RFID Project Technical Report

RFID Project Technical Report Bernard Ghanem 1. Introduction The use of Radio Frequency Identification (RFID) technology extends historically to the 1970 s, when the first commercial products emerged. The introduction of this technology motivated future research that sought to integrate RFID into various applications ranging from inventory control/monitoring to human/animal tracking. However, the practical significance of this technology was not fully realized, until mass manufacturing of RFID tags and readers, at different scales and for different ranges, was made possible. Currently, RFID-based systems are being proposed as alternatives to previously established technologies (e.g. Wi-Fi for indoor identification and tracking) and as extensions to others (e.g. GPS localization and tracking in indoor settings). A basic RFID system is comprised of three main elements, as follows: RFID tags/transponders: These are simple communication devices that are activated by incoming registration from the RFID reader, when it is in its operating range (e.g. 3m-10m). They usually contain a small amount of readable/writable memory, which is used to save a unique ID and other important information (e.g. temperature logs). RFID reader: On one side, the reader communicates with and accesses the RFID tags via an antenna, whose size is dictated by the desired range of the reader and its power constraint. On the other, it connects to a processing unit, which initiates read/write operations and is responsible for post processing of the data acquired. Processing unit: This can be either a stationary (e.g. PC) or mobile platform (e.g. PDA). It integrates the reader-tag subsystem into a higher level application (e.g. tag deployment, temperature monitoring, etc.). The concept of RFID systems has been available for a long time, yet only recently has its potential been manifested in applications of ever-growing demand. The recent popularity of RFID technology is due to the following fundamental properties that characterize the production, deployment, and functionality of RFID systems. (1) The recent ability to acquire low cost and power efficient RFID equipment has supported a novel trend in RFID system research that seeks to replace current technologies. (2) Its non-contact, non line-of-sight functionality allows for remote access of data through a variety of substances such as concrete, paint, and wood, which pose impeding obstacles to its optical counterparts (e.g. barcodes). (3) Due to its data storage capacity and the addition of diverse sensors to RFID tags, this technology has the potential of bridging the growing gap between the digital networked world and the physical world [3]. This provides a low cost means for ubiquitous sensing of environmental parameters, which is not restricted to measurement alone, but it can also render intelligent monitoring of these parameters and their temporal fluctuations. For example, RFID tags equipped with humidity and temperature sensors have been utilized during the transport and delivery of perishable products including fruits, vegetables, and meat (e.g. Sensitech Inc.). In this report, we describe details of an ongoing project that aims at integrating RFID technology into building infrastructure non-invasively. We present the significance of utilizing RFID within buildings as a cornerstone for 1

ubiquitous sensing and describe three applications we have developed: RFID tag management, continuous temperature monitoring, and real-time indoor location sensing. These applications are user-friendly implementations that facilitate the manipulation of the data stored in the RFID tags especially in regards to location information and temperature logs. 2. Current System Setup We are using an RFID system based on Intelligent Long Range (ILR) technology provided by IDENTEC SOLUTIONS. Figure 1 illustrates the overall setup of our RFID system. We make use of two types of semi-passive RFID tags (i.e. i-d2 and i-q32t), which share common properties, but are characterized by complimentary functionality. In what follows, we provide a brief description of each tag. i-d2 tags are used for location sensing; they can be detected by a mobile reader unit up to 6m away. These tags can be mounted on the walls of corridors and/or rooms in order to span the required space. They contain a small amount of volatile memory (i.e. 64 bytes), which is used to save its unique ID, location information, and other local properties. They are equipped with an LED that facilitates identification of the tag. i-q32t tags log temperature measurements over time at a user defined rate and transmit them to a reader up to 100m away. They contain a relatively larger memory capacity (i.e. approximately 6000 bytes). They also have embedded LEDs for visual identification. Figure 1. (a) shows the two types of tags currently being used. The gray tag on the right is an i-d2 tag, while the black one is the i-q32t tag. (b) shows an i-cardiii RFID reader, which can be embedded into a PC or PDA processing unit via a PCMCIA type II slot. (c) shows two types of antennas that are used depending on the application. The longer antenna on the left increases the range of the reader to 100m and is primarily used with a PC configuration. On the other hand, the shorter antenna is meant for mobility and is connected to a reader embedded in a PDA. The range of this antenna is an order of magnitude smaller than the longer one. Since indoor settings are being considered, we require the tags to be spatially fixed, while the reader can be either spatially fixed or allowed to roam. Consequently, the tags are non-invasively attached to building infrastructure at locations that are convenient for its underlying functionality. For example, since i-d2 tags will be used for location sensing and tracking, they are positioned uniformly along walls, such that the tag-to-tag distance allows for adequate RF coverage. On the other hand, since the i-q32t tags will be used for temperature logging, they are positioned in strategic positions within rooms to account for local temperature distribution. In Section 3, we describe how we exploited this simple setup to implement an RFID system, which is capable of three main applications: tag deployment, temperature monitoring, and location sensing. 3. Description of System Components In this section, we illustrate how an RFID distributed system, as described previously, can be applied to three real world problems prevalent in indoor settings: (1) tag management/deployment, (2) continuous temperature monitoring, and (3) distributed location sensing. Figure 2 illustrates an overview of all three applications. (1) is

essential since the other applications assume that the tags have been deployed successfully within the building. (2) & (3) build on the location-based information available through tag deployment to respectively manage higherlevel local information (i.e. temperature measurements) and deduce more global characteristics (i.e. sensing the position of a handheld RFID reader from the deployment information of tags in its vicinity). Figure 2. Application (1) is responsible for storing the user-defined deployment information on the corresponding tags. Application (2) allows for continuous monitoring of room temperature based on the temperature measurements relayed by the i-q32t tags. Application (3) performs real-time location sensing of an RFID reader, given the location and power information of the tags in its vicinity. 3.1. Application (1): Tag Deployment/Management Inherently, each RFID tag is given a unique ID, which enables the reader to distinguish it amongst other tags. Although this uniqueness is vital for any RFID application, it does not give much insight on the local properties of the tag, most importantly its locale. Consequently, we propose to make each tag aware of its local position, which is determined by its image coordinates within a map among a set of predefined maps. In fact, each tag stores these local coordinates in addition to the map name, which uniquely identifies that map (e.g. floor number within a multi-floor building). For this purpose, we developed a user interactive application, which allows building administrators to manage the deployment of RFID tags as they are installed within the building. Figure 3 gives an overview of this application s functionality. Indeed, it is responsible for transferring the desired deployment information (i.e. position information and tag options) to the tags themselves. Figure 4 depicts the interface of this application. The user simply chooses the tag to be positioned and identifies its local position by clicking on the corresponding position on the given map. In this case, there are four maps to choose from. Also, navigation between tags and maps is facilitated, as the selected deployed tag is bounded by a green box. All deployed tags are represented by red blobs at their corresponding positions within the map.

Figure 3. Flow chart for the functionality of Application (1) When an i-q32t tag is being installed, its temperature logging interval (i.e. the interval of time between any two time stamps corresponding to consecutive temperature measurements) should be set by the user. The default log interval is one minute. Here, we note that a secondary application was developed to allow the user to define the rooms designated for RFID usage (color coded on the map in yellow) and set their individual properties (e.g. temperature thresholds that will be used in Section 3.2). Figure 4. The user interface to the deployment application. Tags are identified by their type (i.e. id2 or i-q32t) and their unique ID. A deployed tag is marked by an asterisk and the selected tag is identified by a green box. The Deploy button is responsible for storing the user-defined location and map ID onto the selected tag. The Undeploy button reverses the results of the Deploy button. The Scan button permits additional scanning when new tags are introduced to the vicinity. This interface allows for three main functions, which hide the underlying operations required to successfully

complete them. (1) By clicking the Deploy button, the label of the chosen map and the local coordinates in this map (as denoted in the X and Y coordinate labels) are stored on the respective tag by writing them as a single string to that tag. A deployed tag is identified by an asterisk. For an i-q32t tag, the specified logging interval is also written to it. (2) When the Undeploy button is clicked, the deployment information of the selected tag is cleared from its memory. This function is essential for the purpose of redeployment or decommissioning of tags that are to be relocated or disposed of. (3) The Scan button allows the user to rescan the vicinity for all possible tags. This function is required in order to deploy new tags that were not visible previously. Moreover, the deployed tag information can be stored offline in a tag description database, which must be kept up-to-date to the current tag status. In summary, this tag management tool provides user-friendly visualization of where a tag is (or will be) deployed within a map. More importantly, it forms a basis for the higher-level applications to be discussed in Sections 3.2 and 3.3. 3.2. Application (2): Temperature Monitoring In this section, we show how an i-q32t tag s ability to measure temperature in its proximity is combined with its location information to render a continuous temperature monitoring system. Such a system finds important use in case of fire, when a locally powered distributed system can be exploited to give a rapid assessment of the extent, history, and location of the hazard. In Figure 5, we illustrate the flow of information and decisions within the framework of this application. Basically, the temperature daemon scans the vicinity for deployed i-q32t tags, extracts their temperature logs, and decides whether a thermal irregularity has occurred. Such an irregularity is defined by the inherent properties of the room within which these tags are located. If such an irregularity is observed, a warning log is formed for future reference and an immediate emergency response is initiated (i.e. sound a fire alarm, contact authorities, etc.). In our implementation, we designate a single i-q32t tag to a given room and restrict the emergency response to a simple beeping. Figure 5. Flow chart for the functionality of Application (2) Once temperature logging tags are deployed, the user can easily monitor and visualize the temporal variations in individual rooms, by using the interface depicted in Figure 6. Two modes of operation are made available. (1) In manual mode, the user can scroll the time line through the temperature logs of each room over the last 24 hour period. These variations are depicted via a color map, which defines thresholds for the three regions of

operation: cold, moderate, and warm. Also, the user can view temporal variations of temperature logs for a single tag by right-clicking the tag in the list and viewing its temperature plot, as shown in Figure 7. (2) Continuous temperature monitoring is the second mode of operation for this application. In this mode, only the most recent temperature of the rooms under study is portrayed, while the tags are scanned at a periodic interval specified by the user. This mode supports the detection of any temperature irregularity, which might occur in case the threshold assigned to a given room in the map is exceeded. The application checks for such irregularities at every new scan of the tags in the vicinity. When the permissible threshold is exceeded, a warning is issued and it is saved to a warning log file. This warning identifies the location, time, and extent (most recent temperature recorded) of the irregularity. Figure 6. This figure shows the user interface for Application (2). 3.3. Application (3): Indoor Location Sensing The location information stored on the i-d2 RFID tags can be exploited to implement a location tracking system, which estimates the approximate position of the RFID reader relative to the positions of the tags detected in its immediate vicinity. Previous work on RFID-based, indoor, location sensing/tracking [1, 2] aimed at tracking an RFID tag relative to stationary reference RFID readers. Our goal is just the reverse, where the tags are fixed to the building s infrastructure and the reader is the mobile agent, whose relative position needs to be estimated. Deployment information accessible from the tags permits estimation of the global position of the reader with

Figure 7. A temporal plot of the temperature measured by the selected i-q32t tag. respect to the map that they belong to. This application is of significant importance during emergency situations, where an agent requires real-time localization and possible path planning towards a certain destination (e.g. an exit during a fire). Figure 8 illustrates the main components of this application. In essence, this application is loaded into a mobile agent (a PDA in our case), which is equipped with an RFID reader and a short range antenna. It scans for deployed i-d2 tags in its proximity. From their deployment information, it deduces which map it should belong to, populates this map with the positions of the tags, and proceeds to estimate its own position. We propose to use a form of triangulation to render an adequate approximation of the reader s global position. Our approach utilizes a direct neighborhood algorithm that takes into account the ranges of the detected tags, and their received power profiles. In order to provide a more accurate estimation, we conduct offline experiments to determine the power vs. range relationship of a typical i-d2 tag. Figure 9 shows the transmission power profile of an RFID reader through a range of distances. So, from the received power signature of each scanned tag (i.e. the transmission power setting of the reader once the tag is detected), an estimate of its distance to the reader is attained according to this defined mapping. By assuming omni-directional transmission, we can draw a simple geometric equivalence from the problem of estimating the reader s global position. Each tag can be represented as a disc, whose center is located at the respective tag position in the map and radius is the estimated distance from the reader. This distance is proportional to the uncertainty that the reader lies within this range. Hence, location sensing becomes a problem of determining the intersection of all these discs. Obviously, if more tags are deployed in the vicinity of the reader, the intersection area will be smaller and the position estimate will be more accurate. Figure 10 shows the result of a location sensing task, with three deployed tags in the reader s range. There are three partially visible circles of different radii, whose intersection identifies the region enclosing the reader. We define the largest enclosed disc (color coded in green) within this intersection region to be the area within which the reader is located (approximated as

Figure 8. Flow chart for the functionality of Application (3) Figure 9. Experimental results for the relationship between the maximum power needed by the RFID reader to successfully access an i-d2 tag located at a certain distance from the reader. We resort to linear interpolation to acquire a continuous mapping between distance and transmission power. its center).

Figure 10. Interface of Application (3) on a PDA equipped with an RFID reader. 4. Conclusion and Future Work In this report, we have described a user-centric RFID-based system for indoor environment monitoring and location sensing. Currently, three user-friendly applications have been implemented that hide the underlying complexity of operating the RFID equipment, while providing the desired functionality. For future work, we aim at conducting more extensive experiments to validate the accuracy of our location sensing implementation. In fact, these experiments will entitle the following scenario: a number of tags are placed uniformly along a hallway, which the hand-held reader traverses. In this controlled environment, we will compare the groundtruth localization to that estimated by our application. Further, experiments should be conducted to estimate the number of uniformly spaced tags needed to approximate a reader s position with a given precision. Moreover, we aim at embedding the location sensing application into a real-time tracking system and extending it to include a path planner. This provides the user with not only his/her locale, but also navigational information for a given pair of destination and source locations. These extensions will incorporate more detailed information about the surroundings and map structure. References [1] D. Cox, V. Kindratenko, and D. Pointer. Intellibadgetm: Towards providing location-aware value-added services at academic conferences. Proc. of the International Conference on Ubiquitous Computing, pages 264 280, 2003. [2] L. M. Ni, Y. Liu, Y. C. Lau, and A. Patil. LANDMARC: Indoor location sensing using active RFID. Proc. of the International Conference on Pervasive Computing and Communications, pages 407 415, 2003. [3] R. Want. Enabling ubiquitous sensing with RFID. IEEE Computer Society: Invisible Computing, 37(4):84 86, 2004.