Abstract

Computer based controllers are being used increasingly in every aspect of modern life and have been integrated into everything from simple thermostats to the most complicated fighter jets. Multiple methodologies can be used within these controllers including differential equations, discrete difference equations, and fuzzy logic. Also, they can run on platforms ranging from scaled down versions of consumer operating systems to directly on embedded processors. What's missing is a low cost framework that links the variety of methodologies used in controllers and the variety of platforms that controllers can run on. My thesis is that it is possible to create a real-time control systems framework that is portable across platforms, is extensible, and has a low cost. The approach to make the framework multi-platformed and to allow it to use multiple control methodologies was to build a software system with a layered design. This was done so that the platform specific information could be isolated in its own layer, the control methodology information could be isolated in its own layer, and the services provided by the framework could be loosely couple from both of them. To solve the problem of price it was decided to make the project open source, which had the added benefit of allowing students, scientists and even corporations to improve upon the system in the long run. The solution that was implemented is called RTCSF, Real-Time Control Systems Framework. It runs on two different platforms, RTLinux, and a preemptable version of the base Linux 2.6 kernel. Two control methodologies were also implemented in RTCSF, one uses discrete difference equations, and the other uses fuzzy logic. It reached the goals of being multi-platformed, multi-methodology, and low cost, although it has limitations such as being only soft real-time on the 2.6 preemptable kernel and when running the fuzzy logic methodologies.

Library of Congress Subject Headings

Real-time control; Cross-platform software development; Open source software

Publication Date

2005

Document Type

Thesis

Student Type

Graduate

Degree Name

Computer Science (MS)

Department, Program, or Center

Computer Science (GCCIS)

Advisor

James Vallino

Advisor/Committee Member

Leonid Reznik

Advisor/Committee Member

Juan Cockburn

Comments

Physical copy available from RIT's Wallace Library at TJ217.7 .T43 2005

Campus

RIT – Main Campus

Download

Share

COinS