Post-Doctoral Research

RFID (Radio Frequency Identification) can play a primary role in achieving the visions aimed at by recent computing trends such as pervasive computing, Internet of Things (IoT) and tangible computing. RFID can transform ordinary physical objects into 'digital' objects  by simply attaching to them RFID tags, thus enabling to explore a physical approach, where information storage and processing are supported by physical objects themselves. 

 

Nevertheless, current RFID systems are not utterly reliable as they may fail to identify all the objects present in their RF field, particularly when dealing with bulk object identification, in ad hoc situations (where conditions are unfavorable for RFID tag reading, and thus for object identification). 

 

To address the above issue, I investigate the implementation of distributed data structures over RFID tags, that can be used to identify missing objects and recover lost data, as well. More specifically, high-level information distributed on RFID tags are used to describe various properties of physical objects. These properties may include: integrity information, group information, and the intrinsic properties to objects that are relevant to user application.

 

In this context, I use erasure coding techniques (generally used for reliable data transmission over the network) to implement the concept of object grouping, i.e., logically coupling objects thus forming a consistent group. This concept allows to determine missing objects by identifying only a subset of the group.

 

My research lead to the introduction of EraRFID, an erasure coding approach for RFID object grouping. EraRFID models the problem of identifying missing objects as an erasure coding problem, and solves this problem using the well-known Reed Solomon (RS) codes. The efficiency of EraRFID in terms of timeliness, storage overhead and object identification rate, is validated through extensive experimental studies performed using a sophisticated RFID testbed.