|
Instructors: Dr. Annika Peters; Dr. Manfred Eppe
Event type:
Project
Displayed in timetable as:
PJ Cogn.Robot
Hours per week:
6
Credits:
9,0
Language of instruction:
German/English
Min. | Max. participants:
- | 15
Registration group: Anmeldegruppe Projekte
Comments/contents:
The foundations of human cognition are often an inspiration for algorithms in Artificial Intelligence and Robotics. At the same time, the interaction between robots and humans plays an increasingly important role in our society. This leads to interdisciplinary research and applications in the areas of computer vision, signal processing, machine learning with neural networks, as well as the evaluation and design of human-robot interaction.The goal of this project is to investigate these aspects and to learn to know the cognitive principles of human-robot interaction. Herein, the participants of the projects will realize human-robot interaction scenarios. Due to the Corona pandemic, this year’s project will not use a physical robot but simulated robots using the CoppeliaSim software. The scenarios will be inspired by the RoboCup@Home competition, where robotic key capabilities like face detection, language processing, and navigation must be implemented and integrated. Depending on the interest of the students, we will develop a scenario, such as a domestic clean-up task, where students will have the opportunities to take the first steps with neural networks and other AI methods for robotic applications.
Learning objectives:
The project group will be organized into subgroups, according to the individual’s interest with a focus on pair programming. This technique stimulates continuous communication and knowledge transfer between the project partners. As every group is part of a team, the students get to know important project soft skills like time management, teamwork, and project management. After the successful completion of the project, the students gained competence in their project field and learned how to communicate their development steps to the group helping to solve a complex problem.
Didactic concept:
The first part contains an introduction to the topic and a presentation of techniques required throughout the project. Individual open questions like the scenarios to be realized will be discussed with the students. The practical part of the course will consist of a single large project, where students will subdivide into smaller subgroups to tackle individual project parts. A weekly team meeting led by a rotating team leader from the student group will provide time and space for a progress report and solution suggestions for current problems. After the successful integration of the modules, the students will test the scenario and evaluate several interaction aspects like system response time, reliability, etc. The results will be documented in a project report and demonstrated in a final presentation.
Digital Realization of the Project:
The project including all meetings, tutorials, and lectures will exclusively be held via Zoom.
The link to the Zoom room will be announced before the first meeting.
The first meeting will be at 10:00 am 8. April 2021
Literature:
- Raul Rojas, Neural networks: a systematic introduction (Inf-Bib: A ROJ 36077)
- Daniel T. Kaplan, Simon D. Levy, Kenneth A. Lambert: Introduction to scientific computation and programming in Python (Inf-Bib: P KAP 52918)
- Python Einführung online: http://www.scipy-lectures.org/intro/
- CoppeliaSim: https://www.coppeliarobotics.com/
- We will use the latest Edu version: CoppeliaSim Edu
- RoboCup@Home Example: https://www.youtube.com/watch?v=PZmKzngDegk
Additional literature will be made available during the course of the project.
|
