1. Introduction

The contactless Radio Frequency Identifcation (RFID) technology holds promise for object tracking, supply chain, smart payment and security management. At the early of this year 2004, Gillette purchased 500 million radio frequency ID tags from Alien Technology at a reported cost of less than 10 cents each for its shavers and related products. The world's largest retailer, Wal-Mart, and the world's largest household products manufacturer, Procter & Gamble, are also planning to run a field test of smart shelves that can read radio frequency waves emitted by tags embedded in millions of cosmetic and related products. In the last year, Visa has already announced that it plans to use smart cards fitted with special tags to allow its customers to conduct transactions such as buying a soda without having to fish for change or swipe a credit card. While the market for RFID chips is still small now, its potential is of signi¯cant value and the market for RFIDs is estimated to be $700 million in revenue and is expected to grow to$2 billion by 2007, according to research firm VDC. In this May, seven companies, including In¯nite Power Solutions, Cymbet, Graphic Solutions, KSW Microtec, Parelec, Cryovac and Power Paper, announced the formation of a nonprofit organization called the Smart Active Label Consortium to promote the benefits of RFIDs and develop standards to push their adoption.(www.news.com)

Compared with barcode, the signifcant advantage of all types of RFID systems is the noncontact, non-line-of-sight nature of the technology. Tags can be read through a variety of substances such as snow, fog, ice, paint, crusted grime, and other visually and environmentally challenging conditions, where barcodes or other optically read technologies would be useless. RFID tags can also be read in challenging circumstances at remarkable speeds, in most cases responding in less than 100 milliseconds. The read/write capability of an active RFID system is also a signi¯cant advantage in interactive applications such as work-in-process or maintenance tracking. Though it is a costlier technology, RFID has become indispensable for a wide range of automated data collection and identification applications that would not be possible otherwise.

RFID is a proven technology that's been around since World War II, however, due to issues like cost and performance, it has not achieved mass-market applications. While a RFID system design largely depends on its applications, the key criteria in evaluating RFID systems are the lowest cost, consistence with standards and regulation and satisfying performance. Also the security and privacy issues have received many concerns recently. In this report, several issues regarding RFID system design are discussed. A reference block diagram of the firmware design is proposed and some future research directions of RFID are discussed at last.

2. RFID System

A typical RFID system consists of three components: an interrogator (RF transceiver), a transponder (RF tag) and a back-end application system. In RFID systems, antennas are the conduits between the tag and the interrogator, which controls the systems data acquisition and communication. Often the antenna is packaged within a interrogator, which can be configured either as a handhold or a fixed-mount device. The interrogator decodes the data encoded in the transponders integrated circuit (silicon chip) and the data is passed to the host computer for processing. RFID transponders are categorized as either active or passive. Active RFID transponders are powered by an internal battery and are typically read/write, i.e., transponder data can be rewritten and/or modified. The transponders are extremely low functionality devices that communicate with readers only and are unable to detect the communications from other transponders. The transponders do not actively send out a communication signal. Instead, they use either load or backscatter modulation to communicate with the interrogators. Furthermore, the transponders perform operations only under the direction of the interrogators. The information obtained from the transponders is transmitted from the interrogators to an application system, such as an inventory management system or a quality control system [1-4].

The interrogator-transponder communication in RFID systems uses a distinct localized communication medium, where interrogators are locally communicating with transponders and not other interrogators. This is di®erent from ad hoc wireless networks, where devices communicate with other devices and act as communication routers [5].

3. Physical Layer Design

The main components of a RFID system are the interrogator, transponder and their antennas. In a typical communication between a interrogator and a transponder, the interrogator emits a continuous radio frequency wave, which may be modulated or not. When a transponder enters the interrogation zone, the RF field of the interrogator, the transponder receives energy and possible instructions from the field. After received su±cient energy and information, the transponder resonates and modulates the carrier signal according to the stored data and sends it back to the interrogator. The interrogator then demodulates and decodes information from the received signal and send it to the back-end application server. A complete block structure of a RFID system without application server is shown in Fig. 1.

In the physical layer design of a RFID system, the main issues are regarding how to efficiently transmit energy and signals between interrogator and transponder, how to modulate/demodulate and code/decode information on RF sine wave, and how to reduce the cost of transponders and interrogators while keeping robust performance.

3.1. Antenna Design Issues

Most RFID systems operate at 120-140KHz(LF), 13.56MHz (HF) or 902-928MHz(VHF) frequency band. As shown in Fig. 2, there are two classes of antennas, one based upon inductive coupling and one based upon propagating coupling. Passive transponders typically work at low frequency using inductive coupling. Most active transponders work at high frequency using propagation coupling. As the frequency for RFID system rises into the microwave region 2.4-5.8GHz, the problem of antenna design becomes more acute. The issues in antenna design are the choice of antenna type, the impedance matching issue, radiation patterns design, range and so on. Usually, the antenna for RFID must be small, cheap and robust enough, have good coverage and can maximize power transfer [6, 7].

When a RFID system works at a low frequency (13.56MHz or lower), the coupling element normally consists of a coiled antenna and a capacitor. Usually, inductive coupling works within the near-field, the radian sphere of ¸\lambda/(2\pi)of the electromagnetic signal, where \lambdais the carrier wavelength. Here, the quality factor, Q, of the coupling element defines how well the resonating circuit absorbs power over the relatively narrow resonance band. Though Q value is demanded to be reasonable high, when Q value is too high, it may result in a too narrow resonance band, which can cause the distortion on received signals.

When the operation frequency is high, the type of antenna can be chose from dipole, folded dipole, printed dipole, printed patch or log-spiral antenna. The gains and radiation patterns of these antennas are different and effect the range of the RFID system. For a RFID system, the radiation pattern of both interrogator and transponder can be seriously distorted by other objects around. A directional antenna is normally used since it is more robust and resist to the distortion, compared with a omnidirectional antenna. Furthermore, for maximum power transfer, the input impedance of the following ASIC should be matched to that of the antenna.

3.2. Signal Processing In Interrogator Design

In a RFID system, the transponder resonates and modulates carrier signal from interrogator. Having received the backscatter or load modulated carrier signal, the interrogator needs to demodulate and decode the information [4].

In the near field, the modulation in transponder is achieved with changing the impedance of the transponder. It is called load modulation. In the far field, the modulation is transponder is done by changing the cross-section of the transponder antenna. It is named backscatter. In either load modulation or backscatter, the choice of modulation and coding scheme is based on signal power spectrum, signal bandwidth and bit-error rate (BER). Both modulation and coding schemes can change the spectrum of modulated signal. (The coding schemes discussed here are generally called modulation coding or line coding.) The popular modulation schemes are AM, FSK and PSK. AM is widely used in load modulation at HF band. PSK is normally used in backscatter modulation. The most common coding schemes for RFID, as well as for other wireless communications, are NRZ, Manchester and differential coding [7].

In Fig. 3, an example of the block diagram for demodulating QPSK signal is presented. In actual systems, it may be simpler than this if BPSK modulation is used. The received signal by interrogator passes through a band-pass filter to boost received signal-to-noise ration (SNR). Then, it is mixed into two baseband signals using IQ orthogonal demodulation followed by the decision block. The carrier signal and symbol timing can be recovered from received signal using phase-lock loop (PLL). But, there is another problem. since in the reeived signal from the transponder is normally mixed up with the exciting signal from the interrogator, there must be a filter to cancel the exciting signal [8]. It may be called Exciter Cancellation and it may be implemented within the band-pass filter or the low-pass filters.

3.3. Transponder Design

The choice of transponder has great e®ects on RFID system design and applications. Normally, a complex transponder means it has more programming option, more functions and more security, compared with a simple passsive tag used in supermarket. However, increasing complexity of the transponder will in°uence the cost of both transponders, reader and programmers. Normally, it means more expensive, much bigger and harder to maintain. Hence, when designing a transponder, there are lots of factors and trade-offs to be considered. A block structure of transponder is shown in Fig. 4. It normally consists of four parts: antenna, power supply, control and memory.

Based on how a transponder is powered, there are two kinds of them: passive and active. For passive RFID transponders, it needs to consider, how to utilize the active energy, how to resonate the carrier signal and how to modulate and encode the resonated signal. For active transponders, how to design a battery is also a big issue. Having a battery or not can affect the range of a RFID system [4]. Generally, passive transponders are cheaper than active transponders.

The control part manages all the data and control activity in transponder. It controls the switching between local battery and absorbed energy, obtains synchronization and timing information from received signal, modulates the impedance or the cross-section of the transponder antenna, provide security and privacy protection on the data stored in memory. It basically like the central processing unit (CPU) in computer systems. Security control and privacy control are two important sub-blocks with in it.

There are three kinds of memory can be used in transponder: read-only memory (ROM), random-access memory (RAM) and non-volatile programmable memory. The ROM-based memory is used to accommodate security data and the transponder operating system instructions. The RAM-based memory is used to facilitate temporary data storage during transponder interrogation and response. The non-volatile programmable memory may take various forms, electrically erasable programmable read only memory (EEPROM) being typical. It is used to store the transponder data and needs to be non-volatile to ensure that the data is retained when the device is in its quiescent or power-saving "sleep" state.

4. Anticollision Issues In MAC Sublayer Design

Now we discuss a RFID system that consists of multiple transponders and interrogators as in Fig. 5. Each interrogator works with the transponders in its interrogation zone. These interrogation zones may have some overlaps. When there are more than one transponders communicating more than interrogator(s) at the same time or using the same frequency, there will be various collision happened among transponders or interrogators. The collision among the transponders is called Multi-transponder Access Collision. The collision among interrogators is called inter-interrogator collision. This problem is very similar to the multiuser access interference (MAI) problem in mobile communications and the collision problems in computer networking. It is a channel allocation problem. In practical, there are many factors affecting anticollision scheme design such as the cost, the communication data rate, engaged bandwidth, how much information the transponder can provide and the range. In this section, we discuss the anticollision issues in the MAC sublayer design for a RFID system.

4.1. Transponder Collision And Selection

Reading more than one RFID tag at the same time is problematic because there will be MAI among transponders if all the transponders resonate the carrier signal back to the reader at the same time. There is the transponder selection problem [9, 10].

Beacon Approach

One of the popular schemes to solve this problem is called beacon scheme, which is similar to the random access schemes, ALOHA. In this scheme, each transponder repeatedly broadcasts its identification message to the interrogator at regular intervals until the interrogator acknowledges receipt without garbling from another transponder's message. The interrogator then concentrates on a one-to-one data exchange with the transponder concerned. Depending on the method employed transponders are either muted after being read or left to continue broadcasting. There are several variations of beacon schemes depending on how much control the interrogator has over the transponder.

Tree-Walking Approach

Another popular transponder selection scheme is based on the fast binary tree-walking search of uniquely numbered transponders. Here, either the transponders are each assigned a unique permanent ID number or they are given a temporary queue ticket number. The communication between transponder and interrogator is then started as the interrogator is search each existed ID. At each node in the tree, the interrogator checks for response. Only the transponders whose identi¯er is a child of the checked node respond. Otherwise, the sub-tree is empty. This scheme is also similar to the time division multiple access (TDMA) scheme. However, the original TDMA is hard to be implemented here since the synchronization among transponders is very difficult.

Though there are some discusses based the principles frequency division multiple access (FDMA) and code division multiple access (CDMA), they may operate best with large bandwidth available at UHF and microwave frequencies. There are some discusses about space division multiple access (SDMA) techniques, too. However, it is known to be too slow for most application. In practice, most RFID transponder anti-collision scheme are a mixture of beacon and tree-walking schemes and most of them require a unique ID for each transponder.

4.2. Interrogator Collision

An interrogator may interfere with the operation of other interrogators in the RFID system. There are two principal types of inter-interrogator interference. When two or more interrogators are closed to each other and use the same frequency at the same time, there is possible frequency collision among these interrogators. When two or more interrogators are attempting to communicate with a particular tag at the same time, there is tag collision [11]. Interrogator anticollision problem is similar to the inter-cell interference problem in mobile communications. Due to the limited functionality in transponders, normally it requires additional functions in interrogators or the communications among interrogators.

Besides the application of the ideas of TDMA, FDMA or CDMA schemes in mobile communication systems and random access schemes for interrogator collision avoidance, there are many variants of the MAC schemes for wireless ad-hoc network have been studied [5]. Basically, there are two classes of interrogator anticollision schemes, either synchronized or random, based on how the interrogators access the common channel. For synchronized schemes, there is a central coordination unit to synchronize the operation of interrogators for collision avoidance, such as TDMA, FDMA or CDMA-type schemes. For rand access schemes, each interrogator will compete for occupying the resource, such as ALOHA variations.

5. Security And Privacy Issues

The security issue in RFID refers to the allowance of access to the data stored on a transponder(Secrecy), the assurance that the stored data are kept from unauthorized changes (Integrity) and the "authenticity" of data being retrieved from a tag as having originated by the claimed source (Validity). Except the data security issues of RFID, privacy activists worry that the unchecked use of RFID could end up trampling consumer privacy by allow ing retailers to gather unprecedented amounts of information about activity in their stores and link it to customer information databases. They also worry about the possibility that companies, governments and would-be thieves might be able to monitor people's personal belongings, embedded with tiny RFID microchips, after they are purchased. They need be sure that the customer has the control over the data stored in the RFID tags. The optional security and privacy mechanism should pose minimum burden (and cost) on tag electronic, and most of the complexity is added at application level or at the interrogator level, depending on the access control to the interrogators. It should be able to manage frequent changes in the "keys". The key management should be handled at the application level, and not at the tag level.

5.1. Deactivation Approaches

The straightforward approach for protecting consumers' privacy is to disable, kill or destroy RFID tags with the request of consumers. After this, the tags will not longer be read by any interrogator. One of the simplest ways to achieve this, the cashier places the product with the tag into a deactivator that can generators a su±ciently high magnetic field to destroy some capacity in the ta g [12]. However, since this approach is too simple, it can be easily performed by unauthorized person, too. In AutoID center, another approach is proposed to deactivate a tag by sending a special KILL command, including a short 8-bit password, to it [13].

It is known that, the deactivation approach is not su±cient and undesirable for privacy enforcement. In many cases, consumers may actually wish RFID tags to remain some functions while in their possession. Consumers may want easy product information access, missing-protection and easy return in the future with active tags.

5.2. Encryption Approaches

Protecting privacy using cryptography methods is very challenging, providing the very limited computation resource and severe cost constraints on most RFID tags. Hash-lock [14] and re-encryption [15, 16] are two of the most discussed approaches.

In Hash-lock approach, the ID stored in transponder is locked using Hash function computation. However, the key management will become a big problem for customers. Using re-encryption approach, though it solve the key management problem may be solved, how to get the necessary computation resources for encryption becomes another problem.

Also, in order to secure the communication between the interrogator and the desired transponder, there is an approach called Silent Tree-Walking, which can encrypt the transmission of interrogator to prevent eavesdroppers from interfering the communications [14].

5.3. Blocking Approaches

With a blocking approach, the whole or partial ID information will be blocked from be accessed by interrogators. A interesting approach to achieve this is called Faraday cage, in which special containers are used for preventing interrogator from reading tags. However, it is known that not everything can be held in the containers. Another interesting approach is to use some machine to generate interference or jamming signals to prevent interrogators from working well. However, it is known that there are many restrictions on using this kind of machine. Recently, an approach called blocker tag is proposed to block certain information bits on tag IDs when tags rely on tree-walking as a anti-collision technique [17].

6. Regulation And Standards

By now, the operation of RFID systems is regulated by local governmental bodies, which control the electronmagnetic spectrum in a region. A degree of uniformity is being sought for carrier frequency usage, through three regulatory areas, Europe and Africa (Region 1), North and South America (Region 2) and Far East and Australasia (Region 3). It is expected to achieve some uniformity by the year 2010. Currently, most RFID systems operate in so-called Industry-Scientific-Medical (ISM) bands, which are designated by the International Telecommunications Union (ITU). The most commonly used ISM frequencies for RFID are 13.56MHz worldwide and 902-928MHz in US.

By now, all major RFID vendors o®er proprietary systems, with the result that various applications and industries have standardized on di®erent vendors competing frequencies and protocols. The lack of open systems interchangeability has severely crippled RFID industry growth as a whole, and the resultant technology price reductions that come with broad-based inter-industry use. However, a number of organizations have been working to address and hopefully bring about some commonality among competing RFID systems, both in the U.S. and in Europe where RFID has made greater market inroads. Meanwhile in the U.S.A., ANSIs X3T6 group, comprising major RFID manufacturers and users, is currently developing a draft document based systems operation at a carrier frequency of 2.45 GHz, which it is seeking to have adopted by ISO. (www.rfid.org)

Since DSP based architecture provides lots of °exibility and the performance-to-price ratio of DSP hardware price keeps decreasing, it is widely used in many communication system design as well as interrogator design. In Fig. 7, the interroagor is designed around a powerful DSP, which handles all the modulation, demodulation, anticollision search functionality, security and privacy control and communication protocols in software. The other hardware of the interrogator consists of a RF module and oscillator connected to the DSP system. These modules are simple up/down converters that convert signals from the operating frequency to baseband for the following DSP system.

In uplink, the received signal is converted into base-band signal before entering DSP. It is then digitized via A/D to become digital signal. After this, the digital signal will be demodulated and decoded into bit frames. These frames is sent into data-link layer for anticollision control. After anticollision control, the decoded information goes through security and privacy control so that the 'hidden' information from transponder is decrypted.

After the interrogator obtains the required information from transponder, it may be shown on the interrogator display via output display driver or put into IP package to be relayed back to the application server via network interface card.

The downlink communication is basically the same to the uplink communication. The instruction either from application server or from interrogator itself is encrypted, coded and modulated before transmitted to the transponder within its time slots. The time slots for each interrogator are allocated by the anticollision control block.

8. Conclusions

RFID system is adopted in more and more application areas because of its contactlessness and good penetration in low frequency. In this report, several issues regarding RFID system design are discussed. A reference firmware design is proposed, too. It is expected to help with the development of a secure and robust RFID system in the future.

References

[1] K. V. S. Rao. An overview of backscattered radio frequency identification system (RFID).

Microwave Conference, 1999 Asia Pacific, 3:746~749, December 1999.

[2] A. Riabtsev;I. Zakopailo; U. Piletsky; V. Irinarhov; V. Goncharov;V. Istratov and

A. Barcovsky. The versatile RFID system. Science and Technology, 2000. KORUS '99.

Proceedings. The Third Russian-Korean International Symposium on, 2:709~711, June 199.

[3] N. Raza;V. Bradshaw and M. Hague. Applications of rfid technology. RFID Technology

(Ref. No. 1999/123), IEE Colloquium on, 1:1~5, October 1999.

[4] T. Flor ; W. Niess and G. Vogler. RFID: the integration of contactless identification

technology and mobile computing. In Telecommunications, 2003. ConTEL 2003. Pro-

ceedings of the 7th International Conference on, pages 619~623, June 2003.

[5] J. Waldrop; D. W. Engels and S. E. Sarma. Colorwave: a mac for RFID reader networks.

In Wireless Communications and Networking, 2003. WCNC 2003, volume 3, pages 1701~

1704, March 2003.

[6] P. R. Foster and R. A. Burberry. Antenna problems in RFID systems. RFID Technology

(Ref. No. 1999/123), IEE Colloquium on, 3:1~5, October 1999.

[7] S. C. Q. Chen and V. Thomas. Optimization of inductive RFID technology. Electronics

and the Environment, 2001. Proceedings of the 2001 IEEE International Symposium on,

pages 82~87, May 2001.

[8] J. Engel. DSP for RFID. Circuits and Systems, 2002. MWSCAS-2002. The 2002 45th

Midwest Symposium on, 2:227~230, August 2002.

[9] P. Hawkes. Anti-collision and transponder selection methods for grouped vicinity cards

and RFID tags. RFID Technology (Ref. No. 1999/123), IEE Colloquium on, 7:1~12,

October 1999.

[10] P. Hernandez;J. D. Sandoval;F. Puente and F. Perez. Mathematical model for a mul-

tiread anticollision protocol. Communications, Computers and signal Processing, 2001.

PACRIM. 2001 IEEE Pacific Rim Conference on, 2:647~650, August 2001.

[11] D. W. Engels and S. E. Sarma. The reader collision problem. Systems, Man and

Cybernetics, 2002 IEEE International Conference on, 3:6, October 2002.

[12] Klaus Finkenzeller. RFID Handbook: Fundamentals and Applications in Contactless

Smart Cards and Identi¯cation. John Wiley & Sons, Ltd, 2003.

[13] S. E. Sarama; S. A. Weis and D. W. Engels. Radio-frequency identification systems.

CHES 2002, pages 454{469, 2002.

[14] S.A. Weis; S. Sarma; R. Rivest and D. Engels. Security and privacy aspects of low-cost

radio frequency identification systems. In First International Conference on Security in

Pervasive Computing, 2003, 2003.

[15] P. Golle; M. Jakobsson; A. Juels and P. Syverson. Universal re-encryption for mixnets.

to appear., 2002.

[16] A. Juels and R. Pappu. Squealing euros: Privacy protection in RFID-enabled banknotes.

In Financial Cryptography, 2003.

[17] Ari Juels; Ronald L. Rivest and Michael Szydlo. The block tag: Selective blocking of

RFID tags for consumer privacy. to appear., 2003.