Florian GunzerTo explore new terrain and perform tasks autonomously, be it in the air, on the ground or under water, robotic vehicles need to know where they are and what their surroundings are like. As part of his research for his doctorate, Christian Brommer, a member of the research group ‘Control of Networked Systems’ at the University of Klagenfurt, has developed a modular and robust framework model that is designed to maximise the autonomy of robotic vehicles exploring unknown terrain.

‘Until now, most technology development has been tailored to very specific problems. These technologies work very well in their particular scenarios, but in some cases they are over-engineered and require refinement by experts,’ Christian Brommer explains. The overall aim of the contributions to his doctoral thesis was to develop a universal approach. Christian Brommer continues: “In an ideal scenario, a localisation system should allow for the autonomous integration of information, with little or no human interaction. This capability would be particularly important for long-term autonomous exploration missions, given that it is not possible to know in advance which sensors will be available at the location and which measurements can be taken and in what quality.”
To implement such a system, what is needed is a modular and robust multi-sensor fusion framework. Christian Brommer’s aim is to enable “true modularity”. The aim for the future is to make it possible to add or remove individual sensor state elements in terms of position, rotation and speed in the form of isolated blocks of information. The algorithm should then be able to determine what type of sensor is involved and how it can contribute to the modular filter. In one of the subprojects, Christian Brommer was able to demonstrate that it is even possible to reduce the computational complexity required for localisation without negatively affecting the accuracy of the estimates. This is particularly important for smaller robots with a low payload and longer operating times.
Christian Brommer successfully completed his doctoral studies – during which he conducted research in the ‘Control of Networked Systems’ research group – in late summer 2024. The framework he co-developed with colleagues has also successfully undergone all validations, as Brommer recaps: “Thanks to its reliability, real-world applicability and real-time capability, the framework has proved a success in numerous international projects and for various types of vehicle, ranging from underwater vehicles and ground robots to unmanned small helicopters.”
In addition to securing a patent, the research group’s work on the framework earned them the prestigious 2023 Carinthia Research and Innovation Award, as well as a nomination for the 2024 Houska Prize.
Christian Brommer with his family after his defensio | Foto: KK
 
Der Beitrag Location-aware robotic vehicles: a universal approach paves the way for a wider range of applications erschien zuerst auf University of Klagenfurt.