart cards The smart card is an electronic data storage system. The characteristic feature of it is an integrated circuit called a chip that is incorporated in the card. The chip has components for storing, transmitting and processing data. Data is transmitted when a smart card is placed in a reader, which makes either a galvanic connection on the contact surfaces or an electromagnetic field without any contacts. The most significant advantage of the smart card is that the stored data can be protected against undesired access and manipulation. Smart card technology is also reliable and has a long lifetime. The development of chips is very fast and their capacities have been multiplied with every new generation of chips. It is possible to divide smart cards into two different categories based on their functionalities. These are 18 memory cards and microprocessor cards. Memory cards contain EEPROM memory. The memory is for the program code needed by the application. These cards are usually specified for some particular application, which cannot be changed later on. On the other hand, memory cards are inexpensive. For that reason they are typically used in price-sensitive applications, such as prepaid mobile phone cards. (Rankl & Effing, 2004: 17-19) Microprocessor cards contain a microprocessor, which includes ROM, RAM and EEPROM memories. Contest of ROM is defined during the manufacturing and its purpose is to incorporate the microprocessor and operating system. This memory cannot be overwritten. The RAM is temporary working memory and does not maintain the data, since the card is disconnected from the reader. The EEPROM is for application data and can be controlled by the operating system. Primarily microprocessor cards are aimed for security sensitive environments and can include a number of applications. Examples of usage are credit cards and mobile phone SIM cards. (Finkenzeller, 2004: 6) 19 4. BAR CODE SYSTEM Automatic identification can be conducted by bar coding. It is an optical read-only procedure, which is based on the reflection of light off a printed pattern. The dark bars or areas of the pattern absorb light and the intervening spaces reflect the light. The contrasting absorption and reflection is observed by the reader, which decodes the information. The pattern, which is the real arrangement of parallel bars and spaces that encode the data, is usually called a bar code. A bar code system conventionally consists of three components: the code itself officially known as the bar code symbol, the reading device and the printer. There are some international bar code standards, which determine generic rules for defining bar code symbols, as well as specific rules for some particular application. (Muller, 2002: 90) 4.1 Bar code symbol Conventionally, a bar code symbol consists of bars and intervening spaces to encode the data. A given number of bars and spaces build up a character and a given number of characters build up a bar code symbol. It is appropriate to differentiate between the terms “code” and “symbol”. A code refers to the actual data contained in characters, whereas a symbol refers to the actual arrangement of sequential bars and spaces. In this respect, the pattern should not be called a bar code but a bar code symbol instead. Some symbols can encode only numbers, whereas other symbols encode alphanumeric characters. Bar code symbols can be divided into linear and 2D categories. (Palmer, 2001: 15-18) 20 4.1.1 Linear bar code symbol A linear bar code symbol, picture 2, is a single row of bars and intervening spaces. This is the oldest form of bar codes. It has a low capacity, typically from 15 to 50 characters, depending on which type of symbol is in question. The data can be encoded either in a width-modulated or a height-modulated way. The first one means that bars and spaces vary in width and in the latter they vary in height. Further width-modulated symbols falls into discrete and continuous types. Characters in a discrete code stand alone. There is a gap between each character. It starts with a bar and ends with a bar. This feature makes discrete code easier to read and print. Characters in a continuous code start with a bar but end with a space. There are no gaps between characters. This feature allows the insertion of more characters in a smaller space, which is a good thing with a bar code label with limited space available. The generic structure of linear bar code symbols includes three parts: • There is a quiet area at both side of the symbol. Its purpose is to distinguish a symbol from its background in order for readers to identify the code accurately. • There is a start and stop character. The start character indicates the starting point of the symbol and, as the name suggests, the stop character indicates the point where the symbol ends. That symbols enable scanning devices to read a bar code symbol from left to right and vice versa. • Between start and stop characters are actual data characters, which compose the message. (Palmer, 2001: 16-24) Picture 2. Linear bar code symbol. 21 4.1.2 2D bar code symbol 2D bar code symbols were developed in an attempt to reduce the room typically needed by conventional linear bar code symbols. Because the more data is encoded into a linear symbol, the taller a symbol would be. In many cases traditional symbols act as a license plate to reference information stored in a database. 2D symbols are able to do that with significant less space or even function as the database itself, because of their higher data storing capacity of up to 2000 characters. 2D bar code symbols can be divided into two categories, which are 2D stacked and 2D matrix symbols. These are illustrated in picture 3. 2D stacked symbols are basically very long linear bar code symbols cut up into shorter linear lines and stacked up in a multi-row arrangement. All the rows are the same length and either touch each other or include a single bar separating them. This is printable with similar techniques as linear symbols. More sophisticated are 2D matrix symbols. They do not consist of rows of bars similarly as stacked symbols but rather of a grid of square cells. Thus, they are read independent of orientation. Matrix symbols have even higher data capacity than stacked symbols, but require special printing and reading equipment due to the resolution requirements. (Palmer, 2001: 17, 48, 58) Picture 3. 2D stacked and matrix bar code symbols. 22 4.2 Industrial bar code symbols All different bar code symbols have their own fixed alphabets made up of bars and spaces coupled with the rules for how they are presented. Some of them employ only numbers, whereas others also employ alphabets and even special characters. Some widely used industrial symbols are presented below. (Muller, 2002: 95) Code 39 Code 39 is popular in industrial bar code applications, such as warehousing, tracking shipments and manufacturing. It is illustrated in picture 4. It contains 43 data characters; numbers, all uppercase letters and seven special characters. The code can be printed easily by most software available and can vary in length. It is also self-checking and discrete. Self-checking means that a single printing error cannot cause a character to be mixed with another valid character in the same symbol. All of its characters consist of 9 elements, five bars and four spaces. Three of these elements are wide and six are narrow, making up its name code 39. (Palmer, 2001: 27-30) Picture 4. Code 39. Code 128 Code 128 is a linear symbol, which is increasingly used, for instance, in retail distribution applications for serialised carton tracking. It is illustrated in picture 5. It is continuous and can vary in length. This code type supports all the ASCII characters, so all alphabets can be used in upper and lower case letters, as well as numbers and all special characters. The code 128 has three alternate character sets. 23 Each of them includes shift and start codes to control which set is used. This feature permits changing a character set inside a symbol, in order to express the encoded message as shortly as possible. Each character of the code consists of 11 elements with equal width, which further build up totally three bars and spaces. For example, a single bar could be from 1 up to 3 elements wide. (Palmer, 2001: 34-37) Picture 5. Code 128. Code 49 Code 49 is a 2D stacked symbol. It contains two to eight adjacent rows separated by a single-module separator bar. Each row has eight characters consisting of 70 elements, which equal 17 bars and 17 spaces eventually making up four words. Rows can be in any order, because each row contains a row number and the bottom row encodes the total number of rows in the symbol and check characters. The number of rows depends on the number of the characters being encoded. For example, eight rows provide enough room either for 49 alphanumeric or 81 numeric characters. The code 49 supports ASCII characters. (Palmer, 2001: 48-50) 4.3 Bar code reader A device that reads bar code symbols is called a bar code reader or a bar code scanner. The purpose of these devices is to determine the exact width of bars and spaces and decode that information into digital forms that a computer can understand. This is accomplished by projecting a tiny laser beam that crosses the bar code symbol and then measuring the amount of reflection off bars and intervening 24 spaces. A reading device is able to transmit decoded data instantly to the attached computer or can interact with an application program that is resident in the reading device itself. To avoid misreads and compatibility problems with given bar code symbology the beam of light of scanning devices must not be larger than the X dimension of a bar code symbol. That dimension is the width of the narrowest bar, as well as space in a given type of symbol, such as code 39. In addition, readers for 2D codes are able to read linear codes, as well but that does not work the other way around. Bar code readers fall in the following subcategories. These are contact, non-contact and conveyor readers. (Muller, 2002: 101-103) 4.3.1 Contact readers As the name suggests, contact readers physically touch the symbol that is being scanned. Typically those kinds of devices come in the form of a light pen or wand and are used in an office environment to scan bar code symbols on the papers substituting a manual key entry. The beam of the device is fixed and the scanning motion comes from the user, who manually passes the device across the bar code symbol. Despite contact with the symbol it still has some depth of field, meaning that a thin laminate can be employed to protect a bar code symbol being scanned. (Palmer, 2001: 126-127) 4.3.2 Non-contact readers Non-contact readers can either be handheld or stationary. Handheld means that users need not write but must place the reader near the bar code symbol. Stationary is for automatic reading, then the object must be placed under a scanner in some predetermined position within a given distance that a direct line of sight exists between the object and the reader. In addition, non-contact readers are able to read codes on soft or irregular surfaces even through thick laminates or windows. Typically non-contact readers come in the shape of a pistol. These readers employ 25 fixed beam, moving beam, and charged couple device CCD and rastering laser technologies: • Fixed beam scanners use a stationary beam to scan a bar code symbol. The operator provides the scanning motion. • Moving beam scanners use a moving beam to scan a bar code symbol and no additional moving motion is required. • CCD scanners scan the light path as a whole with electronic detectors. • Rastering scanners have higher horizontal and vertical scanning amplitude in order to capture a rectangular area, which matches to 2D symbols. (Palmer, 2001: 128-132) 4.3.3 Conveyor bar code readers Bar code readers alongside a conveyor face some special challenges. Basically, the location and orientation of symbols on the object are fixed but the position and orientation of the object on the conveyor are unknown. In addition, the speed of the conveyor must be taken into account as well. Existing device alternatives are orientation-dependent and omnidirectional laser scanning. Orientation-dependent laser scanning utilises a fixed-mount moving beam technology. If the object and symbol orientation are fixed, then those types of devices can be used. Also the scanning line length, scanning rate, bar height and conveyor speed should be such that a scanner has the minimum of four opportunities to scan the object as it moves by. If a rastering scanner is utilised, a successful scan is more likely, because it moves the scan line direction perpendicular to the scanning motion. Then poor symbol placement is better compensated. The probability of a successful scan is far better if omnidirectional laser scanning is utilised. It is a sophisticated version of fixed-mount moving beam technology. The primary idea is to project a series of straight or curved scanning lines of varying directions in a star-shaped form over the object. Then at least one of the scanning 26 lines will be able to cross the symbol, no matter what symbol orientation or location. It must be mentioned that 2D symbols cannot be read by omnidirectional laser scanning. (Palmer, 2001: 147-150) 4.3.4 Vision-based reading As the name suggests, vision-based reading devices take electronic pictures including a symbol of the object. No laser is used. Then special software detects and decodes the symbol from the electronic picture. This procedure is independent of careful position and orientation of either the object or the symbol. It has the ability to interpret basically all conventional linear symbols as well as 2D stacked and matrix symbols. Three basic types of devices are available: • Handheld vision scanners take a picture when the operator discharges a trigger. • Fixed-mount vision scanner using 2D imager can be used to take a snapshot of the screen or a continuously strobed 1D imager can be used to snap continuous images of passing objects. Both can be done unsupervised. • Fixed-mount vision scanners using linear imagers are typically custom built. They can be used to scan symbols in a 2D format on rapidly moving conveyor lines. This happens automatically and does not require an operator. (Palmer, 2001: 153-155) 4.4 Printing of bar code symbol In a bar code system everything starts with a symbol. That symbol must be generated in some practical form available. Usually this is done by printing a symbol on the label or various kinds of similar substrates, which are attached to the object. Bar code printing can be divided into offsite and onsite printing. 27 Offsite printing Offsite printing conventionally takes place at a location different from the one the bar code symbols will be actually used. These techniques are designed for mass production of identical or sequenced symbols. Often this kind of printing service is purchased from specialised service providers, such as print shops and bar code symbols are included in the actual carton or packaging material permanently. Onsite printing It conventionally takes place at the time and place symbols are to be used. This enables the encoding of unique data into each symbol on demand. On the other hand, it is not a competitive solution in terms of speed for large-scale printing of identical symbols. The absence of this feature is not a problem in its typical application in warehouse and retail industry. (Palmer, 2001: 159-165) Some basic onsite printer techniques are described below: Direct Thermal • Printer forms symbols on a paper by selectively heating localised areas of paper. This is done by the elements in the printhead, which is in contact to the paper. Thermal transfer • Printer forms a symbol on a paper from a ribbon that is heated by the elements in the printhead. Ink jet • Printer has a fixed printhead, which sprays tiny droplets of ink on paper. Laser / xerographic • Printer has a controlled laser beam, which creates a symbol on paper. (Aim, 2008) 28 4.5 Standards Standards are very important for the acceptance and adoption of new technologies. Consequently, there are three main standards to ensure successful operation with bar codes. These are symbol, print quality and application standards. In addition, there is also an organisation called Association for Automatic Identification and Mobility (AIM) to control and develop this area. Standards for bar code symbols define the appearance of the certain bar code symbol. In other words, it determines the exact width of bars and spaces and how the data is encoded. Standards also carefully determine all other features of bar code symbols such as spots, voids, reflectivity, contrast and edge roughness. For example, there are standards for all commonly known linear and 2D symbols, such as code 39, code 128, code 49 and data matrix. (Aim, 2008) Print quality standards define bar code symbol measurement methods. Consequently, the standards are used to check whether some particular bar code symbol fulfil the requirements of its standard or not. There is one standard recognised worldwide and it is referenced in all symbol and application standards. It is called ANSI X3.182 bar code print quality guideline published by American National Standard Institute (ANSI). (Aim, 2008) Application standards are specific to certain industries. They define how technology is used to boost productivity by achieving desired scan rates. Conventionally application standards consist of a symbol based on AIM standard and print quality level based on ANSI X3.182 standards. Often application standards relate to the distribution of an item in an open system. Open systems have several parties which all are able to operate under a given standard. There is one ultimate “shipping label” standard recognised worldwide called ANSI MH10.8M for unit loads and transport packages. (Palmer, 2001: 105) 29 5. RFID RFID stands for Radio Frequency Identification and it is an automatic identification procedure. The basic idea of this system is to identify objects at a distance without requiring a direct line of sight by using radio waves. The two most talked-about components are a tag and a reader. The tag is an identification device containing the data and it is attached to the item. The reader can recognise the presence of nearby tags and read the data stored into them. The reader communicates with software called middleware, which connects the RFID system to an application, such as an enterprise resource planning (ERP) program. These main parts are described in more detail in coming sub-chapters. (Glover & Bhatt, 2006: 1) 5.1 Tag The purpose of an RFID tag is to attach encoded data to the object, since then the object is recognisable whenever necessary. Tags are available in a wide variety of shapes and sizes as well as for different usage environments. However, a connective feature for all tags is that they have some internal system for storing data and are attachable to the object in some way. Each tag must also be able to communicate with the reader on some radio frequency. This is why tags always include some kind of antenna or coil. In addition, many tags may have one or more of the following features: Kill / disable • Some tags can cease to function permanently by a command of the reader. After this the tag will never respond again. Write once • The data is permanently encoded into the tag during the factory manufacturing process and cannot be changed. If the data for some reason needs to be changed, the only way to do it is to replace the tag. Write many 30 • Some tags allow users to rewrite new data into the tag over and over again. Anti-collision • Sometimes readers may find it difficult to separate tag responds from each other if many tags are in very close proximity to each other. Thus, tags with an anti- collision ability are able to queue and respond in turn. Security and encryption • Some tags will only respond to readers that can identify themselves by giving a password. Other tags, on the other hand, are able to communicate under some encryption. Standard compliance • A tag may have been manufactured according to a standard so it is able to talk with readers within the same standard. (Glover, 2006: 55-57) 5.1.1 Tag categories Conventionally, tags are categorised into 3 types depending on their power source. A new arrival is the so-called two-way tag. In addition, the power source has major influence on the price and lifetime of a tag and partly with operational frequency impact on the read range. Passive tag obtains all of the required energy by a method of transmission from the reader and because of this it has a virtually unlimited lifetime. It does not require maintenance and is the cheapest (20 – 40 cents) but read range is also shortest, varying from a very short distance to a maximum of 10 meters, depending on the operating frequency. A passive tag requires a more efficient reader, since they have no battery. Passive tags can operate at a low frequency LF (124 kHz, 125 kHz and 135 kHz), a high frequency HF (13,56 MHz) and at an ultra-high frequency UHF (860 – 960 MHz). Different frequencies have different properties, in terms of read range, permeability of material and speed of data transfer. For example, LF and HF tags 31 form a magnetic field with the reader for communication purposes. This method of communication is called inductive coupling. In such a system the tag must be fairly close to the reader, which limits the read range of the system. Simultaneously, tags with LF are well suited for applications, where an item needs to be identified through some material but not work accurately through metal or water. LF tags can be read within 0,33 meters. Tags with HF have a read range of maximum 1 meter and tags with UHF have a read range of up to 3,3 meters, because they use radio waves instead of magnetic fields in communicating with a reader. This is called propagation coupling. HF and UHF have greater data transfer capability than LF but cannot penetrate materials as well and especially UHF tends to bounce off many objects. However, LF / HF tags are good for the identification of individual items, whereas UHF is good for pallets and shipping units. UHF is the frequency area in which the most modern systems operate. Picture 6 illustrates a passive UHF RFID tag. (RFID Journal: 2 A) Picture 6. Passive label shaped RFID tag. (Vilant) Active tag includes its own battery to power communications, processor memory and possible sensors. This is the most expensive type of tag, the price of which ranges from a couple of euros to up to dozens of euros. Conventionally, active tags operate at 455 MHz, 2,45 GHz or 5,8 GHz frequencies and have extremely long read range, varying from 20 to 100 meters. Despite of the long read range, they can be read reliably, because they broadcast a signal to the reader differently from passive tags. Active tags can even perform some activities without the presence of a reader, such as environmental measuring by an included sensor. Additional features 32 requiring power highly affect the operational time of tags, which may due to this reason be only around 10 years. (RFID Journal, 2008: 1). Two-way tag includes a battery and is capable to initiate communication with other tags of its kind without the support of a reader. (Glover, 2006: 58) A Semi-passive tag is a recent term for a tag that includes its own battery to power some functions but powers communication with the energy of the readers. The read range and price are somewhere between passive and active tags. (Glover, 2006: 58) 5.1.2 Smart labels Smart labels combine a RFID tag and bar code, as well as human-readable text into the paper label. In other words, it allows a user to encode a RFID tag with the identity and also print a bar code and / or human readable text on to the paper label. Therefore, the basic anatomy of a smart label is that the surface of the label is for a bar code and label text. The backside of the label has an adhesive coating, so that the label is attachable to the object. Thus, the RFID tag is extremely thin and sandwiched in the middle. Smart labels are currently one of the most commonly used tags in RFID applications. It is probably the easiest way to get into the world of RFID, because it can be encoded and printed on-site, based on the users´ needs. It is also reliable, because printing devices verify that all of smart labels function correctly before being attached to the item. (Kleist, Chapman, Sakai, Jarvis, 2004: 66-68) 5.1.3 Standards Some tags operate internally in some applications of a single company. Thus it is not so important which standard the company has decided to use. Some other tags must share information with partners in open logistic supply chains and in a situation like this it is extremely important to use standardised tags validated by the field of business. Generally, the purpose of standards is that different parties use 33 certain kinds of tags and related devices that partners are able to understand. In addition, many manufactures could manufacture tags according to a standard, so users are not dependent on a certain manufacturer. In the long run, standards will also drive the cost of tags down and boost their utilisation. (Glower, 2006: 71) Until recent years, the RFID industry has been driven by two different proposed standards. The first one is based on the Electronic Product Code (EPC) system that has digits to identify the manufacturer, product category and the individual item and a storing capacity ranging from 64 to 256 bits. The EPC is being developed by EPCglobal, a non-profit organisation founded by EAN International and the Uniform Code Council (UCC). The second standard is being developed by the International Organisation for standardisation (ISO). The two competing standards were limiting the worldwide adoption of RFID, as end users were reluctant to invest on either one of them, because it was unclear, which would become the leading standard in the end. No matter, that plenty of effort has been put into finding a way to blend these two standards into one. (RFID journal, B) Basically, standards in the RFID environment are actually seeding a new industry, rather than describing existing practices and technologies. (Glower, 2006: 215) One major challenge slowing down the development of new standards is the different radio spectrum regulations from country to country. EPCglobal is maintaining a list about UHF regulations worldwide. The goal is to harmonize those regulations into a range of 860 to 960MHz. For example, most of the European countries conform to the frequency range 865.6 – 867.6 MHz but China has decided to use 840.5 – 844.5 MHz, as well as 920.5 – 924.5 MHz. The USA has a frequency range of 902 – 928 MHz. (EPCglobal, 2008, A) EPCglobal defines a combined method of classifying tags that specifies frequencies, coupling methods, types of keying and modulation, information storage capability and modes of interoperability. All of its tags are intended to carry the Electronic Product Code (EPC). The different classifications of tags recognised by EPCglobal are as follows: 34 • Class 0, passive read-only • Class 0+, passive write-once but using class 0 protocols • Class 1, passive write-once • Class 2, passive write-once with extras, such as encryption • Class 3, rewritable, semi-passive with integrated sensors • Class 4, rewritable active, two-way, powering their own communication • Class 5, can power and read class 1, 2 and 3 tags and read class 4 and 5 and acting as class 4 themselves (Glower, 2006: 72) ISO has developed standards for the RFID automatic identification; item management and air interface protocol how tags and readers communicate. The standards for tracking goods in open supply chains are known as the ISO 18000 series and are aimed to cover major frequencies used in RFID systems around the world. They are as follows: • 18000-1: generic parameters for air interfaces for globally accepted frequencies • 18000-2: air interface for 135KHz • 18000-3: air interface for 13.56 MHz • 18000-4: air interface for 2.45 GHz • 18000-5: air interface for 5.8 GHz • 18000-6: air interface for 860 to 930 MHz • 18000-7: air interface for 433.92 MHz EPCglobal class 0 and 1 tended to be used in similar supply chain applications although they are interoperable, i.e. the user must have two types of devices to operate. Thus, in early 2004 there was a need to develop a new class to replace classes 0 and 1. This means that one reader could read all EPC tags. At the same time ISO was developing its 18000-6 series. Some major vendors also worked with these two coming standards and put huge pressure on the two standard organisations to merge them. It was a great opportunity to create one standard, because the new EPC class seems to be quite close to ISO’s 18000-6 series. (RFID journal, B) 35 As a result, in summer 2004 EPCglobal has announced class 1 Generation 2 (Gen2) standard and in summer 2006 ISO has amended its ISO/IEC 18000-6 type C, formally known as 18000-6C, to include Gen2 standard. In practise, this means that RFID systems using Gen2 standard are also compatible with systems made up according to ISO 18000-6C standard. End users finally have a clear vision about which type of RFID devices to invest, in order be able to operate successfully in global open supply chains. (RFIDUpdate, 2006) As a summary, Gen2 is a standard which combines EPC classes 0 and 1 and ISO 18000-6C standards. It defines the physical and logical requirements for a passive, reader (interrogator) talks first (ITF), RFID systems operating in the UHF (860 – 960 MHz) frequency range. RFID systems in this context mean transactions between tags and readers allow the maximum read range of up to 10 meters. (EPCglobal, 2007, B) It must be highlighted that there are no standards for active and HF tags. Each supplier of active tags has their own standard and related devices, which are not compatible with systems of other suppliers. In addition, EPCglobal is working on an HF standard but it is still in process. 5.1.4 Information storage capacity RFID comes in a wide range of storing capacity. The simplest tags store only 1 bit. These tags can only recognise the absence or presence of the tag and can not identify individual items. These are conventionally used in libraries and clothing stores to prevent theft. Some tags may store kilobytes of data but typically a tag carries no more than 2 kilobytes of data. That is enough to store some basic information about the item the tag is attached to. Larger data storage capacities require active tags. The higher the data storage capacity, the more the tag costs. In order to reduce cost it is recommended to store only an identifier on the tag and look up rest of information on a database. (Glower, 2006: 67-68) 36 As mentioned earlier, the EPC tags have a user memory ranging from 64 to 256 bits, which allows user-specific data storage. For example, commonly used EPC standard tags have 96 bits of user memory, which can store 24 HEX-digits. HEX digits consist of numbers 0-9 and letters A-F. Generally, 1 digit equals 4 bits, so a tag with 256 bits can store 64 digits. (Lahtinen, 2008) 5.1.5 Tag protocol A tag protocol is a set of formal rules describing how to transmit data between readers and tags. Some important terms related to tag protocol are sinqulation and anti-collision. Singulation refers to a procedure that reduces a group of things to a stream of things that can be handled individually. Anti-collision is a term that describes a set of procedures to prevent tags from interrupting each other and talking out of turn. (Glower, 2006: 77-78) The ways in which readers and tags communicate can roughly be categorised as tag- talk-first (TTF) high-end active tags or reader-talk-first (RTF) smart labels and other passive tags. However, it would be simplest if tags arrived on the scene to announce their presence to all involved. In practise, this will cause reading problems, unless tags are able to speak taking turns, instead of all speaking at the same time. This is why RTF protocols are preferred. The most common of these protocols are: Slotted aloha With this anti-collision protocol tags begin to broadcast their ID´s as soon as they have arrived at the read range of the reader, then they are able to obtain energy from the reader signal to energise them. The reader only receives the signal and does not reply in any way. This is simple and fast but unworkable with more than 12 tags. Thus, adding even some concept of sinqulation and requiring tags to broadcast only at particular time slot remarkable cuts the chances of a collision. Variations of slotted aloha is used for ISO 18000-6 type B and Gen2 RFID types of tags. 37 Adaptive binary tree EPC class 0 and 1 UHF tags use this sinqulation and anti-collision protocol. In this protocol a binary search is used to find one tag among a bunch of tags. At first the reader sends a query asking, “Does any tag have an ID beginning with a bit 1?” Tags that answer “No” then step out of conversation, whereas tags that answered “Yes” are asked similar question about the next bit. The tags are narrowed down until only one tag is left. Slotted terminal adaptive collection Part of the EPC specification for HF tags is described with the abbreviation STAC. This protocol is especially suitable for the sinqulation of a large tag population, because it provides up to 512 slots. A group of tags or a single tag is selected based on matching lengths of tags with an EPC code beginning with the most significant bit (MSB) and ending in the least significant bit (LSB). Since an EPC code is organised by header, domain manager number, object class and serial number from MSB to LSB, this protocol can easily select tags belonging to some particular group, such as a certain object class. EPC Gen2 The protocol has three alternative ways for communication between readers and tags. A reader may select tags by asking them to compare themselves to each other. A reader may inventory tags by sinqulating them, until it has recognised each tag within range. A third way to communicate is to access tags. That includes reading stored data, writing new data and killing or locking some memory sections of the tag. This protocol also allows devices to operate both under European and USA radio frequency regulations, which is impossible for EPC class1 gen1 devices. (Glower, 2006: 87-96) 38 5.2 Reader The purpose of RFID readers is to recognise the presence of nearby RFID tags. Typically, the reader transmits signals and tags inside the operating range pick up the signal. The signal is sufficient to power the semiconductor chip inside a tag, which stores the identity of the tag. After this, the tag returns the identity to the reader. This is only one way, in which readers and tags interact and some others may work in slightly different ways. Readers are available in various kinds of shapes and sizes and can be found in stationary as well as portable handheld selections. In addition, readers are devices that connect tags to a network. (Glower, 2006: 36-37) Readers for passive tags cannot recognise active tags and vice versa, because in passive systems readers talk first, whereas in active systems tags talk first. Thus, transmitted and received signals do not come across in an appropriate way. Therefore, if passive and active tag systems are needed to run simultaneously, a double number of devices must be purchased. Physical parts of the reader are antenna subsystem, controller and network interface. An antenna subsystem enables interaction between reader and tag. Some readers may have only one or two antennas, one to transmit and one to receive signals, whereas other readers may use many antennas at remote locations. A controller implements communication protocols and controls the transmitter and also determines when information read is worth sending to the downstream of network via middleware. A network interface enables readers to communicate with middleware or other devices. Furthermore, readers have four internal functions to perform within a controller, which is capable to operate with tags and middleware. These are application programming interface, communications, event management and antenna subsystem. 39 Application programming interface (API) • Creates messages and parses received messages from middleware. For instance, parsed messages might be a request for tag inventories, monitoring the health of the reader or control configuration settings, such as power level. Communication • Handles details of communication, made up by the API, over any transport protocol the reader may use to communicate with middleware. Event management • Most of the time many tags are visible to a reader. This is called observation and observation, which differs from previous observations, is called an event. Event management is a tool, which defines events and further determines which of the events are valuable to send forward to the middleware. Antenna subsystem • This component must implement tag protocols and it consists of the interface and logic that enables the RFID reader to interrogate the RFID tags and controls the physical antennas. (Glower, 2006: 108-110) 5.2.1 Layout for readers and antennas Since the purpose of RFID systems is to sense the presence or absence of items, the environment dictates the details of any installation. Possible variations are infinite but the most typical layout for readers and antennas is portal. Thus, an RFID portal is an arrangement of antennas and readers designed to recognise items with attached tags entering or leaving through a doorway. This is widely used in warehouses and factories, where items move between different sections of the factory. A new much talked-about application is called smart shelves. These are shelves with antennas and readers can recognise the arrival and departure of items from the shelves. These kinds of shelves are capable to do inventories on demand and also match item IDs against a database to find oncoming expiring dates, for example. (Glower, 2006: 113-116) 40 5.2.2 Reader protocol A reader protocol is a set of formal rules defining how one or more readers and hosts communicate within a network. The host can be either application or middleware. Communication includes alerts and observations from a reader to a host and commands from a host to a reader. An alert is a message with information on the condition of a reader. An observation is a record about a movement of tag at some point of time and place. A command is a message that causes a reader to perform some actions, such as read or write tag information. These methods of communication are the basis for all reader protocols. In addition, communication can be either asynchronous, meaning that a reader can contact a host any time or possibly at scheduled times or synchronous, meaning that a reader waits until a host sends a command requesting pending, as well as immediate information. (Glower, 2006: 119-122) In the past standardisation has mostly focused on tags and tag protocols and not much attention has been paid to reader protocols, this was the result in many proprietary protocols developed by vendors. Recent years have witnessed plenty of ongoing work to set up global standards for reader protocols. For instance, EPCglobal and Internet Engineering Task Force (IETF) were developing separate standards. (Glower, 2006: 122-125) However, IETF gave up the developing of a standard called The Simple Lightweight RFID Reader Protocol (SLRRP), whereas EPCglobal ratified a standard in spring 2007. This standard will most likely become a global one. EPCglobal reader protocol In April 2007 EPCglobal announced a standard called the Low-Level Reader Protocol (LLRP). It is called low-level, because it provides control of RFID air protocol operation timing and accesses to air protocol command parameters. LLRP specifies an interface between reader and host. In practise, this ensures some basic cross-compatibility among readers and host devices provided by different vendors. In addition, they are able to extend the protocol to vendor-specific features, in order 41 to tailor devices to match various requirements of RFID end users. LLRP also provides maximum support to the earlier published tag-to-reader Gen2 protocol. (EPCglobal, 2007, C) LLRP consists of three layers. These are reader layer, messaging layer and transport layer. Below is a structure and short explanation about its functionality. 1. Reader layer: defines the allowable content and format of messages sent between the reader and host. This layer is divided into four subsystems. 1. Read subsystem: responsible for reading tags and supplying tag information to the event subsystem. It also determines the timing of reads and filters valuable read, which differentiates a given pattern set by a host, from non- valuable reads to be submitted to event subsystem. Read subsystem cannot tell the difference between newly arrived and already detected tags. 2. Event subsystem: responsible for turning reads into meaningful events, such as arrival of a new tag or the absence of a tag that was read previously. 3. Output subsystem: decides what data the reader will report by comparing data against the filters set by the host. Then it buffers the data until the host has requested it or send it at scheduled times (trigger). 4. Communication subsystem: responsible for implementing reader layer to a combination of messaging and transport layers. It also packages and translates the data to comply particular transport layer requirements. 2. Messaging layer 1. The layer provides three messages channels, a basis for all reader protocols, such as commands from host to readers and alerts and observations from reader to host. 3. Transport layer 1. The layer controls transports between a reader and a host. (Glower, 2006:125 – 133) 42 5.3 RFID Printer for smart labels RFID printers include a RFID reader function, as well. The reader can both read tags and write tags that allow writes. Consequently, RFID printers are devices that make smart labels by encoding tags and printing bar codes, as well as human-readable texts to paper labels that house the tags. RFID printing devices can also include an apply device. In a low volume application an operator is capable to manually attach smart labels to the item but in high volume applications this must be done automatically. Thus, the printer can be considered to consist of four separate functions. There is the reader that is already familiar to us and the printer, which is no different than any other bar code printer. In addition, there is a verifier and an applicator. The purpose of a verifier is to check that both RFID and bar code are functioning properly. The same reader, who wrote the tag performs the RFID verification. Bar code verification is performed by an optical scanner located just after the printer. If the tag fails it will be discarded, before it is attached to the object. An applicator is a function that attaches the smart label to the passing object. This can be done either with the a wipe-style, in which tags peel off a roll directly on the items or pad-style, in which has a pneumatic arm to press tags to the items. In bar code systems, typical applicator solution blows the tag onto the object with compressed air but it is not convenient with RFID, because it may damage the electronic components inside an RFID tag. However, it is used in the small scale. (Glower, 2006: 111-112) 5.4 Middleware Middleware refers to a server with software that is located between readers and enterprise applications. The purpose of middleware is to filter data from readers and pass on only the useful data to enterprise applications. In addition, middleware can 43 also be used to manage readers and query RFID observations within a network. Thus, RFID middleware consists of application-level interface, event manager and reader adapter. In practise, companies utilising RFID have readers from one or more vendors. Each vendor has specified a reader interface that should be adjusted to the enterprise application separately. This awkward feature creates the need for a reader adapter, which provides means to eliminate differences of reader interfaces and expose a single interface to enterprise applications. Readers make a huge amount of observations every second. Because of sheer volume of the data, it must be further processed to be meaningful to enterprise applications. For example, owing to the physics of radio frequency communications current readers operates at a 80 to 99 % of read rate. If there is more than 100 object to be identified, at least one of them will be missed in every read cycle. Passing that information directly upstream will cause continuous fluctuation to the inventory level, which further loads down the enterprise application and may even bog it down. Therefore, there is the event manager to control raw observations coming from the readers and lets only the application-relevant ones pass. Middleware also has an application-level interface that allows the enterprise application to manage readers, such as set up event processing methods to filter data and make queries about filtered RFID observations. EPCglobal has developed a standard called Application Level Events Specification (ALE) for that purpose. In practise, the ALE standard allows an application to describe what information they are interested in and how they wish to receive it, without worrying about the physical RFID infrastructure. In addition, one of the most significant benefits of ALE-compliant middleware is expandability to interface with devices other than RFID readers. The ALE standard also separates the interface from implementation, which means implementation details are left to the vendors. For example, this approach allows the vendor to decide on technology platforms, deployment options and add-on features. (Glower, 2006: 137-140) 44 6. HOW TO DETERMINE INVESTMENT ADVISABILITY This chapter introduces two methods that can be used to evaluate how well or how poorly a planned investment will turn out. Since the current process exists for a reason it is absolutely important to figure out whether the investment actually improves the process and is reasonable in financial terms, instead of just installing it without adequate insight on the results. 6.1 SWOT analysis SWOT analysis is a simple framework used to evaluate the internal aspects of the company as Strengths or Weaknesses and external situational factors as Opportunities or Threats involved in a project or in a business venture. SWOT is applicable in both corporate level and business unit level and especially popular for marketing plans. The first part of any SWOT is to collect a set of key facts about the company and business environment with regard to the overall objective. The second part of the analysis is to determine whether they constitute one of the following: • Strength is a resource or capacity that the company can take advantage of to achieve its objectives. • Weakness is a limitation, fault or defect that prevents the company to achieve its goals. • Opportunity is any favourable situation in the business environment, which may assist to obtain goals. Usually, it is new technology that can be taken into use or a trend or change in customer behaviour, which increase demand. • Threat is any unfavourable situation in the business environment, which has the potential to harm, constrain or cause problems to achieving the goals. Since the classification has been completed and items prioritised within classes, it is possible to consider interactions of items with regard to the objective and decide 45 whether it is attainable or not. In case the objective seems attainable, the outcome of the analysis can be used in forming possible strategies. In general, an effective strategy is one that takes advantage of the opportunities of the company by emphasising its strengths. In addition, an effective strategy avoids threads and corrects or compensates for weaknesses. The picture 7 illustrates SWOT analysis. (Kotler, 2006: 52-56; NetMBA) Picture 7. Illustrative diagram of SWOT analysis. (Wikipedia) 6.2 Pay-back time calculation Pay-back time is the period of time it takes until positive cash flow generated by the investment covers the negative cash flow. This is a very simple method and particularly suited to the initiate stage of the project when cost and savings are only outlined. In addition, it is good guide at any given stage of the project when modification is suggested. Appropriate pay-back time is naturally specific to context and is calculated as follows: savingsAnnual InvestmentofCost = Pay-back time 46 This method is very simple and widely used. However, it has one basic deficiency, because it does not automatically pay attention on interest rate. Naturally this is easy to fix, because annual net savings can be discounted to the present value of investment moment by applying the market interest rate. Then it is possible to observe, how many years of discounted savings are needed to cover the cost of the investment. According to the pay-back method, the investment becomes even more advisable the faster it is able to cover the investment cost. (Aitken, 2000: 105; Neilimo, 2005: 223) 47 7. PROCESS DESCRIPTIONS As mentioned in the introduction, the purpose of this study is to figure out whether RFID could improve given processes of Metso Paper Service. The Service falls into many sub-processes but only some of them come in question when searching for ones which will benefit the most from RFID. Therefore, the processes eligible for process description phase must conform to some of the following criteria, in order to be potential applications for RFID. The criteria are great volume, critical process for customer, feasibility, cost savings and increase of sales. • Great volume criterion stands for high item turnovers with remarkable financial value. • Critical process for customer means that output of the process is extremely important for customers and failures may cause unplanned shutdowns. • Feasibility criterion means that, in theory, RFID is possible to set up for the process or it could simplify things. • Cost saving criterion means that RFID would make the process more effective, faster or reduce routine work or risk for identification errors. • Increase of sales criterion refers to a hypothesis that RFID would provide new data about product life cycle, which could be used to support sales. Based on the above criteria, the following processes are eligible for the process description phase. These are: warehouse operations, direct delivery process for Sizer consumables and Doctor Blades, consignment stocks for Doctor Blades, spare part packages, return and repair orders, roll coatings and roll workshop. Data was collected by interviewing key persons of the processes and other people involved, in order to understand how each process works and to discover problems and reasons behind these problems. The interviews were conducted by using semi- structured theme questionnaires tailored to suit a particular process. It was not possible to make only one set of questions relating to every process, because they 48 are so different. Basically, the following questions were asked concerning all the processes: • At what stages of the process and how identification is conducted at the moment and what information is collected? • What additional information would be useful to collect? • What kinds of problems are faced with the current practice? • What stages (routines) of the process would be good to automate? • Does automatic identification provide any development of the process? However, this method allowed detailed answers and it was possible to get some examples about the problems and further needs for development. In addition, process descriptions were drafted to make them more understandable. At the beginning, every eligible process is broadly presented and its potential for RFID stated. Then process characteristics and data transfer actions are described, ending up with stating possible problems and needs for additional data collection. Finally SWOT-analysis has been carried out to evaluate how RFID could influence the processes in operational, economical, technical and organisational ways. Siegel sign-test is used to indicate the magnitude of items in SWOT. Two plus signs (++) mean great positive magnitude, whereas two minus signs (--) mean great negative magnitude and single signs are mean values of those limiting ends. Based on the SWOT, the number of processes is narrowed down at the end of this chapter by discarding processes, which do not seem to have potential for RFID at this point of time. 49 7.1 Warehouse operations Metso Paper has four logistic centres for spare part deliveries. They are located in Kerava, Sundsvall, Beloit and Shanghai. Until spring 2008, one centre in Finland was located in Jyväskylä but that has been out-sourced and moved to Kerava. It handles spare part deliveries to customers in Finland and Europe but supports other market areas, as well. Beloit and Shanghai mainly focus only on their own continents. The logistic centre in Finland is described first. Its average volumes in 2007 were about 2850 received and 3900 shipped lines per month. The higher rate of shipped lines is explained by the fact that a received line may include many units, which spread out to several sales orders. The logistic centre always keeps 3000 of its highest circulating or critical items in stock, whereas rest of the items will be purchased-to-order and may have very short turn-around times. Since the volumes are great, there is a need for efficient material handling. Thus, it is topical to go through how different operations are currently managed and consider whether they might perform even better with the help of RFID. Appendix 1 displays warehouse operations. 7.1.1 Characteristic Receiving The purpose of receiving is to receive incoming goods, check their quantities and general quality and attach Metso item labels, measure weight and place them on storing locations, as well. Suppliers tend to mark shipments in their own style and the corresponding Metso purchase order can sometimes be very challenging to interpret. In the perfect world supplier packing lists indicate the Metso PO number and item codes and items themselves have the codes as well. In case the items are marked appropriately, the identification process takes an average of 3 minutes. 50 Bar coding is not used but it is notable that there are devices for bar code utilisation. This would require a paper called “a receiving note” to be printed out for every single PO in advance to indicate bar codes, because suppliers are not capable to mark them. No wonder employees find it inconvenient to use and carry out receiving in the traditional way. Receiving is manually entered into Baan, which is simple, as well, but makes employees go in and out the office during the operation. The receiving note is in Appendix 2. In this stage Metso item labels is attached to items, which makes them recognisable. The label is in picture 6. It indicates the item code, date, and possible SO number and storing location. Most of all, the item code is expressed in the form of a bar code (code39) as well, which will boost forthcoming warehouse operations. Figure 6. Item label. Shelving and stocktaking Shelving is very simple and fast thanks to bar codes. If a recently received item has no existing storing location, the employee can store it wherever s/he considers best. In that case, s/he reads bar codes from the location and item with a handheld mini- computer, which includes bar code reading / writing applications and types the quantity. The logistics centre has on-going stocktaking, the purpose of which is to 51 go over some predetermined locations every week. Inventory records, printed on paper from Baan, are compared against actual items at the location and differences can be fixed right away with the mini-computer. Bar codes have improved these two operations the most. Short reading distance is not a slowing factor, because stocktaking is based on visual count. Picking and packing The Baan automatic application prints the picking list a couple of days before the planned delivery date, but only if items are on hand. This is the input for delivery actions. The picking list is in Appendix 3. It indicates the SO number, items and storing locations, quantities, delivery date and the customer. Packers print sales order labels for items, which also include customer item codes if sales have inserted those into Baan while entering the order into the system. After items have been collected and packed, the SO must be updated in Baan. This is done manually, the function erases allocated inventories per items associated to that particular SO and prints the address label. The packer attaches it onto the package and delivers it to the dispatching area. The address label is in picture 8. Picture 8. Address label. 52 Picking lists and item labels have bar codes and updating the SO is also possible with a mini-computer, but practise has revealed this impractical. The process works manually, as well. In addition, usually only few items need to be picked up for the SO and physical packing itself definitely takes most of the time, so employees have not considered bar coding attractive at this stage. Shipping and Forwarding The input for forwarding actions is the packing list, which is to be printed out whether domestic or export forwarding at the time the packer has updated the SO in Baan. The packing list is in Appendix 4. Forwarding in Kerava handles domestic shipments and export forwarding in Jyväskylä takes care of international shipments. After delivery has been booked, a waybill number is entered into the packing list, which will be printed out again. Forwarding at Kerava attaches the list on the package and it is ready for courier. At this stage operations are carried out in the Baan system. 7.1.2 Problems Bar code devices exist for warehouse operations but are not widely utilised. Devices have mainly been set up for the existing way of operating and do not provide actual benefits over traditional ways of operating. Especially in receiving operations bar codes have a lot of potential to speed things up. But receiving cannot be completely carried out with handheld mini-computers in the receiving area. In any case, employees have to print item labels and “the receiving note”, go to the office and pick them up. Printing the receiving note is an additional task compared to traditional receiving. Basically, it may help if the printer is re-located close to the receiving area and the mini-computer would have a wider screen for added ease of use. It would significantly help, if suppliers could set up bar codes for Metso item codes. However, the target is that new receiving employees in Kerava begin to use mini-computers, but time will show if that proves successful. 53 Sometimes items cannot be received based on packing documents, because items are marked insufficiently. Then the supplier must be contacted. Unfortunately, this tends to take a long time, because difficult cases are usually postponed until other shipments have been received. Occasionally, receiving employees ask the associated purchaser to figure things out, but it may not be any faster. Basically, warehouse operations run fine. 7.1.3 SWOT The purpose of this SWOT is to assess how RFID would change warehouse operations. The premise is that bar code devices are replaced with RFID devices and every item has an RFID tag. After this all operations are performed with the help of RFID. Strengths ++ Passive RFID system is convenient, because there are no needs for tags to continuously send signals to readers as active ones do. ++ The reading from a distance feature would be speed stocktaking up to a degree. But it cannot totally take the place of visual observation, because every unit cannot have an RFID tag. Weaknesses -- In practice it is impossible to attach a RFID tag to every item, such as small ones like screws, springs and nuts, which can be sold individually. -- Generating and printing of bar code symbols is basically free of charge but convenient passive RFID tags costs 10 – 20 cents. -- Impossible to use in receiving until suppliers could fit items with RFID tags. At the moment they cannot do that even with bar codes. -- Very low benefits in terms of time savings for picking and packing, due to the fact that the SO typically consists of few items and the packing itself takes most of the time. -- The re-write feature is wasted, if tags are not erased during the shipping phase and then re-used. 54 - Generally warehouse operations work fine the way they are and there is no need to change them. Opportunities ++ Automatic receiving, since suppliers could fit items with RFIDs. ++ If it would be possible to attach a tag to every unit, stocktaking would happen automatically within seconds and inventories would be accurate all any time. + Possible to use RFID only in some areas of warehouse operations. Threads -- Bar code devices are not widely taken advantage of, this could happen to RFID, as well if employees do not see any significant operational benefits. It seems RFID is not attractive for warehouse operations, because it has some remarkable weaknesses, which decrease its feasibility. Currently bar codes are used to a degree, but they have faced some reluctance hampering the ultimate breakthrough, because they have not benefited operational work much. Besides, it is quite obvious that RFID would be extremely difficult and expensive to set up successfully and would not offer additional improvements over bar codes in that kind of a warehouse environment. Since the suppliers could fit items with tags, RFID becomes extremely useful, but this does not seem to be the case in the short term. Because of this it is my opinion to keep focusing on bar codes at this point. 7.2 Direct delivery processes Metso has considered it reasonable to always ship some parts directly from the supplier site or production unit to customers instead of taking them to a logistic centre first and then shipping them forward to the final destination. This is called direct delivery process. Sizer consumables and Doctor Blades are delivered that way but practices are not identical. The purpose for checking direct deliveries is to figure out whether it occasionally happens that parts are shipped but not invoiced, because the SO is not updated in Baan. Additionally, volumes for both products are great and 55 parts critical for customers, so it is worth considering whether RFID could benefit the deliveries. 7.2.1 Sizer consumables Sizer is a coating unit at the end part of paper machine. It plays an important role relating to the coating and surface sizing of paper to produce excellent quality. Sizer has wearing parts, which are generally categorized as Sizer consumables. These consist of rods manufactured by Komas in Jyväskylä, and rod beds manufactured by Finn-Valve in Joutseno. These are always made to order. Parts are critical for customers, because the paper manufacturing process must be shut down, if parts run out. Rod and rod bed are illustrated in picture 9. Picture 9. Rod and Rod bed locations in coating unit. 56 These two parts have high rates of item turnovers. For example, volumes at 2007 were 3683 units of rods and 2685 units of rod beds. These built up 625 shipments, 313 for rods and 312 for rod beds. The length of Sizer consumables vary between 3,5 – 11,5 meters, depending on the width of the paper machine. The parts are used together, but it is not necessary that parts are combined to each other, until they reach the customer’s site. Since the length of parts is great, and suppliers have good quality control, it is reasonable to deliver them directly from production sites to the customer’s site, in order to reduce costs and speed up the delivery. 7.2.1.1 Characteristics The direct delivery of Sizer consumables includes many steps and seven participants. The process is pretty much identical with both suppliers and is presented in Appendix 5. The sales receive the customer purchase order and create an SO in Baan, then items show up on the purchase queue. Purchasing creates a PO into Baan and sends it to the supplier by email or EDI. At this stage, only typical sales information, such as customer PO, SO and item identifiers and quantities are transferred between participants. Basically, Baan controls the whole process according to delivery date set by sales. The next impulse, which moves things forward, is notice of readiness of the PO. The supplier emails it to forwarding, after he has finished the parts and the delivery process begins. The notice contains appropriate order and packing information, such as the PO number, weight and size. If forwarding does not receive the notice, the parts will never be shipped and invoiced and the process stops. Besides, suppliers do not get paid, until Metso has updated the PO in Baan. Thus, it also in the supplier’s interest to ensure the process continues. Based on this notice, forwarding receives the PO into Baan and updates the SO that launches invoicing, arrange transport and sends packing documents to supplier by email. Parts themselves stay at the supplier site all the time. The packing documents are important for transport, but also for identification. Packing list and item labels also indicate items with bar codes (Code 39). 57 Unfortunately packing documents include bar codes only for the Metso SO number and part numbers, not for customer ones. However, bar codes exist but there is no need to use them in direct delivery processes, but theoretically customers can use Metso bar codes in receiving. After the courier has picked up the delivery, its journey can be monitored on the courier’s web page, where the actual shipping status is updated in real time, thanks to the courier’s internal bar code system. Normally, Metso checks the shipping status only, if the customer has complained about a late deliver,, although the goods left on time, or if the shipment is very critical for some reason. Parts are very sensitive and require careful packaging to avoid damages during transportation. This is why sturdily built plywood boxes are used. The boxes are relatively expensive and Finnish customers are requested to return them. The return address is painted on boxes permanently and the supplier pays the shipping fees. 7.2.1.2 Problems Possible problems mainly relate to customer orders, which are requested with too tight a schedule compared to parts lead times, because the customer is running out of parts. Currently, there is no way to follow up customer consumption or inventory levels. For this reason it has been suggested that it would be useful to be aware of it in some way and be able to make refill reminders. Sometimes product updates cause problems, if they are not updated into Baan correctly or the customer has not informed, when he has started to use a revised version of the part. However, the direct delivery processes themselves work fine and without any problems. 7.2.2 Doctor Blades Doctoring in paper manufacturing process means removing all kind of contamination such as fibres, water and fillers from the surface of rolls and prevents a sheet from wrapping around a roll causing serious damage. For this reason, there has to be a doctoring device for each roll in the machine. In general, optimised doctoring will result in better runnability and easier maintenance of the machine and 58 ensures even paper surface and further improves paper quality. Doctoring devices consist of many parts, but one of the most critical parts is a Doctor Blade. It is a wearing part, because it touches the roll and therefore needs to be replaced regularly. At 2007 Metso delivered around 36,000 units of Doctor Blades. The delivery unit is 10 pieces, which equals 3,600 transport cases. An average length of Doctor Blades is about 7.5 meters, but blades can be rolled up in order to fit a transport case in size 90 x 90 x 10 cm, which enables economical delivery. Doctor Blades are always made to order, because they must be specific for certain rolls and usually rolls tend to differ from each other. For example, distinctive factors are the width of the roll, lo