A novel image processing algorithm for pattern recognition based robotic applications, Chandresh Chaudhari

The proposed thesis is aimed at developing a new image processing algorithm for video based indoor robot navigation. A novel image processing approach will be researched and analyzed to simplify the task of image understanding and thereby make the robotic system cost effective without compromising performance. The autonomous robot will be tested for the task of numbered door identification in the indoor environment using the researched techniques. The new technique will be compared with existing techniques in the field.

H-infinity control applied to mobile robots, (pdf 940KB), Nuha Nawash

This is a project to implement embedded controls in a Microchip microcontroller. Control algorithms that are being investigated include nonlinear PID and H-infinity methods.

Fuzzy logic control for an autonomous robot, (pdf 4.08 MB), Vamsi Mohan Peri

Objective of my research is to build a robot with fuzzy logic to control its motion. The robot in discussion has to follow a predefined path with minimum of deviation. A Microchip PIC16F877 microcontroller is being used in this process integrated with a few ultrasonic sensors (to measure the distances).

The robot I initially designed had no motion controller built in it and thus was prone to some errors in its movements thus slowing it down. To avoid this I am working on trying to incorporate a fuzzy logic controller into it. The purpose of choosing fuzzy logic instead of a proportional controller was when I simulated my robot model I saw that fuzzy logic was giving me better results.

I initially built a Simulink model to simulate the robot and then I tested its performance using two different controllers (P and Fuzzy Logic). After it was clear that the fuzzy logic controller was working better I programmed my microcontroller. This was done using a PIC C compiler. Due to memory and speed constraints I had to add another microcontroller and I am testing the whole system right now.

This is the result obtained when comparing fuzzy and proportional controller. It is clear from the above graph that the system has reached a steady state lot faster using a fuzzy controller than a Proportional controller. The P- controller was tuned with Genetic Algorithms for optimum results.

Mailing Address
Cleveland State University
Embedded Control Systems Research Lab
2121 Euclid Avenue
Cleveland, OH 44115-2214

Phone: 216.875.9670
Fax: 216.687.5405

Web Content Contact
Saurabh Jain

Web Technical Contact
Sunil Gavini