RFID Data Processing & Management
Call for Participation: EII Workshop on RFID Data Processing and Management
Dr Xue Li
Tel 07-3365 2379 Fax 07-3365 4999
School of Information Technology & Electrical Engineering
The University of Queensland
Brisbane Q 4072
Dr Michael Sheng
SENIOR RESEARCH ADVISOR/MENTORS
Prof Peter Green The University of Queensland email@example.com
Dr Kerry Taylor, CSIRO ICT Centre firstname.lastname@example.org
Dr Elizabeth Chang Curtin University of Technology email@example.com
Prof Jiawei Han University of Illinois at Urbana-Champaign, USA firstname.lastname@example.org
Dr Tao Lin SAP Research Centre, USA email@example.com
Dr Chengfei Liu Swinburne University of Technology firstname.lastname@example.org
Dr Bela Stantic Griffith University email@example.com
Dr Dongming Xu The University of Queensland firstname.lastname@example.org
Prof Chengqi Zhang University of Technology, Sydney email@example.com
Dr Shichao Zhang University of Technology, Sydney firstname.lastname@example.org
Workshop on RFID Data Processing and Management DATE: Friday 31 August 2007
Radio Frequency Identification (RFID) is a wireless communication technology that uses radio-frequency waves to transfer information between tagged objects and readers without line of sight. This creates tremendous opportunities for linking various objects from real world. These objects are numbered, identified, catalogued, and tracked. In recent years, RFID has gained a significant momentum and is emerging as an important technology for revolutionizing a wide range of applications including supply chain management, retail, aircraft maintenance, anti-counterfeiting, baggage handling, healthcare, just to cite a few.
RFID technology is bring a revolution to the way that we track items in business applications such as manufacturing, the logistics distribution, various stages of supply chains, retail store, and asset management applications. The ubiquitous computing and sensor networks are now generating a very large volume of temporal and spatial data that are often with noises and duplicates. RFID tags are much more than barcode replacement in terms of base technology and the applications. RFID technology also provide many more advantages than the smart card technologies and poses many interesting challenges to the current data processing and management systems.
Understanding these new technology implications and its new applications is a key to enable Australian ICT professionals to face these new challenges to our society.
New Applications of RFID Technology: Once the cost of RFID tags is pushed down to its low prices as a couple of cents per tag, almost every human-made product or any item that moves can be tagged. Explosive number of applications would be identified. From airplanes maintenance, Supermarket shopping activities, to the cyber fridge that records milk or the frozen food package to communicate the appropriate cooking instructions to microwave oven. Even PDA can have built-in readers for personal interactions when visiting the places of museums, hospitals, airports, and tourist attractions. An RFID tagged world is emerging.
Scalability Issues: Large-scale RFID data will exist everywhere. There may be noises and many duplicates read by the readers that continuously read RFID tags in their ranges. Data are collected in a large volumes such that in a containers or in pallets during the transportations. Also all RFID data are temporal and spatial. The tracing of their individual paths would become a problem when the number of tags becomes large. An individual RFID tag has only primitive data to be recorded at a specific time point on its EPC, timestamp and its reader's ID. To an application, this piece of information is an important trigger that can bring an event to its corresponding actions to be carried out by the application system.
RFID Data Privacy Issues: Protecting privacy is ever becoming an objective and is getting much difficult when the RFID technology is becoming widely adopted. If an application can only read the RFID tags and the object that it owns, the privacy issues might be simplified. However, many applications do not in this category. For example in a supply chain or in a shopping centre, the ownership may need to be transferred or the tags are to be disabled. So the legislation issues, the data privacy issues must be resolved before this technology can be widely applied in the business world.
RFID Data Query Model and Event Processing: Probably the most critical problem to the RFID data is to have a computational model of the RFID data that can lead to a complete computational system to query the RFID data for their temporal and spatial properties.
RFID Data Warehousing and Data Mining: Warehousing RFID data is very different from traditional data warehousing as mentioned by Jiawei Han. Discovering patterns from a large collection of RFID data also presents a new challenge to data mining community.
Good data management methods are needed for these RFID applications, and some are available from database management technologies such as those from active databases, object-oriented databases, and spatial-temporal databases. However, such methods are not directly applicable to RFID data since its data type is simple (i.e. with the attributes of: EPC, Time, Location) and the data are dynamically collected as the both readers and tags may be constantly changing their locations.
A critical issue for both RFID data processing and management is on how to define a computational model that can process and manage RFID data in an evolving environment where the configuration of RFID readers is dynamic, the RFID tagged objects are constantly moving, and queries on the RFID data can be exhaustive: from the RFID network topology, RFID tags group behaviors, to the individual whereabouts and their movements histories.
In this task force, we will gather researchers from different disciplines and domains such as businesses, information systems researchers, computational and mathematical scientists, to initiate collaborative research in processing and managing RFID data in large-scale applications.
The main outcome of the taskforce will be continuing and cross-disciplinary research with a common theme of large scale RFID Data processing and Management.
This task force will NOT investigate on the RFID hardware technology or its working principles, hardware manufacturing, and implementation issues.
Processing and managing large and complex RFID Data and together with their underlying dynamically configured networks is a challenging research topic. More specifically we can identify the following research significance:
RFID Data Privacy and Policies: the main challenge of this research is the Data Privacy of corresponding policies. For example, after a customer purchased the items from a supermarket, the RFID tags should be permanently disabled by the supermarket readers. Different framework of privacy protection can be designed for the different types of RFID networks. In different RFID applications the privacy requirement may be different. For example when stray pet animals are in the public areas, the RFID readers should be allowed to read their tags for the identifications.
RFID Data Processing Scalability: the second challenge for this taskforce is the scalability. When a large quantity of items are moved or shifted together, the individual reading and processing of tags is time consuming and unnecessary. The design and mapping of the GIDs (Generic IDs) and their consistency checking, cross-reference indexing requires semantics and ontological support for the hierarchical data processing. A centralized approach normally suffers scalability problem. A distributed approach may suffer a search problem to satisfy queries.
RFID Data Management: A critical issue for the success of an RFID application is its computational data model and the query tools. That is, the RFID data read from a dynamically established RFID network need to be coherent in the data collection, cleaning, consistency checking, and data persistency management. Path queries on the RFID data temporal and spatial data and their histories present a challenge for an effective and efficient computational data model of RFID data. A full-fledged DBMS would be too much for managing RFID data. Thus a data management platform specialized in RFID data management will need to be considered.
RFID Network Dynamics: Real world RFID networks are always changing over time. Many RFID readers, like those built in non-moving objects, such as item shelf, change relatively slow over time. The dynamic configuration of RFID networks of their readers in topology and connections will make the data collection, cleaning, GID mapping become more complicated. In some cases, such as the fast moving objects, a very fast-streamed graph of RFID data can be expected. Effective and efficient modeling and analysis for dynamic RFID networks are challenging research topics.
- We will identify research opportunities in RFID Data Processing and Management for the EII network, focusing on the Australian context.
- We will form a research community with cross-disciplinary collaboration, including business, computer science, information systems, and mathematics, with a focus on problems in data processing and management issues.
- We will spawn continuing cross-disciplinary collaboration (that is, continuing after the lifetime of this taskforce).
- We will assist emerging researchers to find support for research in RFID Data Processing and Management. This involves helping with linkages to international researchers, industrial funding, and ARC Discovery-style grants.
- The main long-term objective is to spawn a viable high-end industry in software to handle large and complex RFID networks.