RobSwarm: Deep reinforcement multi robot mapping
- Problem: A group of robots, called a swarm, is placed in an unknown environment and is supposed to explore it independently. The goal of the exploration is the creation of a common map. Implementation - Equipping six Kobuki robots with appropriate sensor technology, a large battery, a router and the Jetson board - Setup of the Jetson-Boards with self-made ROS2 nodes and the set up mesh network - Writing of launch files for the common start of all functions - Reinforcement learning is used to train an AI that controls the swarm by selecting points for the robots to approach and navigating to them and navigating them there. - Setting up a responsive website using Angular and the Bootstrap Framework.
Author: | Julien Meine, Niklas Digakis, Niklas Dietz, Karol Rösner, Sarah Brüssing |
---|---|
URN: | urn:nbn:de:hbz:1010-opus4-43099 |
Document Type: | video |
Language: | German |
Date of Publication (online): | 2023/02/02 |
Date of first Publication: | 2023/02/02 |
Publishing Institution: | Westfälische Hochschule Gelsenkirchen Bocholt Recklinghausen |
Release Date: | 2024/01/26 |
Departments / faculties: | Fachbereiche / Informatik und Kommunikation |
Licence (German): | Creative Commons - Namensnennung |