2006-12--RFID-Hardware--Reid-Phillips

University of Arkansas – CSCE Department
RFID Agent Middleware – Final Report – Fall 2006

RFID Hardware
Reid Phillips
Abstract
The focus of this project has been primarily centered on working with the agent
wrappers for the RFID hardware devices. From testing and debugging to development
a wide range of issues were partially addressed in this project.
The hardware wrappers contained within the RFID code are essential to the project as a
whole. These wrappers provide the interface by which the overall architecture can
communicate with devices in the outside world. This functionality gives a purpose and
thus overall direction to the entire system and therefore its correctness is essential.

Problem
The problem to be addressed is twofold: first it is necessary to ensure that the existing
code base remains up-to-date and that any discovered bugs be fixed, second it is
important to develop new code to support new hardware devices. The focus of this
project was to address these two problems.
These two issues are important for the project especially as it is going open source
soon. The first issue ensures that the established code base is in fact useful and
relatively error proof. The second issue improves the overall quality of the project by
expanding its capabilities. These features are important to any package being released
whether it is open source or not.
The RFID package would not be taken as seriously until these features were
addressed. If the released software was full of bugs it would be seen as
counterproductive to expand its functionality until a stable version was reached. Also, if
the software provided almost no functionality then again it may not be seen as
worthwhile to build off the existing architecture.

Objective

The objective of this project is to both test and debug existing code while also
expanding the code base by writing new agent wrappers.

Design

Page 1 – <December 05, 2006>


RFID Hardware

The existing architecture was used. The changes made involved adjustments to
individual lines of code within the code base.

Background
The problem is fairly straightforward. For most of the accomplished work, the majority
of the code already existed. Only minor changes were identified. Though the amount of
code written was small the necessary research required took more work. This required
access to the API’s of both the Symbol and Impinj readers and the Zebra printer. Until
this project I had not actually worked with the readers therefore requiring slightly more
research than the printer which was relatively fresh in mind from previous work done.
Beyond previous work done on the Zebra printer no background existed for this project.

Related Work
For communicating directly with the devices, generally related work consists of
proprietary solutions which typically entail a web page interface.

Requirements
Zebra printer
ZPL definition document
Alien EULA

Impinj contact (Mike Grobler)

Impinj, Inc.
5000 Birch Street, Suite 7100
Newport Beach, CA 92660
Tel: (949)553-0650x1056
Fax: (949)553-0615
Symbol reader (RFID lab)
Symbol reader API

Risks
No risks were considered.

High Level Design
No changes were made to the supporting architecture, only code revisions were
accomplished.

Tasks (as executed)
1. Test the Zebra printer at the RFID lab …
2


RFID Hardware

2. Obtain the Alien EULA …
3. Understand the Impinj architecture …
4. Discuss code details with Impinj contact …
5. Adjust Symbol reader agent wrapper …
6. Document …

Risks
N/A.

Detailed Design
The changes made to AR400ReaderAgent.java include altering the command issued to
the reader. Originally the “queryTags” command was issued with associated
parameters; now the “queryEvents” command is issued. This in turn changed the
response sent back from the reader to the agent and subsequently involved making
slight adjustments to the parse method that interprets the incoming message.

Implementation
Implementation involved adjusting the code within the existing code base.

Test Plan
Tests will be run in the RFID lab as soon as is convenient to those working there. Tests
will involve using the agent to make correct delta reads rather than system snapshots.

Results and Analysis
No problems were found using the Zebra printer agent. Several tags were printed
correctly while at the RFID lab. The agent performed as specified in the existing
TagCentric documentation.
Through the assistance of David Cromhout and Nick Jones at the RFID lab the Alien
EULA was obtained to ensure that distribution of some of their resources was possible.
Some time was spent with Mike Grobler understanding the Impinj API and how to work
with their libraries. Nothing ever came of this as the issue of using their products is still
being negotiated. Another associated issue is the understanding of JNI or some
comparative technology to enable Java to talk to a .dll or .so.
Also the API for the handheld Symbol reader was never obtained. Sean Bruce, through
his connections at the lab, attempted to obtain this document but was ultimately
unsuccessful.

Finally the Symbol reader agent, AR400ReaderAgent.java, was changed to allow for
delta reads rather than complete system snapshots. This code is finished but the
testing is yet to be performed.

Conclusions

3


RFID Hardware

Tests were run on the Zebra printer in the RFID lab indicating correct performance. The
Alien EULA was obtained for consideration concerning the move to open source. Only
a general understanding of the Impinj architecture was gleaned as most of the time
spent dealing with them was spent through lawyers. Finally the Symbol reader agent
has been edited to allow for delta reads rather than complete system snapshots.
I think an important organizational feature is to make sure all the documentation is
available before assigning a project. A lot of time was spent doing nothing while waiting
for API’s and other similar documents. Though it gave me more time to work on other
projects I did not accomplish as much for the RFID class.

Future Work
Once access is granted to the Impinj resources an agent should be developed for it.
This will require a different approach than exists in the AR400ReaderAgent.java. This
reader responds to http communication which is vastly different than what is required for
the Impinj reader.
Also development of agents to support handheld / mobile readers is an issue yet to be
addressed. Specifically the Symbol handheld reader should not be too different from
the existing AR400ReaderAgent.java.

Deliverables
•

Code – Changes were made to AR400ReaderAgent.java file that support delta
reads.

•

Final Report – This document.

Key Personnel
Student name – Phillips graduate Computer Science major in the Computer Science
and Computer Engineering Department at the University of Arkansas. He has
completed two semesters of this class working with both Impinj and Symbol readers and
a Zebra printer. He was responsible for the changes described in this paper.

4