|
Summaries and links to the papers - both published and
yet-to-be-published along with my personal comments.
Author:
Oleksandr Mylyy Summary: RFID make it possible to create a
"physically linked world" and does not require either line of sight
or contact. It mentioned the components of an RFID system,
characteristics of RFID data - simplicity, large volume, temporal,
inaccuracy and so on - along with principles and a proposed layered
architecture of effective RFID data management system. It moves on
to discuss algorithms for data cleaning - including smoothing and
duplicate elimination - after which, it talks about redundant reader
elimination (or power saving) and unsatisfactory results of current
algorithms (NP-Hard). It finally formalize the RFID event detection
through a proposed language and concluded.
Author(s):
Hector Gonzalez,
Jiawei Han, Xiaolei Li,
Diego Klabjan
Summary: They propose a warehousing model to preserve object
transition compressing and using path-dependent aggregates. (Items
moving together and application requirement dictate bulky abstracts)
It discusses different forms of compression that can be achieved in
SCM systems at the cost of processing powers etc. It listed out:
- Structure of the input data - and data cleaning method
through merging the readings to get a time-in/time-out tuple
instead of storing individaul periodic records.
- Compressing and generalization methods for RFID warehouse
design - through taking advantage of bulky movements, data
generalization and collapsing of path segments.
- Algorithms for constructing one - from lower level ones,
- Methods for efficient processing of RFID queries and
- Experiment results
|
