19-09-2012, 03:13 PM
Rfid in pervasive computing: State-of-the-art and outlook
Rfid in pervasive computing.pdf (Size: 1.24 MB / Downloads: 60)
ABSTRACT
rfid has already found its way into a variety of large scale applications and arguably it is
already one of the most successful technologies in the history of computing. Beyond doubt,
rfid is an effective automatic identification technology for a variety of objects including
natural, manufactured and handmade artifacts; humans and other species; locations; and
increasingly media content and mobile services. In this survey we consider developments
towards establishing rfid as the cost-effective technical solution for the development of
open, shared, universal pervasive computing infrastructures and look ahead to its future.
In particular, we discuss the ingredients of current large scale applications; the role of
network services to provide complete systems; privacy and security implications; and how
rfid is helping prototype emerging pervasive computing applications. We conclude by
identifying common trends in the new applications
Introduction
At first glance rfid and its application to pervasive computing appear to be simple and straightforward to implement.
Yet, the opposite is true: rfid is a technology that requires the development of distinct systems, software and networks
of considerable complexity. It entails antenna design, radio propagation analysis, low-cost integrated circuit production
techniques (increasingly focused on printed electronics), receiver design and data encoding mechanisms, lightweight
encryption and security protocols, materials technology, network discovery and information services, and novel interaction
design approachesto mention only some of the numerous engineering and computing disciplines involved in the
development of complete rfid systems.
It is not surprising then that following Stockman's statement in 1948 [1], rfid required almost 60 years of development
to find its way to large scale applications [2]. And also, the fact that so many disciplines are involved implies that a full
treatment of every aspect of rfid cannot be contained in a single survey. Instead, this paper addresses those aspects of rfid
that are useful to the developer and the researcher of pervasive computing systems. We hope that the survey will be useful
both as an introduction to the technology and as a guide to emerging research trends and novel applications.
Rationale and overview
Recent years have observed an explosion of interest in rfid. The reason for this is twofold: first, the availability of
very low-cost passive rfid tags that require no battery to operate; and second, the wider availability of robust internet
infrastructures that can provide networked services to complement rfid and thus provide complete system functionality.
These developments have allowed large-scale commercial applications in the supply chain [3,4], ticketing [5], asset
tracking [6], maintenance [7], retail [8], and personal identification [9]. Due to these applications, rfid has become one of
the most numerous computing platforms in use today: IDTechEx, a market research firm specializing in rfid, estimates that
more than 3.7 billion rfid tags have been deployed in the field by mid-2007, with more than 1.6 billion new tags employed
in 2006and this trend is accelerating. The increasing popularity of rfid permits further cost reductions, and has revived
interest in the use of rfid in pervasive computing research due to the unique opportunities it offers for low-cost large-scale
experiments of novel systems and applications.
Yet, the same features that make rfid such a popular technology are also complicating its use. To offer battery-free
operation and low cost, passive rfid tags have extremely limited capabilities often being able to hold and in fewer cases
protect only a simple entity identifier and potentially a very limited amount of contextual information. This code is often
employed as the handle that links a specific physical entity to its stored information. As a consequence, a considerable
proportion of system functionality must be located on the network. For example network directory services could be used
to relate the code retrieved from a tag to entity-specific metadata including the description and associated attributes of the
tagged entity.
Ticketing
One of the earliest and most successful large-scale applications of rfid has been in metropolitan public-transport
ticketing. Today, such systems are deployed in numerous cities across the globe, in some cases having been in operation
for a decade. In ticketing applications, rfid tags are often embedded in credit-card sized reusable tickets, which store either
a seasonal pass or credit that can be used against travel. Such tickets hold a code that identifies uniquely the ticket or the
passenger, and in some cases also maintains personal information about the holder and a record of the most recent trips.
Tickets are validated and updated in real-time at gates fitted with readers during entry and exit to the transport system.
One of the largest rfid-based ticketing systems is the Oyster card [5] in London, which supports more than 10 million
active passengers and has deployed over 27 thousand readers. Other popular systems include the Octopus card in Hong-
Kong and the Suica card in Tokyo, both of which are notable for the fact that their use has been extended beyond ticketing
to general-purpose micropayments accepted at a large number of retail outlets.
Supply chain management
Supply-chain management (SCM) has been one of the main drivers of rfid technology in recent years. SCM deals with
the movement of goods between organizations across a supply chainfrom raw materials obtained by the manufacturer
to finished products delivered to the consumer. Each supply chain is distinct and reflects the unique needs of the range of
products that are processed, ranging from the delivery of fresh food from the farm to the supermarket shelf to army uniforms
transferred from the manufacturer to the soldier in the desert. Nevertheless, all supply chains share a common goal: to keep
the process simple, standard, speedy, and certain. To achieve this objective, it is necessary that all participating organizations
exchange accurate information at frequent intervals and that related costs be unequivocally identifiable at all times. This in
turn requires the use of open, worldwide data standards for globally unique product identifiers and product classification
schemes, combined with internet-worked information services that can be used to track and trace goods and services [12].