Learning Robots

Year:
2nd year
Semester:
S1
Programme main editor:
(I2CAT)
Onsite in:
UBB
Remote:
ECTS range:
3-7 ECTS

Professors

img
Professors
Birte Glimm
UULM
img
Jeremy Van Gorp
CNAM
img
Professors
Mathieu Moze
CNAM

Prerequisites:

Programming in Python or C++, basics of ROS and simulation environments, understanding of machine learning concepts (e.g. Neural Nets, Reinforcement Learning, Computer Vision) may be helpful.

Pedagogical objectives:

Students are able to conduct a project-oriented, scientific work in the area of robotics. They have the ability to define an innovative project topic, acquire skills and technologies required for the project, find related works and base their own work on them where appropriate. The students are further able to work independently in a team and to apply state-of-the-art methods to develop concepts and solutions for the project topic. They can document the results in the form of a scientific report and present their findings in the form of a presentation. They are able to acquaint themselves with a new topic and to conduct a scientific project.

Evaluation modalities:

Project presentation and report.

Description:

At the beginning of the project, the range of contents that should be covered in the project will be discussed in detail and the relevant methods will be identified. This is accompanied by a literature search, reading, and discussion phase. In this phase relevant mathematical and computational methods will be selected and discussed. In the second phase, the team starts to implement the experimental design on a simulated robot. Regular team meetings and supervisory consultations lead to an iterated improvement of the software. In the third phase, the thoroughly tested software will be transferred to the robot and tested in a real or simulated environment. The results are written up in a final project report and are presented in a final project presentation.

  • Definition of concrete project idea.
  • Project plan incl. systematic literature review.
  • Evaluation of suitable technologies.
  • Self-learning of the required technical foundations.
  • Architecture design.
  • Implementation.
  • Integration.
  • Test.
  • Deployment.

Required teaching material

Literature: • Thrun, Burgard, Fox “Probabilistic Robotics” • Bishop “Pattern recognition and machine learning” • Russell & Norvig “Artificial intelligence. A modern approach” • Goebel “ROS by Example INDIGO – Volume 1”

Teaching volume:
lessons:
Exercices:
Supervised lab:
15 hours
Project:
15 hours

Devices:

  • Laboratory-Based Course Structure
  • Open-Source Software Requirements