A Werner von Siemens Award goes to FAV. Jakub Matoušek placed third for his dissertation research

Jakub Matoušek, a graduate of the Faculty of Applied Sciences at UWB, made it into the top dissertations on the topic of Industry 4.0 in the Werner von Siemens Award. He succeeded with a proposal for algorithms for more accurate processing of sensor data, taking third place in this category.

Reliable aircraft navigation, the ability to track multiple objects in autonomous transportation, or in medicine and biology. In all these fields, experts use sensors that often do not provide accurate or reliable data . This was the challenge for Jakub Matoušek, who based his dissertation at the Faculty of Applied Sciences at the University of West Bohemia in Pilsen (FAV) on the effort to improve the accuracy of this essential data. His work has now caught the attention of the expert jury of the Werner von Siemens Awards.

A graduate of the Cybernetics program addressed the issue of inaccurate sensor data using advanced algorithms. From the imperfect data, he estimated the magnitude of the measured quantity and even the potential margin of error in that estimate. “I used grid-based algorithms—you can imagine it, for example, as if we were trying to find pieces on a chessboard and assigning each square a probability that a piece is located there based on observations made, say, by a low-quality camera,” explained Jakub Matoušek.

In today’s complex geopolitical situation, where GPS signals are being jammed, this method can, for example, reliably navigate aircraft without relying on external signals. According to Jakub, this is often not possible using existing methods due to their high computational demands or insufficient accuracy and robustness. Jakub has simplified the calculation—which was previously feasible only on supercomputers—so that it will be possible to perform it on a standard laptop in the future.

The topic of improving existing algorithms is nothing new to Jakub. He has previously been involved in developing algorithms to enhance operational safety and reliability. He is currently working as a researcher at the University of Texas at Austin. At the end of the year, he will return to FAV, where he will continue his research. Another challenge he is currently working on, after completing his dissertation, is running these calculations on microcontrollers, such as the Raspberry Pi. “I want to achieve this using effective tensor decompositions—essentially data puzzles that allow us to describe and process a system’s behavior across multiple dimensions simultaneously, rather than just in 2D as with vectors or matrices,” he explained.

The Werner von Siemens Award has been organized for 28 years by Siemens Czech Republic in partnership with leading representatives of universities and the Czech Academy of Sciences, who also serve as guarantors for the individual categories and participate in the evaluation of the best works . This year, 56 experts and representatives of the academic community served on the independent juries. In terms of scope, financial rewards, and history, the Werner von Siemens Award is one of the most significant independent initiatives of its kind in the Czech Republic.