Bachelor Artificial Intelligence

Course Autonomous Mobile Robots

This is the information of Fall 2013

The information of the previous year could be found here

Description

The description is available in the course catalogue with code AUMR6Y. The course is a free choice in the Bachelor Artificial Intelligence Curriculum.

This course gives an introduction in the fundamentals of mobile robotics, spanning the mechanical, motor, sensory, perceptual, and cognitive layers the field comprises. The focus will be on the mechanisms that allow a mobile robot to move through a real world environment to perform its tasks. It synthesizes material from the fields of kinematics, control theory, signal analysis, computer vision, information theory, artificial intelligence and probability theory.

This course is based on the book 'Introduction to Autonomous Mobile Robots', from Prof. Dr. Roland Yves Siegwart, Prof. Dr. Illah R. Nourbakhsh and Prof. Dr. Davide Scaramuzza.

Schedule

The official schedule should be found here. The Studio Class Room is scheduled on Monday and Thursday, from 9u00 to 13u00. The Studio Class Room will be a combination of lectures, book exercises and assignments. The course will take place in A1.30. A list of Frequent Asked Questions will be maintained. Chapter 1-4 of the book will be introduced by Toto van Inge. Chapter 5-6 will be covered by Arnoud Visser.

Students, who were not able to attend a lecture, can catch up by listing to the recordings of my lectures (in Dutch). Download Lecturnity Player and listen to lecture, synchronized with the sheets.

For the assignments not only a solution is expected, but also a rational. The experiments performed to solve the given problem should be described in a lab report, which will be graded based on the following criteria.

Week 44: Chapter 1 & 2 - Introduction & Locomotion
Solve OpenLoop steering assignment, including RWTH Toolbox Installation Instructions.

Week 44: Chapter 3 - Kinematics

Week 45: Chapter 4.1 Sensors for Mobile Robots

Week 45: Chapter 4.2 Fundamentals of Computer Vision
Solve Assignment 3.

Week 46: Chapter 4.3-4.5 Feature Extraction

Week 46: Chapter 4.6-4.7 Place Recognition

Week 47: Partial Exam

Week 48: Chapter 5.1-5.5 The Challenge of Localization
Solve Localization assignment, Matlab code, color picker and two locally recorded datasets (one for training and one for testing).

Week 48: Chapter 5.6 Probabilistic Map Based Localization

Week 49: Chapter 5.6.8 Kalman Filter Localization
and Kalman Filter Geometric Approach (Slides and Dutch recording)
Only slides 9-45. Note the slightly different notation for the intermediate prediction;
x(k+1|k) by Choset et al.and x(t) by Thrun/Siegwart et al.

Week 49: Chapter 5.8 Simultaneous Localization and Mapping
Solve Assignment 4, with provided Logger, Example log, Matlab files and a locally recorded dataset (dataset and route (displacements of 15cm at the straight lines)).

Week 50: Large-Scale 3D Point Clouds - Introduction, Basic Data Structures, Registration and Reconstruction.

Week 50: Chapter 6 - Planning and Navigation

Week 51: Partial Exam, December 20th, 13:00-15:00, G3.02

Literature

Roland Siegwart, Illah R. Nourbakhsh and Davide Scaramuzza 'Introduction to Autonomous Mobile Robots', 2nd edition, The MIT Press, 2011.

Reading guide

Monday October 28th: page 56
Wednesday October 30th: page 99
Wednesday November 6th: page 194
Friday November 15th: page 264
Thursday November 28th: page 321
Thursday December 5th: page 368
Tuesday December 10th: page 424

Embedding in AI curriculum

This course is supported by the following chapters of 'Artificial Intelligence - A Modern Approach' 3rd edition, by Stuart Russell and Peter Norvig:

Chapter 13: Quantifying Uncertainty
Chapter 14: Probabilistic Reasoning
Chapter 15: Probabilistic Reasoning over Time
Chapter 24: Perception
Chapter 25: Robotics

Evaluation

The course is this year evaluated by the participants with a 6.4:
.

Software toolkits

Chapter 4, section 2.6 (page 186) - Structure from Motion:

FIT3D, by I. Esteban Lopez.
Structure from Motion toolbox by V. Raboud.
Matlab Functions for Multiple View Geometry by A. Zisserman.
Structure from Motion Toolkit by P. Torr.
Matlab code for Non-Rigid Structure from Motion using Factorisation by L. Torresani.

Chapter 4, section 5 (page 234) - Interest Point Detectors:

Intel's OpenCV library, with e.g. Harris, MSER, FAST, SURF, ... Library maintained by Willow Garage.
SUSAN.
FAST.
SIFT by D. Lowe.
reimplementation in C and Matlab of SIFT, MSER, e.a. by A. Vedaldi.
3D object recognition toolkit by Autonomous Systems Lab at ETH Zurich.
GPU implementation SURF by Vision Lab at ETH Zurich.

Chapter 5, section 8 (page 365) - Simultaneous Localization and Mapping algorithms:

OpenSLAM, a list of SLAM software, list maintained by C. Stachniss.
real-time monocular visual SLAM by Davison.
real-time monocular visual SLAM algorithm PTAM by Klein and Murray.
Matlab EKF SLAM simulator.
RawSeeds collections of benchmarked datasets used for SLAM.

Bibliography (page 444) - Referenced webpages:

CVonline: On-line Compendium of Computer Vision, maintained by R.B. Fisher.
The Intel Image Processing Library / Integrated Performance Primitives (Intel IPP)
CMvision source code
Newton Labs website
For probotics
Intel's OpenCV library, maintained by Willow Garage.
Passive walking.
Passive walking, the Corner Ranger.
Computer Vision industry
Camera Calibration Toolbox for Matlab
List of camera calibration software
Omnidirectional camera calibration toolbox from Christopher Mei.
Omnidirectional camera calibration toolbox from Joao Barreto.
Omnidirectional camera calibration toolbox from Davide Scaramuzza.
OpenSLAM, a list of SLAM software, list maintained by C. Stachniss.
Open source software for multi-view structure from motion.
Microsoft Photsynth.
Photo Tourism.
Voodoo Camera Tracker: A tool for integration of virtual and real scenes.
Augmented-reality toolkit (ARToolkit).
Parallel Tracking and Mapping (PTAM).

Last updated January 6, 2014

This web-page and the list of participants to this course is maintained by Arnoud Visser (arnoud@science.uva.nl)
Faculty of Science
University of Amsterdam

Course Autonomous Mobile Robots