Professor: Marek A. Perkowski, Electrical and Computer Engineering.

ROBOT LEARNING, VISION AND PERCEPTION

This class is the same as Intelligent Robotics 2. Just this new title reflects better what is our concentration in the class. We cover also topics not directly related to vision and perception.
We start with easy and very practical methods for early robot vision. We present basic Machine Learning methods used in robotics. Class covers also Neural Nets, Deep Learning and search. Next, more advanced methods of Machine Learning and Robot Vision are covered, that can possibly lead to thesis, for instance advanced topics such as Fast Fourier Transforms and Radon Transforms. All these theories are used in practical robot vision projects. The last part of the class is robot navigation and localization, but this may be not finished this term and will continue to the next term.

Goal of the class.

Students should learn elementary, medium and some advanced topics in robot vision and perception to be used in software, hardware and integration/robotic projects. Students shoul continue learning about robot design and some Machine Learning aspects of robotics. See the syllabus.
Official Syllabus.
This is the second class in robotics but can be taken with no prerequisities in case that the students knows programming. Students should get practical understanding how robot vision and navigation algorithms are used in robot systems. They should also learn fundaments of probabilistic.

Where is this knowledge useful?

Previous students who have done projects found this class useful to find industrial positions in the following areas: embedded systems, robot vision, flexible automation systems with vision, medical image processing, hardware design of image processors, bank image processing, general software design, industrial board-level design, design automation, industrial robotics.

Software.

Software used in class depends on every year projects. This quarter the projects will use Matlab, Python, OpenCV, Visual Basic and C/C++.

Grading.

Grading is based in 85% on project. Project is splitted to three "homeworks". The remaining 15% are based on quizzes, class attendance and presentations. Class attendance is expected and checked. This time I already allocated students to project teams. In most cases they follow the Fall projects.

SLIDES AND INFORMATION FOR CLASS.

PROJECTS RELATED MATERIALS
PROJECTS FINAL.docx
1. Funny Robot Communicator Monster.docx This is a prototype paper to be implemented in secveral projects of this class and completed. Students have to present this to me and others, so I will be sure that everything is clear and everybody understands the project.
2. Robot Theatre.docx This is a prototype paper to be implemented in secveral projects of this class and completed. Students have to present this to me and others, so I will be sure that everything is clear and everybody understands the project. Similar to the above paper.
Team 7 Frankenstein Monster.docx Old report that will be enhanced.
Dancing_Hexapod_Final_Report (2).docx For the teams working on Monster and Copernicus.
quantum robots for teenagers.pdf For the teams working on Monster and Copernicus.
theoretical advances AIS.pdf This is a paper for the Team A of Mohammed.
85.-Leo_Deng-Extended_Abstract.pdf For all teams, example of a paper written by a high school student, good quality.
PROGRAMMING RELATED MATERIALS
GOOD_AI in PROLOG, LISP, JAVA - Luger_0136070477_1.pdf Great source of codes and ideas.
GOOD_Russel=Norvig=Artificial Intelligence A Modern Approach (3rd Edition).pdf Famous textbook, many ideas.
CLASS OVERVIEW
2017_0002_Overview of vision.ppt No details. What are the problems and ideas in vision.
MATLAB RELATED MATERIALS
This material is auxiliary for students who want to improve their knowledge of Matlab.
writing fast Matlab code tutorial.pdf
2017_0010_Matlab_tutorial 1.ppt
2017_0011_Matlab_tutorial 2.ppt
2017_0012_other files in Matlab.pptx
2017_0017_Matlab for vision.pptx
external interfaces to Matlab tutorial.pdf
Math review for Matlab users.pdf
Matlab programming_tips.pdf
matlab_programming_tutorial.pdf
MORPHOLOGICAL METHODS
Class material. Some will be skipped, use it only if project related.
2017_0022_intro binary morphological-dilation_erosion.ppt
2017_0027_big example.ppt
2017_0028_hit and miss - thinning - thickening.ppt
2017_0029_Thinning - advanced.ppt
2017_0034_morphology mathematics Theory.ppt
2017_0035_MATLAB MORPHOLOGICAL FILTERING.ppt
2017_0054_Advanced Morphology.ppt
2017_0122_grey scale morphological (1).pptx
MACHINE LEARNING
Decision-trees.ppt this is example of one ML method but with discussion of general methods, useful for projects even if you will use only Decision Trees.
Occam Razor vs Vapnik in ML with Kmaps.pptx Here we show how to explain ML concepts using Kmaps.
logic synthesis approach to ML.ppt Here avarious ML ideas are explained and illustrated with simple examples. SOP algorithms. This can be used in advanced projects and MS/PHD theses.
brief survey of ML.ppt Survery of all ML methods, just to give you hints about more advanced software that you may used in projects.
Intro Decomposition.ppt Introduction to Ashenhurst-Curtis Decomposition and similar methods.
Column Multiplicity - Coloring and Decomposition.ppt
AC decomposition-for-ML -advanced.ppt
review of probability.pptx These are mandatory slides about probabillity and its use in several ML methods.
probability exercises.ppt
naive Bayes classifier.ppt
naive Bayes 2.ppt Additional Reading.
basic Probability and Bayes - Criticism.pptx Additional Reading.
QUIZZ theoretical
QUIZZ software

AUXILIARY SLIDES AND INFORMATION.

MORPHOLOGICAL ALGEBRA AND IMAGE PROCESSING

2015_0003_Introduction_to_Morphological_Operators.ppt The most basic introduction to morphological image processing.
0002. Mathematics of Binary Morphology. ppt
0002_A. Binary Morphology and Use in Robotic Path Planning. ppt
0003.A. Big Example of morpology and measurements on image. ppt

LOW LEVEL IMAGE PROCESSING

SHAPES, COLORS AND HISTOGRAMS

EDGES and CONVOLUTION.

FUNDAMENTS OF SPECTRAL METHODS

SPECTRAL TRANSFORMS FOR IMAGES.

PROBABILISTIC ROBOTICS AND BAYES CLASSIFIERS

MOBILE ROBOTS, KALMAN FILTERS AND SLAM

BAYES NETWORKS and PARTICLE FILTERS.

ROBOT THEATRE AT PSU

Individual reading, not covered in class.

Text of the play "Quantum Consciousness" for our robot theatre. We are realizing Act1 of this play in Winter quarter.
Text of research paper about robot theatre, written by Marek Perkowski and Dorota Zukowska. Some futuristic ideas of robot theatre. We will realize some of them during this and next quarters.
Text of PhD dissertation proposal of Mathias Sunardi. Do not share it outside the class. This is a theory behind artistic emotional robot motions and behaviors.
Intervewiew with Marek Perkowski for the international magazine about puppet art. State of the art of robot theatre from my perspective. Other interesting papers about robot theatres in Poland. Every paper is in both Polish and English language.
2015_0047_Some Plays of Portland CyberTheatre.pptx
2015_0050_Towards Future Robot Theatres.ppt

HISTORY AND TYPES OF ROBOT THEATRES WORLDWIDE

QUANTUM ROBOTS

EXAMPLES OF SIMPLE CODES

SCHROEDINGER CAT DOCUMENTATION

ARMS FOR HUMANOID ROBOTS

Humanoid Robot example.pptx Here are the examples of how to describe kinematics for your robot. FOR HOMEWORK 1.
Examples of Humanoid Robot arms-hands-ex.pptx Just to enthuse your imagination.
About Levers. Screws. Wedges. Principle of MA.ppt High-School stuff to be refreshed. If you work on robot design this quarter: think how you can use this knowledge practically on your robot.
Power, Torque and robot Arms.ppt Helps to design a good robot arm of small cost.
Pulleys.ppt">

CEILING CAMERAS FOR ROBOT THEATRE

Preliminary project assignments for this quarter. Will be updated.
2015_0001_Vision for Curie and soccer.ppt The explanation of some vision projects for this quarter.
2015_0004_Homework 1 for Class.ppt Morphological processing, edge detection and other simple algorithms for ceiling and front cameras for robot theatre