Master Artificial Intelligence

Course Probabilistic Robotics

This is the information of Fall 2018

Note that the course will no longer be given in 2019. This course was given for the first time on Master level in (Fall 2017).

Description

The description is available in the course catalogue of the UvA The course is a constrained choice in the Master Artificial Intelligence Curriculum. More details about the organization of the course can be found in the Course Manual.

Probabilistic robotics is a subfield of robotics concerned with the perception and control part. It relies on statistical techniques for representing information and making decisions. By doing so, it accommodates the uncertainty that arises in most contemporary robotics applications.

This course is based on the book 'Probabilistic Robotics', from Sebastian Thrun, Wolfram Burgard and Dieter Fox. The book concentrates on the algorithms, and only offers a limited number of exercises. Their suggestion is to accompany the book with a number of practical, hands-on assignments for each chapter. The assignments of this course are designed to understand the basic problems concerning mobile robotics.

Schedule

The course will take place in a studio classroom setting. Students, who were not able to attend a lecture, can catch up by listing to the recordings of Prof. Burghard's lectures (English).

The assignments were based on the Octave or Matlab environment, ported to Jypiter notebooks.

Week 36: Chapter 1 - Introduction & Robot Paradigms
Recording Robot Paradigms

Week 36: Chapter 2 State Estimation & Chapter 3.2 - Gaussian Filters - Kalman Filter
Recordings State Estimation part 1 and part 2 & discussed Burghard's lecture

Week 37: Chapter 3.3 - 3.5 Extended Kalman Filters & Geometric Approach
Recording Extended Kalman Filter

Week 37: Chapter 4 - Nonparametric Filters: Discrete and Particle Filters
Recordings Discrete Filters and Particle Filters (part 1 & part 2)

Week 38: Chapter 5 - Wheeled Locomotion & Robot Motion Models
    Recordings Wheeled Locomotion (part 1, part 2, part 3, part 4) and Motion Models (part 1, part 2, part 3).

Week 38: Chapter 6 - Sensors & Robot Perception Models
    Recordings Sensors (part 1) and Sensor Models (part 1, part 2).
    Stanley's RaceDay movie and Stanley's Tech lecture recording (Dutch)

Week 39: Chapter 9 - Mapping with known poses
Recordings (part 1, part 2, part 3, part 4).

Week 39: Chapter 10 - SLAM .
    montemerlo-fastslam-aaai video, corresponding to AAAI'03 paper
    Recordings (part 1, part 2).
    animation of raw data of Intel Research Lab, animation of CMU's Wean Hall

Week 40: Chapter 13 - The FastSLAM Algorithm
Recordings (part 1, part 2).

Week 41: Chapter 11 - GraphSLAM and Chapter 12 - SEIF SLAM.
video GroundHog, video mine map, video GraphSLAM with Segbot

FastSLAM with Known Data Association Assignment.

Week 41: Chapter 17 - Exploration and 3D Mapping
Recordings (exploration and 3D mapping).
video CMU 1999 multi-robot exploration, video exploration with limited communication

Week 42: Final - Sebastian Thrun's TED talk.

Week 43: Exam ('open book') of Chapter 1-13 & 17 of the book, Thursday October 26, 9:00-12:00, A1.02.

Week 2: Reexam ('open book'), Thursday January 11, 9:00-12:00, room A1.06

Literature

Sebastian Thrun, Wolfram Burgard and Dieter Fox, Probabilistic Robotics, The MIT Press, 2005.

Reading guide

Week 36: until section 3.3 - page 54
Week 37: until chapter 4 - page 116
Week 38: until chapter 6 - page 187
Week 39: until chapter 10 - page 335
Week 40: chapter 13 - page 437-483
Week 41: chapter 11,12,17 - page 336-436 and 569-605

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 25: Robotics

Evaluation

The course was evaluated in 2018 by the participants with a 7.1:
.

Course Probabilistic Robotics