The engineer conducts research on signal processing and machine learning and wants to further develop methods for the Scout rescue robot.
Rescue robots can be a great help in disasters. For example, to locate missing persons in destroyed buildings after an earthquake, fire or tsunami. Christoph Löser, research associate in the Robust Data Science Group (RDS) and the Institute for Artificial Intelligent Systems in Medicine (KIS*MED) at TU Darmstadt, wants to optimize methods for the Scout rescue robot developed by emergenCITY so that missing people in need of help can be found more quickly.
He is supporting the LOEWE center emergenCITY in the Cyber-Physical Systems program area since March. His research topics are statistical signal processing and machine learning in the context of detection, parameter estimation and variable selection. His focus is on researching, developing and analyzing statistical error-controlling methods for high-dimensional problems in biomedical applications.
“Specifically, I will be working on applying the T-Rex methodology developed by Jasin Machkour and the Robust Data Science Group, i.e. using it for radar-based localization of people and their vital parameter estimation,” the scientist says. Subsequently, the amount of sensor technology will be used to improve detection by using not only radar, but also other sensor technology with the help of sensor fusion.
The aim is to enable the Scout rescue robot to locate people safely and quickly and to measure their state of health using vital parameters such as breathing rate and pulse. This would allow injured and sick people, for example, to be better monitored and considered first during the rescue.
Christoph Löser has already written his master’s thesis on “Efficient Graph Learning for High-Dimensional Data” in the RDS Group. Topics that he is equally enthusiastic about are non-invasive methods for vital parameter estimation, initially radar-based.
