Sensor Fusion – Architectures and Algorithms

11B1805

Bachelor

German, English

5.0

irregular

1 semester

Similar to our senses, sensors in a robot or vehicle perceive the environment. Applications such as smart homes, automated driving and human-robot collaborations can be realized with a sophisticated sensor set that uses intelligent real-time algorithms to sense the environment of a machine. This course aims to learn about, evaluate, and implement sensor fusion's fundamental architectures and algorithms. When developing the architecture, the dimensions of functional safety, cyber security, sensor costs, software costs, and real-time requirements must be taken into account. With the architecture as a basis, algorithms can be specifically selected and implemented that combine all sensor information, for example, in a map of the environment, where all objects can be reliably localized and recognized. In the practical course accompanying the lecture, the practical application is carried out, in which, for example, indoor navigation is realized using an acceleration sensor, and a GPS receiver in a smartphone or vehicle is localized using a camera and LiDAR objects.

• architecture design for software and hardware systems
• sensor synchronization, event-based data processing
• virtual sensors or how several sensors become a "super" sensor
• functional safety and cyber security vs. software and hardware costs
• classic methods for sensor fusion, physical correlations in defined algorithms
• AI-based methods for sensor fusion, dataset-based learning of correlations

The total workload for the module is 150 hours (see also "ECTS credit points and grading").

• Oral presentation, with written elaboration or
• Project Report, written

• Field work / Experimental work

Presentation: 30 minutes; accompanying essay: 8 pages

Project report (written): 10 pages; associated presentation: 15 minutes

Experimental work: 6 worksheets

Object-oriented programming (OOP), linear algebra, strong interest in algorithms

Students acquire in-depth knowledge of sensor fusion concepts and techniques as well as an understanding of the physical relationships in classical fusion algorithms and working with AI-based methods.

Students are able to use practical sensors and fusion technologies in various applications such as smart home systems or autonomous vehicles. By analyzing and implementing sensor architectures that take into account safety aspects as well as cost and time efficiency, students learn to solve complex problems and apply their knowledge in real projects.

This module encourages scientific innovation by providing the ability to design innovative sensor architectures and develop creative solutions for the fusion of sensor data. By exploring classical and AI-based methods and considering safety and cost aspects, the module encourages students to push the boundaries of current knowledge and develop novel applications in different application areas.

Kommunikation und Kooperation werden gefördert, indem die Studierenden dazu angehalten werden, die Interaktion zwischen verschiedenen Sensoren und Algorithmen zu optimieren. Durch die Entwicklung und Anwendung von Sensorfusionen in Gruppenarbeit lernen die Studierenden, wie sie effektiv zusammenarbeiten können, um komplexe Systeme unter Berücksichtigung von Sicherheitsaspekten und Echtzeitanforderungen zu realisieren.

• Schöning, Julius

• Schöning, Julius
• Stiene, Stefan