Pages: pp. 295-303
Abstract—Due to the increasing demand for RFID expertise and the existence of a knowledge gap between industry and academia in this domain, work has been stimulated to help spread understanding in this field and bridge the gap between theoretical examinations and industrial practices. Among the encouraged work, there is the Integrated Auto-ID Technology for Multidisciplinary Undergraduate Studies (I-ATMUS) project that involved developing an online learning environment for RFID technology with a remotely controllable laboratory system. Technological resources can now be accessed by learners through the Web technology to apply appropriate configurations to the system, conduct experiments using RFID technology, and perform statistical analysis on the acquired data. The developed educational tool was used by two sets of students, who showed improvements in their confidence, knowledge, and skills.
Index Terms—Remote laboratory, learning environment, online teaching, RFID, hands-on education.
For some technologies, the supply of their qualified users struggles to match the pace of the associated growing demand. In other words, the growth of the skilled-user community may not be able to keep up with the rapid evolvement or emergence of these technologies. Companies may use outsourcing or internal training to obtain the necessary expertise, but that does not always solve the problem in a time- or cost-effective manner. While that might be a constraint, a relatively great number of potential expert users of a specific technology are supplied from schools and universities.
One solution would be to expose more engineering students to the newest technologies, such as Radio Frequency Identification (RFID), while they are still in school. This can lead to an increase in the supply of acquainted users, who can become experts at later times. The problem with such a strategy is that a technology might not be available or accessible to enough institutions to see the desired educational results. Whether or not that is due to affordability, novelty, or safety issues, the problem is likely to persist until some type of collaboration is established between organizations that teach different kinds of technologies.
A way to promote educational collaboration and instructional instruments sharing is the utilization of the ubiquitous Web technology. This approach can provide distant exploitation of the scarce technological assets of a certain institution, whether they are hardware devices or software applications. For this reason, it would be desirable to develop learning environments that yield remote access to technological resources as well as collaborative learning. In fact, several schools have already started promoting online educational tools from which students at different locations can greatly benefit. These distant learners are able to visually acquire knowledge and/or obtain hands-on experience through animation-based tutorials [ 1] and practical activities in virtual laboratories [ 2], [ 3], [ 4], [ 5] or remote real laboratories [ 6], [ 7], [ 8], [ 9], [ 10]. Not only that, but Web technology is also capable of providing new teaching techniques that are appealing to students [ 11].
On the other hand, RFID is one of the new technologies that is more visible than ever and has a high potential of being used extensively in the near future. The existence of RFID laboratories in educational environments will serve the purpose of providing testing results and conclusions, as well as giving the involved students the opportunity to obtain hands-on experience, making them potential RFID experts and valuable assets to RFID stakeholders. This is important for employers who want to adopt RFID since the majority of them believe that there are not enough RFID-skilled people to hire [ 12].
It is therefore very valuable to have an automated RFID laboratory whose equipment is remotely controlled, and whose graphical user interface (GUI) is linked to a knowledge base about RFID technology and related matters. A learning environment of this kind has the potential of satisfying corporate needs, supporting collaborative educational programs, and promoting RFID technology. The contribution of this paper is 1) the development of an online learning environment that targets teaching RFID with an emphasis on the practical aspect of the technology, 2) automated easy control of the laboratory hardware devices, and 3) the student assessment to the created value of the developed learning environment.
Radio Frequency Identification is a data-collection technology that utilizes wireless radio communication (radio frequency signals) to identify, track, and categorize objects (see Fig. 1). The basic RFID system consists of three main components:
This system can be extended with a set of middleware devices, a variety of soft controllers, a network of readers, and a powerful database management system (DBMS) to ease data acquisition and data management in a large information system.
Figure Fig. 1. Object/device interactions in an RFID system.
With its capability of storing a relatively large amount of data, an RFID tag can outperform a barcode tag, which can identify the kind of an item only, one item at a time, and has to be scanned with line of sight. When an RFID tag utilizes batteries to function, it is called Active, it can be read from far distances (up to 30 meters, or 100 feet), and it uses a specific range of radio frequencies. Passive tags, on the other hand, do not require batteries; they are powered by the electromagnetic waves sent by the reader, and that is why their read distance is limited to a number of feet. The main advantage that Passive technology has over its Active counterpart is the significant cost and maintenance reductions.
Science and engineering fields have been enriched by physical laboratory experiments; however, it is difficult for students to fully comprehend many modern systems due to the limited access to laboratory equipment [ 1], [ 5]. Recent research has revealed that students learn and retain information best through interactive examples and experiments [ 4], [ 13], [ 14]. With the evolution of technology and the Internet, many researchers in all fields are focused on creating Web-based laboratories to ease the learning of students by providing them with the ability of studying anywhere and anytime [ 3], [ 5], [ 15]. Online laboratories also have the advantage of assisting researchers stimulate the interest of learners with new teaching techniques provided by Web technology [ 11].
As the Web has many applications in the education sector, it exploitation includes conveying the knowledge through electronic information, animation, simulation, and virtual laboratories. For example, in [ 15], the researcher discusses how the Internet aids the development of the manufacturing engineering curriculum to augment the students' knowledge and sharpen their technical skills by giving them the opportunity to access subjects that would otherwise be unavailable without the use of the Web. In [ 11], Web-based modules are constructed to enhance the learning experience of students in thermal fluids and complement the traditional engineering lectures. In [ 1] and [ 2], a Web-based system for Programmable Logic Controller (PLC) education is created by using "intelligent tutoring" to convey the knowledge with the animation being the most popular instructional activity. A new mechatronics laboratory is developed in [ 14] to enhance the learning of students in design and development and digital signal processor-based controllers via a "model-based, simulation-oriented approach." In [ 5], virtual laboratories are developed in order to complement the limited classroom material with the laboratory adequate resources in earthquake engineering education. Another online virtual laboratory for instructing the behavior and design of reinforced concrete structures is presented in [ 4].
The employment of the Web technology is taken even a step further by utilizing the Internet to access laboratory equipment remotely. For instance, the researchers in [ 9] show how their remote real laboratories serve students from different majors to conduct experiments in several areas, such as control systems, mechatronics, and robotics and automation. The online system also assists the efforts of integration of laboratories at the University of Illinois to eliminate duplication of facilities and establish across-departmental laboratory instruction. A remote laboratory (ReMLab) used for electrical measurements in Politecnico di Milano is presented in [ 8] with its design and didactic specifications, which help the institution to provide to the students higher availability of experimentation resources with a reduced financial burden. In [ 10], the internet-accessible laboratory (AIM-Lab) provides ways of "instruction that cannot be replaced by simulation software packages" to globally and collaboratively teach and experiment on semiconductor device characterization. In [ 7], a distributed architecture is designed to integrate a set of specialized real laboratories used for teaching electric and electronic measurements in order to economically achieve educational variety and optimize the workload on experimentation assets. Finally, the researchers in [ 16] adopt distantly operated laboratories for their mechanical engineering curriculum and demonstrate through a comparative study that there is no significant difference between the instructional outcomes associated with the students who carried out the experiments remotely and those who performed the experimentation in-person.
From the side of RFID, the research in this field focused on the applications and uses of the technology rather than the education aspect of it. Moreover, while only a limited number of schools and universities instruct RFID, the majority of these institutions teach it through research or projects that are selected by students. In the US, organizations that offer classes in RFID include Indiana University, the University of California, Michigan State University, and the University of Houston. The global presence of RFID education and research is also well represented by laboratories at Cambridge University in UK, Feng Chia University in Taiwan, Fudan University in China, Keio University in Japan, the University of St. Gallen in Switzerland, the RFID Institute in South Africa, and the Australian Universities of Adelaide and Wollongong.
In an effort to teach RFID technology, different institutions follow different approaches. At Indiana University, after the creation of the first working RFID educational model in 2004, students and professors are able to experiment with RFID technology, construct interfaces for RFID-related systems, generate metrics, and teach RFID and its use with EPC systems [ 17]. In fall 2004, a mechanical engineering professor at the University of California established an RFID-focused course called the "Management of Technology, in which students pursuing degrees in business-related topics collaborate with students pursuing technology-related degrees." In this class, students were responsible for designing and implementing new business applications using RFID [ 18].
In East Lansing, the School of Packaging at Michigan State University offers a course on the utilization of RFID technology in packaging. It also has an RFID testing laboratory where both undergraduate and graduate students conduct independent testing for research projects. In 2005, "At least five Michigan State University students have completed their master's degrees in RFID research, in topics such as RFID in warehousing and supply chain applications packaging, and RFID systems design" [ 18].
In 2005, the University of Houston offered a class in RFID Programming as an elective in the Management Information Systems (MIS) Department, and it generated enough interest that led to the enrollment of 24 undergraduate students. The class was under the form of a comprehensive survey of the RFID technology and its business applications. A considerable part of the learning was based on reading material from industry publications and a mixture of corporations' white papers and case studies. Moreover, students took quizzes and carried out collaborative laboratory assignments, in which they experimented with different tag fixed positions and mobile readers, and implemented a back-end software infrastructure using a developer's kit. Furthermore, they wrote data to RFID tags, deployed readers in a laboratory setting using TAVIS middleware and Visual Basic, and learned ways of handling the generated RFID data. By the end of the semester, no student had dropped the class and it was then decided that the course would be offered the following year [ 18]. Although all this effort has been spent, it was not enough to generate the necessary RFID-skilled personnel [ 12], [ 17], and that is due to the fact that only a small proportion of an institution's students is being trained, and only a few organizations are providing this training. The online RFID learning environment developed in this document can provide a way to promote RFID by making it accessible to everybody as well as endorsing collaboration between schools and universities.
Following the footsteps of previous successful work, we provide a similar, but novel, system that targets teaching and evaluating RFID. To achieve that, the I-ATMUS 1 project efforts and its associated funding were spent in developing an educational tool that uses Web technology to give remote access to RFID laboratory resources (see Fig. 2). Such development targeted the following main milestones:
During and after the development stages, the implementers of the I-ATMUS project focused on three different aspects: 1) constructing a cutting-edge architecture for the system, 2) building a robust and inexpensive hardware setup that yields multiple configurations, and 3) developing a complete programming language (called NBL) specific to this system, while keeping its use optional in order to prevent any learning barriers.
Figure Fig. 2. Online view of the developed RFID testing system.
The developed online learning environment involves two educational layers. The first one is attained by providing the students with access to an RFID knowledge base that contains materials about the technology's features, limitations, and applications. This is supported by a website that includes instructional modules, a search engine, message boards, relevant links and articles, tutorials about utilized tools, and other web facilities that let the learner acquire the knowledge s/he needs in an effortless manner. The website can also be easily changed to satisfy new and different teaching needs and adjust to feedback and evaluations. A great advantage gained through this layer of the online learning environment is the interaction of students and stakeholders through discussion boards where detailed technical issues can be discussed in an informal way. Another benefit is the site being a perfect place to find inspiration for future projects and business applications.
The second layer of the learning environment, on the other hand, endeavors obtaining hands-on experience through the remote access of the RFID laboratory. In such a setting, individuals are able to reconfigure the hardware setup, conduct experiments, and acquire RFID read data. Furthermore, the collected data can be used for analysis and testing, which helps learners appreciate the various engineering tools that are utilized in evaluating systems, drawing conclusions, and making decisions.
The hardware design of the RFID laboratory system is based on the different requirements of testing RFID technology. The factors involved during the tests include the motion of RFID tags, their distance from the RFID antenna, the tag density in the RFID envelope, and the angle of the RFID antenna. That is why the system design incorporates degrees of freedom in the hardware that allow treating those factors as variables.
The hardware mechanism chosen to be implemented is in the form of a robotic system that has a set of motors and a control unit that cause the RFID tags to move linearly on parallel train-tracks (see Fig. 3). In addition, the hardware is empowered with a set of smart software applications called agents, which use internet connection to send messages to each other. These agents smoothen the interaction between the different system components such as the robotic hardware mechanism, GUI, database, and RFID readers (see Fig. 4). They also help maintain and upgrade the system in a time- and cost-efficient manner while insuring a high level of compatibility within device communication. This proves very valuable when a certain technological component becomes obsolete and is to be replaced without affecting the functionality of other system components [ 19].
Figure Fig. 3. The hardware setup of the RFID testing system.
Figure Fig. 4. High-level RFID laboratory system architecture.
In order to describe the architectural model (in Fig. 4), consider the following: while the reader agent reads RFID tags in a way that is dependent on its configurable reading mode, the captured reads are broadcast throughout the local (Intranet) network so that all interested parties can obtain such data. Two of these parties are the database agent, which stores the read data, and the GUI listener agent, which displays the data to the user. With the help of its control agent, the GUI can also discover the connected agents in the network, modify the configuration of the reader agent, query the database, and control the robot agent through hardware commands and scenarios (a scenario is a set of commands coded in the NBL programming language to program the motion of the tagged trains over time). The robot agent, on the other hand, receives these commands and scenarios, and takes care of distributing them among its subagents (motor agents and relay board agents) in order to create the desired set of motions and displacements of RFID tags. Note that all forms of communication in this system (commands, scenarios, requests, queries, and data) consist of messages that are encoded in Extensible Mark-up Language (XML). To know more about this agent-based architecture and the thorough details of its implementation, see [ 19], [ 20], [ 21]. Note that [ 19] also describes the used hardware devices and explains the step-by-step implementation process, which eases the task of replicating the overall system.
In such a setting, agents might, and usually do, reside on different computers/servers. As agent communication is a key factor for proper functionality of the involved system, messaging has to be particularly reliable. Two implementation issues may be faced when adjusting to this requirement. First, the security level on the connected machines has to be lowered to allow such messaging, which may cause a whole in the security. A way to alleviate this problem is by keeping the security level high, but manually adding an exception (to the Firewall) for each program that launches one or more agents, so that agents are able to choose a communication port and send/receive messages. The second issue is related to the hierarchy of the computer routing system. To be more specific, if two computers/servers are connected to the same Intranet through different routers, then one of the machines might not always receive messages from the other. To mitigate this risk, the establishment of a sublocal network through one router is suggested.
NBL is a programming language that was tailored specifically for this system to ease the use of hardware while trying to obtain a particular setup (such as the one shown in Fig. 5). Two important points about this feature are to be noted: first, the use of NBL is optional because the system interface includes soft buttons through which the mechanism can be straightforwardly controlled. This eliminates the issue of having learning barriers or attention dispersion for students. Second, NBL is easier to utilize than other more commonly used programming languages, such as C or Java, because it treats hardware scenarios as a set of English statements rather than function calls, and it uses a generalized (or standardized) way of defining objects (and that is simply by using the keyword "var") as opposed to typed languages in which the type of the variables (e.g., integers, doubles, Booleans, strings, lists, functions, objects, etc.) has to be defined. Therefore, the user deals with computational details at a higher (and easier) level and is able to put more focus on the hardware control should s/he opt to follow the programmatic approach.
Figure Fig. 5. The control webpage interface.
Among the benefits that result from the use of NBL are the significant reduction in mouse clicks, the omission of tediousness in acquiring specific setups, the construction of complex testing scenarios, and the option of reading RFID tags while in motion. Moreover, scenarios can be saved in files, and scenario libraries can be established and expanded for easier future use. Furthermore, NBL has computation aspects similar to the other well-known programming languages (e.g., Java); it lets the programmer use arithmetic operators, loops, if-else statements, lists, function calls, recursion, etc. Below is an example of an NBL scenario and its corresponding interpretation:
It is expected from this program to result in the following sequential motions:
The RFID laboratory testing system can be accessed through a control webpage that consists of four main modules (see Fig. 5):
Besides providing users with hands-on experience with RFID, this project is also expected to increase the understanding of the technology and its relevant areas, improve student attitudes about engineering education, and enhance their confidence toward any instructed technology. Surveys were used to assess the developed educational tool and its impact on learners, a fact that provides means to receive feedback and improve the learning environment and the operability of its laboratory testing system.
Junior level engineering students in Industrial Statistics and Manufacturing Systems courses (having 14 and 37 enrolled students, respectively) were assigned different exercises including reading assignments about RFID technology overview, opportunities, limitations and controversies, designing and implementing RFID solutions, major application areas, and emerging trends. In addition, group exercises were assigned to use the developed education tool for hands-on experiments. Student groups were asked to use
The expected outcomes of the assignments included:
These outcomes were measured, and students' learning and performance were assessed using a set of Pre- and Postmodule questionnaire. Students were asked to rate several statements and were given the opportunity to write personal comments about the most helpful aspects and the suggested improvement. Two sets of questions were given to the students. The first set was prepared to assess student attitudes about RFID technology whereas the second set focused on the use and effectiveness of the developed educational tool toward the improvement of these attitudes. The first set of questions was asked to the students before and after the assignments; the second set was asked only after the successful completion. Along with the premodule questionnaire students completed the Index of Learning Styles (ILS) questionnaire [ 22]. Results of the ILS questionnaire describe learning styles across four continua: active versus reflective learners, sensing versus intuitive learners, visual versus verbal learners, and sequential versus global learners. Styles are quantified across each continua on a supposedly interval scale ( $-11, -9, -7, -5, -3, -1, 1, 3, 5, 7, 9, 11$ ).
Industrial Statistics focuses on hypothesis testing, design and analysis of experiment, and regression. During the hands-on assignments, students in this course concentrated on collecting RFID data online, analyzing it using statistical tools, and drawing some conclusions on the capabilities of the technology. On the other hand, the Manufacturing Systems course deals with the analysis of manual, partially automated, and fully automated production systems. Students in this course focused on using the hardware setup and programming the system using NBL language. In the Industrial Statistics course, students were asked to use the webcam module, manual-control module, and RFID system module. In the Manufacturing Systems course, the utilization of the programmatic module was added to the assignment in addition to the above-mentioned ones. Therefore, hands-on assignments for these classes and the set of questions on how the learning environment helped students were slightly different.
In the first set of questions, students were given five options to rate each statement before and after the assignments: very unsure, unsure, neutral, confident, and very confident. The first set of questions includes survey results indicated that activities that involved using our learning environment improved their confidence in general knowledge about relevant topics including:
Having first arbitrarily assigned numerical values 1-5 to ordered categories "very unsure" through "very confident," respectively, we report in Table 1 median confidence before and after Phase 1 activities with respect to each of the topics detailed above. Median confidence improved in 6/8 topics, and average median confidence across topics improved greater than 35 percent. The two topics where confidence was not improved were consumer privacy and security issues (6), and data analysis (7). Consumer privacy and security were not issues we highlighted, and student self-assessment about data analysis while important was already (premodule) observed as "confident."
A more theoretically appropriate treatment of the ordinal data reveals statistically significant differences between pre- and postmodule confidence. Results in Table 2 are $p$ values associated with the binary variable that describes whether or not a response is postmodule. These binary variables of interest were considered independent variables in proportional odds models of responses about each topic controlling for student learning style [ 23]. We also fit an overall model of confidence as a function of whether or not it was postmodule, controlling for learning styles and topic.
Everywhere there was improvement from median pre- to postmodule confidence that improvement is significant ( $\alpha = 0.05$ ) with the exception of Topic 4 (business benefits through RFID). To summarize student confidence in the following five topics was improved after our module:
In the second set of questions, students were asked to indicate their agreement upon the following statements by choosing from five ordered categories: strongly disagree, disagree, neutral, agree, and strongly agree.
The median response to each of the seven items was "agree."
Responses for the second set of questions were collected after activities in which student teams used our learning environment to collect real RFID system data for a larger designed experiment. Teams then estimated, examined for adequacy, and selected linear regression models of the larger RFID system in order to understand read rate variation in terms of angle and distance between tags and antennas. Models also led to discussions about how one could use statistics to understand potential interference among tags. Three students did not participate fully due to absenteeism. They are not considered in our summary of results.
Having assigned numerical values 1-5 to ordered categories "very unsure" through "very confident," we report in Table 3 median confidence before and after Phase 1 activities in the manufacturing systems course with respect to each of the relevant topics. Median confidence improved in 8/8 topics, and average median confidence across topics improved greater than 52 percent. In every topic median confidence improved to the category of "confident."
Significance ( $p$ value) of module in describing pre- and postconfidence among Manufacturing Systems students is given in Table 4. In every topic, there was significant ( $\alpha = 0.05$ ) improvement from pre- to postmodule confidence.
Similar to the other course, in the second set of questions, students were asked to indicate their agreement upon some statements by choosing from five ordered categories: strongly disagree, disagree, neutral, agree, and strongly agree. The statements for this course included:
The median response to each of the 10 items was "agree."
For the hands-on assignment, students worked in groups and were asked to use the camera to acquire a good visual on the hardware platform. Then they were asked to collect data by moving the RFID tags and antennas using manual control. The third setup was to code some scenarios using NBL programming language and collect RFID data while the tags were mobile. Such scenarios had a certain degree of complexity that involved iteration loops (for-loops or while-loops) and variable declarations. Because students did not necessarily have experience with programming, they took a short course during a lecture to learn how to use the learning environment and NBL language.
Median confidence improved in 6/8 and 8/8 topics among Industrial Statistics and Manufacturing Systems students, respectively. Average median confidence improved greater than 35 percent among Industrial Statistics students, and 52 percent among Manufacturing Systems students. Theoretically appropriate statistical models of confidence show significantly improved confidence in 5/8 topics among Industrial Statistics students. Among Manufacturing Systems students there is significant improvement in confidence about all eight topics. Also in both classes, median responses to each of the positive statements about activities were "agree." Finally, there were no consistent, significant results with respect to learning style. In summary, the activities introduced to Industrial Statistics and Manufacturing Systems seem to have improved student confidence about relevant topics remarkably. Student attitudes about the activities were markedly positive. Results of statistical models are not confounded by variable learning styles; the activities themselves do not seem to appeal to certain learning styles.
In this paper, the authors present a learning environment that enables learners to access technological resources of an RFID laboratory through Web technology. Such a project has a highly automated testing system that is easy to use by beginner-level learners. It also has the advanced feature of controlling hardware devices programmatically in order to develop complex test scenarios rapidly and with less tediousness. After engaging students from two different classes in using this educational tool, student assessments were performed to measure its impact on learners. Results show that this online environment helped students learn many aspects of RFID technology and obtain hands-on experience through conducting test experiments, programming hardware scenarios, collecting and analyzing data, and drawing conclusions.
This material is based on work supported by the US National Science Foundation (NSF) under Grant No. DUE0633334. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF. The authors thank NSF for their support to implement the I-ATMUS project. The authors also thank Dr. Craig W. Thompson, Joseph E. Hoag, Alejandro Mendoza, Kevin J. Oden, Jonathan B. Marshal, Qilu Wang, Xavier S. Neely, Korbi E. Orr, and Dewanna Jenkins for their assistance in the I-ATMUS project.