Professor: Marek A. Perkowski, Electrical Engineering.

EE 479/579 Intelligent Robotics II (4).

CATALOG DATA - take from PSU catalog.

TEXTBOOKS:

The same textbooks as in last quarter for EE 478.
Philip John McKerrow, "Introduction to Robotics," Addison-Wesley Publishing Company, 1991.
Perkowski, slides and lecture notes, additional materials for projects.

COORDINATOR:

Marek A. Perkowski, Professor EE.

GOALS:

The course introduces techniques of image processing, computer vision, and planning that are used in robotics.
Combining ideas from computer architecture and AI is emphasized in projects and applications.
Students learn mobile robots, control, interfacing, planning, and using of various kinds of sensors.
Understanding of software/hardware tradeoffs is emphasized in a comprehensive group design project that includes both these components.
Students completing this course have a working knowledge of several advanced and applied concepts of intelligent robotics and are prepared to understand recent research and development results in this area.

PREREQUISITES BY TOPIC:

EE 478/578, or consent of the instructor.

TOPICS IN CATALOG:


1. Sensors. Computer vision hardware.

2. Low-level image processing; filtering, thinning, edge-detection, morphological processors, region growing.
Hardware and software.

3. Medium-level image processing; image transforms, image matching. Hardware and software.

4. Robotic computer vision, use of AI techniques.  Hardware and software.

5. Mobile robots ; guidance systems, path planning, collision avoidance. Use of reasoning and learning.

6. Task planning, robot languages. 

7. Robot programming.

8. Interfacing.

9. Computer architectures for vision and robotics, available solutions and components. Recent research.

10. Robots in health care. 

11. Robots in manufacturing. Manufacturing inspection and integrated systems. Combinatorial problems.

12. Discussion of projects and assignments. System integration.

LAB PROJECTS:

The projects vary from year to year and include: designing and testing specialized computer architectures using Field Programmable Gate Arrays or microcontrollers, writing image processing, path planning, movement/sensor controlling programs, sensor integration. The programs include: image processing, path planning, obstacle avoidance, movement planning, sensor integration. The design projects include specialized computers for solving logic problems, image matching, convolvers, sonar controller, stepper motor controller.

COMPUTER USAGE:

Students use the departmental network of Sun workstations and PC-based computers to solve various problems. Several projects are also related to PSUBOT - a PC-based wheelchair robot designed in our department.

ESTIMATED CONTENT:

Engineering Science: 2 credits or 50%.
Engineering Design: 2 credits or 50%.

Additional information.

I use also the book from Winter quarter in this class as well as my additional notes are available to students in "Smart Copy". The notes include also all projects and project-related reading materials.
The students have to write programs and essays that are emailed to me on a weekly basis. In addition, students work on a comprehensive group project. This project differs from year to year, and usually takes more than one quarter.
This year there are two groups of projects:
- image processing and computer vision of color camera images for mobile robotics applications.
- machine learning by logic decomposition and composition methods.

Base of the grade:

Homeworks and exams: 30 %
Individual project reports and programming assignments: 30 %
Oral presentation of student's research in class: 20 %
Group projects and written group reports: 20 %

MORE DETAILED INFORMATION.

 Sensors.  Touch sensors, magnetic sensors, optical sensors,
infrared sensors, sonars. Control of sonars.
Complete Sonar-Based Imaging system for PSUBOT (PSU robotic wheelchair)
- hardware and software. 
 Computer vision hardware. 
Cameras. Image grabber, look-up table architecture.
Complete imaging system of PSUBOT.

 Low-level image processing.  
Noise removal techniques, Convolution-based methods, Sobel, Prewitt and other linear
filters, dilation, erosion, 
 morphological operations and filtering.  
Discussion of median filtering. Approaches to thinning lines. Labeling.
Fast algorithms for region growing and manipulation.
Methods for edge detection. Hierarchical approaches, hardware realizations.
Programming assignments in C. Integration of tools. Application-related exercises.

 Medium-level image processing.  Hough Transform for lines.  Generalization to algebraic curves. 
Adaptive and hierarchical Hough Transforms. Fast Fourier, Fast Cosine and Fast Walsh transforms.
Pipelined and vector architectures. Hardware realizations. Field Programmable Gate Array computers
for image processing. Approaches to fast image matching; geometrical and topological.

 Robotic computer vision, use of AI techniques. 
Scene recognition.
Semantic networks and reasoning by analogy.
Knowledge-based approaches to vision. 
 Machine learning applied to vision. 
Approaches to machine learning and their hardware/software realization.
Waltz algorithm. Statistical approaches. Fuzzy logic.

Mobile robots; guidance systems,  path planning.  Collision avoidance. 
Artificial Life approaches. Use of fuzzy logic, high-order logic, knowledge-based
and learning approaches. Knowledge representation for planning and collision
avoidance. Corridor following, maneuvering.  Localization.  Optical feedback.

Task planning, robot languages. Examples of VAL programming.
PUMA robot programming. Programming the assembly problem.
Programming navigation in 2D and 3D space.

Robot programming.  VAL and other languages.
Writing robot tasks in LISP, PROLOG, WILD LIFE, C or C++ (depending on
applications). Programming a robot to stack boxes and 
perform simple assembling tasks.  Programming a model of a walking robot.

Interfacing. Use of Field Programmable Gate Arrays (FPGAs) and
Programmable Logic Devices (PLDs). Use of co-processor boards.
Interfacing with FFT Processor from Sharp. Analog and digital boards.
Fuzzy controller. Sonar interface. (Discussion of interfaces varies from year
to year and is related to current project).

 Computer architectures for vision and robotics. 
Pipelined image processors, convolvers.
Neural networks. Fuzzy logic controllers.
DSP processors. Processors for solving logic,
matching and other combinatorial problems: Cube Calculus Machine.

 Robots in healthcare 
Robot for elderly, robots in hospitals, robotic technology to aid surgeans.
Various categories of robots: robots in battlefield.

 Robots in manufacturing. 
Uses of robotic technology in manufacturing: stationary versus mobile robots.
 Applications in scheduling, planning and assignment. 
Incorporation of smart sensors. 
 Integrated manufacturing systems.  
Virtual Reality. Robots in entertainment.

 System integration.  
 Examples of applications.  
Discussion of projects and examinations.
Much time is spend to discuss student progress on projects.
Programming in Lisp and Prolog, understanding of recent research
papers assigned to read and present.
Demonstrations of software, hardware and videotapes with previous
projects from PSU, MIT Robotics Labs, Wright Laboratories, and other.