Multimedia Computing and Computer Vision Lab

Login  

Home

     

Courses

     

People

     

Research

     

Publications

     

Student Theses

     

Source Code / Datasets

     

Contact

     

WS 14/15: Probabilistic Robotics

From Multimedia Computing Lab - University of Augsburg


Registration is now open at Digicampus



Overview

Instructors: Prof. Dr. Rainer Lienhart, Dan Zecha
Time Lecture: Thursday, 10:00 - 11:30, Room: 1057 N, starts Oct 9th
Time Exercise: Friday, 14:00 - 15:30, Room: 1020 N.
The first assignment will be handed out on October 17st; the exercise starts the following week, on October 24th, 15:45 - 17:15.
Credits: 2 + 2 SWS, Schein: yes, LP 5
Exam: 12.02.2015, 10:00-11.30 a.m., Room 2045N
Multimedia Teilbereiche: Multimedia Methoden, Multimedia Anwendungen, Systemnahe Grundlagen von Multimedia

News

  • The exercise lesson on Jan 9th will not take place. Instead, assignment #10 will be discussed on Jan 16th.
  • The exercise has been moved to Friday, 14:00 - 15:30, room 1020N
  • Registration is now open at Digicampus

Synopsis

In the course of this lecture students will learn how robots can estimate their state (e.g. their pose) in a probabilistic fashion, i.e. in the face of uncertainty.

The main focus of this lecture is on the Bayes Filter algorithm which enables robots to estimate their new state after executing a control and to incorporate sensor measurements to update their belief. Various flavors of the Bayes Filter such as the Kalman Filter and the Particle Filter will be discussed in detail .

Furthermore, students will get to know different ways to model robot motion and measuerments of various types of sensors.

The final chapters of the lecture will be on approaches to robot localization, i.e. the problem of the robot having to determine its position on a given map of the environment. Also, the localization problem will be discussed for situations when the robot has to generate a map itself by occupancy grid mapping or simultaneous localization and mapping (SLAM) algorithms.

Exercise and Exam

  • There will be an exercise sheet every week
  • Solutions will be discussed during exercises on Fridays (no handing-in or revision of written solutions)
  • Students are encouraged to present their solutions
  • No admission requirements for the exam

Important Comments

  • Vorlesung wird auf Deutsch gehalten (trotz der englischsprachigen Folien und Literatur)
  • All reading notes are relevant for the exams independent on how thoroughly they have been discussed during the lecture. Thus read and study them carefully.

Literature

Sebastian Thrun, Wolfram Burgard, Dieter Fox. Probabilistic Robotics. MIT Press. (http://www.probabilistic-robotics.org/)
Mandatory to read chapters 1 - 8