University of Dayton
Computer Science Department

Computer Science Colloquium
2005/2006 Academic Year

Fall Semester Dates:     Tuesday, 9/20/2005,  Thursday, 10/6/2005,  Friday, 12/2/2005.  (spring semester tba)

Colloquium Coordinator:  Dr. Jennifer Seitzer

 


Tuesday, September 20, 2005;   3:00pm
Science Center Auditorium
Reception:  Computer Science Conference Room

 

 

Autonomous Mobile Robot using Fuzzy Logic

 

Priti K. Gaonkar

 

Department of Computer Science

Wright State University, Dayton, OH

 

 

ABSTRACT

An autonomous mobile robot implies that the robot reaches a set destination on its own, without human intervention, in an environment not specially designed for it. In this research a fuzzy navigational algorithm using a mapped environment is used to develop the intelligent system. The navigational algorithm uses a layered motion controller. The motion controller is made up of three layers. The first layer, which is the Orientation Layer, assists the robot in keeping it pointed in the general direction of the goal frame to achieve the final destination. The next layer, which is the PD (Proportional-plus-Derivative) Control Layer, directs the robot through passageways efficiently. Depending on the type of passageways there are different modes the robot may use: Wall-Hugging, Centering and Blind modes.  The third layer is the Obstacle Avoidance Layer. Ultrasonic sensors are used for obstacle avoidance and to counter unexpected changes in the environment. A basic platform has been developed for this robot that uses a modular approach. It consists of three separate modules which are the supervisor module, the motor driver module, and the sensor module.

 

 

Ms. Priti K. Gaonkar

Priti K. Gaonkar obtained her BS in electronics engineering from the University of Pune, India, in 2002. She is currently pursuing her MS degree in electrical engineering at Wright State University. Her research interests include fuzzy logic, industrial and mobile robots.

 

 


Thursday, October 6, 2005;  3:00pm
Science Center Auditorium
Reception:  Computer Science Conference Room

 

Data Mining in the Real World


Nick Payne, Ph.D.
 Private Consultant

 

 

ABSTRACT

It is said that Data Mining and Statistical Data Analysis use different tools.  What I hope to show is that one can and should use all the “statistical” tools when mining, just with a different mindset.  Also, a major data analytic issue overlooked in both Statistical analysis and Data Mining is differences that can exist between variables of interest and recorded data.  They often have different characteristics.   We will examine these as well as view an application that is currently being used by the Cincinnati Police Department.

 

Dr. Nick Payne

 

Nick Payne has his  PhD in Chemical Engineering from the University of Cincinnati where he worked in the mathematics of demographic age distributions.  He worked as a research scientist in a long, fruitful career at Procter & Gamble between from 1962 to 2002.  He is now consulting – doing work in data mining and statistical analysis. 

 

 

 

 


Friday, December 2,  2005;  3:00 pm
Science Center Auditorium
Reception: 
Science Center Atrium following the talk

 

 

Evolving Pattern Recognition Systems

Dr. Mateen Rizki

Department of Computer Science

Wright State University, Dayton, OH

 

 

ABSTRACT

The availability of inexpensive sensors coupled with the rise of the internet has led to a rapid expansion in the amount of data available for analysis. Although there are a myriad of uses for this data, one of the most common applications is pattern recognition. The traditional approach to creating pattern recognition systems is human intensive requiring experts with training in pattern recognition to collaborate with experts who have knowledge of the problem domain to develop a custom recognition system for a specific problem. In contrast, our work has focused on the design and implementation of software tools and techniques that use evolutionary computation to synthesize recognition systems from raw data. Thereby reducing the personnel requirements and time needed to deploy new recognition systems. In this talk, I will give an overview of evolutionary computation, describe our approach for evolving recognition systems, and present results obtained using our techniques for a variety of problems in image processing, radar target identification and gene marker selection.

 

Dr. Matt Rizki

 

Dr. Mateen Rizki received his B.S. degree in computer science from the University of Michigan in 1981.  He received his M.S. degree and Ph.D. in computer science from Wayne State University, Detroit, in 1982 and 1985, respectively.  Dr. Rizki was a lecturer at Wayne State University in 1984-85 and join Wright State University in 1985 where he is currently a professor in the Department of Computer Science and Engineering. Dr. Rizki’s areas of expertise include evolutionary computation, data mining, pattern recognition, and image processing.  He has published over 50 refereed articles in these areas and has served as principal investigator on numerous federal contracts and grants. Dr. Rizki serves on conference program committees for SCI, PPSN, LCV, and CEC and is an associate editor for the journals BIOSYSTEMS and IEEE Transactions on Evolutionary Computation.  In addition, he is a member of Tau Beta Pi, ACM, SPIE, IEEE, and Sigma Xi.

 

 


 

Past colloquia:

·        Fall 1999

·         Spring and Summer 2000

·        Fall 2000 and Spring 2001

·        Fall 2001 and Spring 2002

·        Fall 2002 and Spring 2003

·        Fall 2003 and Spring 2004

 

Back to Computer Science Homepage CPS Home