SwarmLab Projects
Lecturer:
- Christoph Steup
- Sebastian Mai
- Michael Preuß
- Sanaz Mostaghim
Language:
The course will be held in English or German depending on the participants preferred language.
Participants:
All students of bachelor and master curriculums of the faculty are eligable to visit the course. The course can be taken as Digital Engineering Project, Inter-Disciplinary Project or Team-Project. The actual type of course depends on the needs of the students and the available Topics. It is mandatory for a participant to have background knowledge in at least one of the following topics:
- Robotics
- Programming in C/C++ or Python
- Theorie and Algorithms of Swarm Intellligence
- Communications and Networks
- Development of Embedded Applications in C
- Robot Operating System (ROS)
- Linux Server Admisitration
First Meeting:
The first meeting will be used to organize each project. Because of the ongoing Corona-Pandemic we wil, not do a general first meeting, but we will collect all interested students and afterwards contact them directly.
Organization:
The course will be taken in groups of 3-5 Students per topic. The students and the groups will be chosen by us depending on your background. The individual topics are not fully fixed, extensions and modifications are possible depending on the skills and interest of participating students. This will be discussed in the first meeting. The result of each project is a working demonstration with commented source code and a written documentation indicating the general concept and a How to to start the demo.
Available Topics:
Teamproject: Automatization of Inventory Management for the FIN
The university administration requires every piece of equipment over 100 Euro worth to be registered and assigned to a structural unit. Currently, this process is done manually by assigning numeric IDs to each piece of equipment and handling those IDs using Excel-Sheets. The purpose of this team project is to automize this process by leveraging the available bar code reader and the attached bar codes of the tags sticked on the pieces. In this project a software solution shall be devloped, which enables adminitsrative staff to scan the tags and output all relevant information as well as updating the last known location. The system should be composed of a database backend together with a pc or smartphone front end. The backend slould provide a well-documented API to integrate additional future services.
There is currently a team of 5 students working on it. It is possible to continue after them approx. mid of SS21.
Contact: Michael.preuss@ovgu.de, steup@ovgu.de
Teamproject / DE-Project: Multi-Source Certainty Grid Cooperative Localization
The general approach of this project is to use certainty grids for teammate localisation. The certainty grid provides a simple way to combine information from multiple sources into a single location estimate. The basic approach of certainty grids is to provide a discretised representation of the spacial probability distribution of the position of the other robots. The underlying grid is aligned to the map the robot uses to navigate (i.e. an occupancy grid created with SLAM). For each cell the probability that the target/teammate is present is computed given the sensor information so each layer is clearly defined by a set of probabilities. The combination of all the layers uses the Bayes Rule for each cell in the grid. The goal of this project is to use and enhance the existing certainty localization grid system on real robots and evaluate the results.
Contact: steup@ovgu.de
Teamproject: Development and Integration of an Evolutionary Algorithm-based Optimizier for Taskordering for the @Work League
The RoboCup Team "robOTTO" of the OvGU competes in the @Work-League of the RoboCup. In this league objects need to be transported from workstation of different shape and complexitz to othe rworkstations. The planning of the transportation tasks is rather complicated, because it resembled the Travelling-Salesman Problem and the Napsack-Problem. The team currently uses a Breadth-First-Search Algorithm to find the optimal order of the tasks, However, this approach generated exponential growing runtimes depending on the number of the tasks. For small tasks this overhead is acceptable, but for larger tasks the runtime tends to outgrow the time for the competion itself. The goal of this project is to design and implement an optimizatiion algorithm tailored to the specific needs of the team based on Evolutionary Algorithms. One important aspect for this algorithm is the incorporation of risk-awareness to generate the maximum amount of points with highest probability.
Contact: steup@ovgu.de
Teamproject / DE-Project: Evaluation of Swarm Behaviours of Quadcopters in a Simulated Environment
Swarms of Quadcopters are interesting research objects, as the provide new means to solve existing problems like area surveillance or search and rescue tasks, but they also open up an area of completely new applications. This project aims to implement various swarm behaviour mechanism in physically correct simulated quad copters to evaluate the resulting (emerging) behaviour. The swarm behaviours maz be adaptive (learned) or designed behaviours and are evaluated in multiple different simulation scenarios.
Project goals
- Implementation of various swarm behaviors of different categories
- Evaluation of these behaviors regarding swarm stability, adaptability and entropy
Subtasks
- Attraction-Repulsion
- Context Maps
- Learning-Classifier System-based Local Interaction Rules
- Definition of Simulation Scenarios
- Evaluation
Contact: steup@ovgu.de