A User Interaction Model for NFC Enabled Applications Yaw Anokwa 1 , Gaetano Borriello 1 , Trevor Pering 2 , Roy Want 2 1 Computer Science and Engineering University of Washington Seattle, WA 98105 {yanokwa,gaetano}@cs.washington.edu 2 Intel Research, Santa Clara 2200 Mission College Blvd. Santa Clara, CA 95054 {trevor.pering,roy.want}@intel.com Abstract Near field communication (NFC) is a short-range wire- less protocol that allows users to connect devices and ac- cess content and services by simply holding enabled devices near each other. This paper introduces a user interaction model for NFC enabled applications. Our model specifies that enabled devices take on the properties and context of the objects required in the interaction. This transformation leverages the existing knowledge users have about certain objects and thus can support a number of different appli- cations tied together with simple, intuitive and repeatable interactions. In this paper, we present an overview of the model and the system we have implemented to enable eval- uation. We also detail some research challenges we are pur- suing. 1. Introduction Mobile devices have become the primary platform of ubiquitous computing. There are now over 2 billion cell phone users worldwide [11] and many of these users carry these devices with or near them [4]. As these devices grow in popularity, technologies that enable more natural interac- tions between users, devices, and their environments have spurred a rich and vibrant research community. It is likely that rather than carrying augmented everyday objects as suggested by Want et. al [10], users will instead carry a single device that has the same functionality as today’s ev- eryday objects. For example, instead of a tagged car key, a cell phone could be authorized to open and start a car. Near field communication (NFC) is a technology evolved from short range radio frequency identification (RFID). Like RFID, NFC works via magnetic field induc- tion and is designed for simple and safe transfer of data be- tween compatible devices. Effective range is limited to 20 centimeters and data transfer rates peak at 424 kbits/s mak- ing it a good technology for scan/touch interactions. NFC goes beyond RFID in that it is a symmetric protocol. Read- ers can read from tags and other readers directly. Bi-directional device to device transfer of information is the key feature that differentiates NFC from RFID. For example, users can use their mobile devices to scan a tag embedded in a “smart” movie poster and either get infor- mation about the movie, view a trailer or purchase tickets to the movie. This scenario can be accomplished with RFID but with NFC, users can then go home, scan an enabled TV with their mobile device, transfer the trailer to the TV and watch it on the larger screen. The mobile device’s ability to act as both a tag and a reader makes this scenario possi- ble. Users can also transfer tickets that were purchased at the poster by scanning their respective mobile devices. Unfortunately, many of the proposed NFC enabled ap- plications are not framed under a general interaction model. This is a significant problem for NFC because it can both be used for simple interactions like touching a secure door with a cell phone to gain access and for more complex sce- narios such as buying a movie ticket. Without a user model in place, NFC enabled applications may end up a mix of poorly thought out interfaces without a unifying interaction model. We solve this problem by relying on a user’s pre-existing cognitive model of the objects with which they are interact- ing. When a mobile device scans an item, the device takes on the properties and context of the item scanned. We call this process a transformation. For example, if a cellphone scanned a smart movie poster, it could take on the proper- ties of a movie ticket. It could contain all the information printed on a standard ticket, it would be used to gain ac- cess to a movie, it could be given to a friend, and finally when authorized at a theater would become a ticket stub. So after a transformation, each interaction with the mobile device is intuitive because the mobile device behaves like the scanned item. Of course, a single device may transform 1