Research

Software Interfaces for Hardware Accelerator in Heterogeneous Systems

The demands on embedded systems are growing rapidly, resulting in the need for ever-increasing processing power, leading to systems with heterogeneous hardware accelerators. Software interfaces must be able to schedule different applications on different accelerators while prioritizing applications on the same accelerator to enable efficient use of hardware resources. At the same time, the interface must be able to adapt to changes in the underlying accelerators, allowing system upgrades without software changes.

Isolated Application Deployment in Automotive Systems

Today's vehicles already use a large amount of third-party hardware and software. Isolating these applications results in easier and faster approval and update cycles, and also improves the user experience by including more third-party applications. However, these applications still require access to vehicle communications. Restricting applications to only the signals they need requires fast, flexible, and secure isolation mechanisms.

Low-overhead, Dynamic, Isolated Serverless Services

Modern embedded systems often use service-oriented architectures to enable communication between different nodes. While splitting functionality into small and separate applications increases reusability while reducing complexity, isolating each application increases communication overhead. Moving applications into serverless runtimes promises to bridge the gap between isolation and performance.

Student Works
Title	Type	Date

HiWi Positions
Title	Starting date

Supervised student work

- MA: AI-Accelerator as a Service for Heterogeneous E/E-Architectures

Publikationen

2023

Policy-Based Task Allocation at Runtime for a Self-Adaptive Edge Computing Infrastructure
Betancourt, V. P.; Kirschner, M.; Kreutzer, M.; Becker, J.
2023. 2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS), 8 S., Institute of Electrical and Electronics Engineers (IEEE). doi:10.1109/ISADS56919.2023.10092022

Modular Hardware/Software Architecture for Edge Units in Highly Flexible Manufacturing Systems
Schade, F.; Kreutzer, M.; Mühlbeier, E.; Gerlitz, E.; Gönnheimer, P.; Fleischer, J.; Becker, J.
2023. Procedia CIRP, 120, 601–606. doi:10.1016/j.procir.2023.09.045

2022

Comparison of Artificial and Spiking Neural Networks for Ambient-Assisted Living
Nitzsche, S.; Pachideh, B.; Neher, M.; Kreutzer, M.; Link, N.; Theurer, L.; Becker, J.
2022. Conference Proceedings: 2022 Smart Systems Integration (SSI), 1–6, Institute of Electrical and Electronics Engineers (IEEE). doi:10.1109/SSI56489.2022.9901412

A hardware/software co-design approach to prototype 6G mobile applications inside the GNU Radio SDR Ecosystem using FPGA hardware accelerators
Karle, C. M.; Kreutzer, M.; Pfau, J.; Becker, J.
2022. HEART2022: International Symposium on Highly-Efficient Accelerators and Reconfigurable Technologies, Tsukuba Japan, June 9 - 10, 2022, 33–41, Association for Computing Machinery (ACM). doi:10.1145/3535044.3535049