Professor: Marek A. Perkowski, Electrical Engineering.
THE TOPICS TO BE DISCUSSED IN CLASS OR ADDRESSED IN PROJECTS.
- INTRODUCTION TO ROBOTICS, MACHINE LEARNING AND ARTIFICIAL INTELLIGENCE.
- Read Luger's book.
- Read description of our class projects funded by Intel.
about our Intel's grant.
- Read recent papers and WWW Pages related to hobby electronics and hobby robotics.
- Read my WWW pages about Data Mining and Functional Decomposition.
- Think why you are interested in robotics, what is your dream project, what you
want to learn in this class, what you want to do as your mandatory project.
Create a short presentation in PowerPoint about these ideas. Send me by email
or give me a diskette in the class. I do not accept other documents.
- STANDARD PROJECT NUMBER ONE: DESIGNING SIMPLE ROBOTS USING TOYS AND KITS.
Project leaders Matt Maple, Michael Kennan.
The material to be learnt in these projects is the following:
- assembling a simple robot from a kit or few kits; 5 - 7 hours at most.
- selecting sensors from the catalog, ordering them and adding to the robotic arm; 2 - 5 hours at most.
- connecting to a PC through parallel port and learning how to use and adapt the existing
software; 2 - 5 hours at most.
- writing new robotics software in C, C++, LISP, Prolog or any other language that you know.
Time depends on the project complexity; 2 - 5 hours at least.
- programming a EPLD, PAL or FPGA with the program of robot's behavior; about 7 hours.
Currently, we dispose the following kits:
- OWI robotic arms.
- Robix robotic kits.
- LEGO Mindstorms robotic kits.
- LEGO add-on kits to Mindstorms.
- Capsela robotic kits.
- Second-hand electric components (motors, sensors, interfaces) and Erector sets.
- Pneumatic hand based on artificial muscles.
- Cars and platforms.
For examples of previous project of this type look to:
Projects from summer 1999
- STANDARD PROJECT NUMBER TWO: ROBOT FOR MANUFACTURING AND TEST.
Project leaders Anas Al-Rabadi, Justin Kam, Ron Fehlen, Xiong (Bear).
These projects use the Rhino robot with connected sensors, assemblies and electronics.
The material to be learnt in these projects is the following:
- Parallel port and serial port programming.
- Simple robotic languages; stand-alone and embedded in Basic and C.
- Simple robotic languages embedded in Lisp and Prolog.
- Control of movement, equation.
- Role of sensors.
- Camera: edge detection, noise removal, finding straight lines using Hough Transform.
- Test; fault models, test generation, fault location, diagnosis, measurements, bed of nails.
- Voice control and voice synthesis.
- Simple system integration.
Variants:
- Dice movement control.
- Test by a robot.
- Fault localization by a robot.
- Self-repair of a simple PCB by a robot.
- Robot using vision solves the "Man, Wolf, Goat and Cabbage" puzzle wihout human programming.
- Robot using vision solves the "Missionaries and Cannibals" puzzle wihout human programming.
- Robot using vision solves the "Wolf and 4 Sheep" game wihout human programming.
- Robot responds to voice in "Block World", using vision it solves "Tower of Hanoi" or similar problems.
- Robot that paints oil painting or spray paintings from examples.
- Robot sculptor: Dremel tools, camera, go.
- STANDARD PROJECT NUMBER THREE: ROBOT THAT LEARNS.
Project leaders Michael Levy, Eli Cabelly.
These projects assembly various human-like or animal-like robots
with connected sensors, assemblies and electronics.
The material to be learnt in these projects is the following:
In all projects many cameras, microphones, sonar sensors, infrared sensors and temperature sensors
are used.
- Robot makes faces, shows emotions while responding to voice sentences from a human.
- Robot responds to voice, recognizes people, mimicques them.
Variants:
- MUVAL: this is an ugly creature with four legs, fat belly and a face from a Halloween mask.
It sits on a desk and responds to your commands.
This is an ugly creature with four legs, fat belly and a face from a Halloween mask.
The name cames from MUltiple-VALued logic, because it reasons using multiple-valued logic.
There is currently a competition for the best head of the MUVAL.
Three students participate.
- GORILLA: the name tells how the robot looks like.
- MECHANICAL ROBOT OF THE THIRTIES: looks like from old movies.
- THREE BEARS: they talk to you, to one another, and they play music.
Variants of learning:
- Constructive Induction by Decomposition of Multi-Valued relations.
- Decomposition of Fuzzy Functions and Relations.
- Decision Trees and Diagrams.
- Probabilistic Neural Nets.
- Rule-Based search with Pattern Recognition used in cost evaluation for rules.
- STANDARD PROJECT NUMBER FOUR: MOBILE LEARNING ROBOTS.
Project leaders none this year.
- Navigation. Maps.
- Use of controllers.
- PSUBOT: A Robotic Wheelchair for Blind Quadruplegic.
- PROGRAMMING IN LISP.
- Read Perkowski/Mishchenko/Al-Rabadi book.
- Read Luger.
- Read KCL manuals on line or manuals of your Lisp on PC or Macintosh.
- PREDICATE CALCULUS AND RESOLUTION METHOD.
- Read Perkowski/Mishchenko/Al-Rabadi book.
- Read Luger.
- Read Prolog manuals on line or manuals of your Prolog on PC or Macintosh.
- STATE SPACE SEARCH.
- Read Perkowski/Mishchenko/Al-Rabadi book.
- Read Luger.
- Read Prolog manuals on line or manuals of your Prolog on PC or Macintosh.
- HEURISTIC SEARCH, METHODS FOR REPRESENTATION OF PROBLEMS AND ADVANCED DATA STRUCTURES.
- Read Perkowski/Mishchenko/Al-Rabadi book.
- Read Luger.
- PROGRAMMING IN LOGIC AND CONSTRAINTS
- Read Perkowski/Mishchenko/Al-Rabadi book.
- Read Luger.
- Programming in PROLOG.
- Programming in WILD LIFE.
- Programming in Constraints.
- Reasoning by analogy and induction.
- ADVANCED ROBOT PROGRAMMING
- Mobile robots ; guidance systems, path planning, collision avoidance.
- Robot programming.
- Task planning, robot languages.
- COMPUTER ARCHITECTURES FOR LOGIC PROGRAMMING AND
ARTIFICIAL INTELLIGENCE
- Read Perkowski/Mishchenko book.
- Cube Calculus Machine.
- Eye for the robotic dog.
- Analog FPGAs.
- IMAGE PROCESSING AND COMPUTER VISION
- Introduction to image processing for robotics application.
- Introduction to machine learning and computer vision for robotics applications.
- Elements of Pattern Recognition.
- Sensors. Computer vision hardware.
- Low-level image processing; filtering, thinning, edge-detection, region growing.
- Medium-level image processing; image transforms, image matching.
- Robotic computer vision, use of AI techniques.
- DATA MINING AND MACHINE LEARNING
- Data Mining and Knowledge Discovery in Data Bases.
- Rule-based systems and expert systems.
- Fuzzy logic and applications.
- APPLICATIONS OF ROBOTIC TECHNOLOGIES
- Manufacturing robots.
- Military, fire-fighting, police.
- Entertainment robots.
- Robots in health care and manufacturing.
- House-hold robots.