The utility value of radio frequency identification (RFID) technology in supply chain management is already well known. The successful commercial deployment of RFID in this arena, which showcases the ease of tagging and wireless identification, has led to the prospect of new and innovative applications, such as wirelessly monitoring the movement of people and personal objects.
However, this potential also raises new challenges such as issues of privacy, security and uncertainty. While the scope of pervasive RFID-based sensing and identification systems continues to attract researchers and businesses alike, such underlying issues and corresponding solutions are difficult to address without a pilot deployment.
In light of this, Frost & Sullivan has taken interest in a project undertaken by researchers from the Computer Science and Engineering department at the Seattle-based University of Washington. They are involved in a large-scale project, having deployed a building-wide networked system of hundreds of RFID readers and thousands of RFID tags. More information can be found here.
The so-called RFID Ecosystem project serves as a living laboratory for sustained research to uncover underlying issues such as security and privacy and devise solutions that enable unique applications. Volunteers wearing RFID tags and their tagged personal objects are tracked inside the Paul G. Allen building on the university campus and the tag read events (TRE) are captured and stored in a central database.
Handling the huge amount of data generated and making sense out of it has been a core aim of this project, as well as coming up with innovative techniques to ensure privacy of the data between the users.
The architecture of this user-centred system consists of RFID readers at the bottom, which forward the TREs consisting of details about the location, identity and time of events to the gateway nodes. The root gateway sends a stream of RFID data to data management systems such as a particle filter for location estimation and a probabilistic event detector for extracting meaningful high-level event data from low-level RFID data.
There are also specific modules which enforce privacy policies using access control systems. An event manager coordinates the data management system and enables the applications and tools to access the data using an event-driven, enquiry-based application programming interface (API). A suite of user level tools simplifies definition of new events and metadata as well as facilitating system administration tasks.
Talking to Frost & Sullivan, Magdalena Balazinska, assistant professor at the University of Washington and project leader, concedes that a major challenge faced in the project was managing the large volume of 'dirty' RFID data and data streams with gaps. "To address these challenges we are building new data management systems that can handle large volumes of dirty RFID data," Balazinska said. "Our systems make it easier to build RFID applications."
Another challenge revolved around privacy issues, for which the researchers developed unique techniques.
The RFID Ecosystem project has also endeavoured to showcase potential applications of such a deployment through their subprojects - RFIDDER, RFID Map Mash-up, Scenic Tool and Digital Diary Tool.
The RFIDDER is a micro-blogging and friend finder tool, which uses the RFID Ecosystem's event detection services to keep in touch with a group of friends using automatic status updates. The RFID Map Mash-up uses a map interface to see locations of friends in realtime if they have agreed that such data be visible during specific times between each other. Both tools support mobile as well as Web interfaces, and are being integrated with a popular social networking and micro-blogging service, Twitter.
The Digital Diary Tool automatically logs detected events to help users track how, where and with whom they spent time. The researchers have also demonstrated this service using Google's Web-based calendar service.
Although the project successfully demonstrates the realm of applications possible with large scale deployments, the difficulty in setting up a network of RFID readers on a large scale does limit the utility value to a certain extent. RFID tags are relatively cheap and can quickly become ubiquitous if there is enough potential, but deployments are only easy for intra-building arrangements and very difficult for a metropolitan area-wide set up.
On one hand, features such as intimating a user of a misplaced object are likely to be very useful in a university/office situation. But on the other hand, features such as automated micro-blogging and diary logging can be made available only to a limited number of subscribers through most social networking sites in which aficionados are likely to be interested. Nevertheless, potential does exist for further development and more feasible commercial prospects are likely to emerge in the near future.
For more information contact Patrick Cairns, Frost & Sullivan, +27 (0)21 680 3274, email@example.com