Using rodents as a model, mobile robots can find their way around their environment more quickly and easily. In his doctoral project, Dr. Muhammad Haris has further developed a machine learning method that enables robots to find their way around as accurately as possible, even outside enclosed spaces. The former doctoral student at Frankfurt University of Applied Sciences (Frankfurt UAS) has now been awarded the university's doctoral prize for his dissertation from 2023. The award, which is endowed with 3,000 euros, was presented for the first time on 8 May 2024 as part of the "Doctorate in Focus" event.
The newly created prize from Frankfurt UAS recognizes outstanding academic work that was successfully defended in the previous year and was carried out either at one of the university's doctoral centers or in cooperation with other universities. It is financed by the Frankfurt UAS Friends' Association. Prof. Dr. Susanne Rägle, Vice President for Research, Transfer and Internationalization at Frankfurt UAS, emphasizes: "With this award, we are taking into account the increasing importance of the topic of doctoral studies at Frankfurt UAS. In this way, the Friends of Frankfurt UAS and we as a university want to recognize and make visible the research achievements of young academics."
The award winner was selected by a four-member jury convened by the university. The evaluation criteria were the innovativeness and social significance of the work as well as its topicality and fit with the main research areas of Frankfurt UAS. The dissertation "Visual Localization and Mapping in Seasonally Changing Outdoor Environments", submitted by Dr. Muhammad Haris at the PhD Center Applied Informatics (PZAI), scored exceptionally well in all categories.
Robust models inspired by rodents save money and time
The scientific work was supervised by Prof. Dr. Ute Bauer-Wersing. Dr. Ute Bauer-Wersing, Professor of Machine Learning and Data-Driven Modelling, as part of the third-party funded project "Unsupervised Learning of Hierarchical Features for Visual Self-Localization and Navigation in Seasonally Changing Outdoor Environments". Mobile robots such as lawn mowing robots, but also self-driving cars, have to constantly reorient themselves in space and assess their surroundings using cameras, among other things. Where am I? What is around me? Especially outdoors, this is made even more difficult by changing weather conditions and sun reflections. At the same time, many of the existing calculation models are very computationally intensive and therefore require expensive hardware, meaning they are not suitable for simple lawn mowing robots, for example. For the new approach that Haris developed further in his doctoral project, he received help from nature. The machine learning process is based on how rodents represent what they see in their brains. "This biologically inspired approach is robust and not so computationally intensive, so it is also suitable for mobile assistance robots with inexpensive hardware. It makes it possible to grasp the environment more quickly and locate oneself in it," says Haris. In addition, a robot is able to learn unchanging features of an environment and thus navigate safely under changing conditions.
"Computer vision" has long been a topic
Haris has been involved in the research field of computer vision since his Bachelor's degree in Electrical Engineering at the University of Islamabad. After completing a Master's degree in Information Technology at Frankfurt UAS, he completed his doctorate there from 2018 to 2023. The practical aspect of the doctoral project also played a role in his decision. The research project was funded by the Honda Research Institute Europe and carried out in cooperation with the industrial research institute. "This allowed me to exchange ideas directly with scientists from industry," says Haris.
Contact PhD Award: Frankfurt University of Applied Sciences, Department of Research, Innovation and Transfer, Head of Department, Dr. Julia Lefèvre, Phone: +49 69 1533-2162, Email: julia.lefevre@fit.fra-uas.de as well as Doctoral Support Officer, Dr. Dominik Wolf, phone: +49 69 1533-2636, e-mail: wolf.d@fit.fra-uas.de.
Scientific contact: Frankfurt University of Applied Sciences, Faculty 2: Computer Science and Engineering, PhD Center Applied Computer Science, Prof. Dr. Ute Bauer-Wersing, phone: +49 69 1533-2793, e-mail: ubauer@fb2.fra-uas.de
Further information on doctoral studies at Frankfurt UAS:
www.frankfurt-university.de/promotionsfoerderung/.
For more information on Dr. Muhammad Haris' doctoral topic:
