Scientific publications

The processing of personal data is becoming a key business factor, especially for high-tech system industries such as automotive and healthcare service providers. To protect such data, the European Union (EU) has introduced the General Data Protection Regulation (GDPR), with the aim to standardize and strengthen data protection policies across EU countries.

The Halide DSL and compiler have enabled high-performance code generation for image processing pipelines targeting heterogeneous architectures through the separation of algorithmic description and optimization schedule. However, automatic schedule generation is currently only possible for multi-core CPU architectures.

Vehicles, as a prime example of high-tech systems, get increasingly connected and data-centric with the need to process personally identifiable information. Often, companies that develop such systems act as integrators and need to comply to adequate data protection requirements. For instance, GDPR requires securing personal data.

Monitoring and managing the health of technical systems with advanced diagnosis and prognosis benefits from fleet analytics: insights on the degradation of other but similar systems help, e.g., to forecast actual issues for predictive maintenance as does detecting and correcting anomalies in usage profiles helps to prevent undue wear and tear.

Tensor Cores (TCUs) are specialized units first introduced by NVIDIA in the Volta microarchitecture in order to accelerate matrix multiplications for deep learning and linear algebra workloads. While these units have proved to be capable of providing significant speedups for specific applications, their programmability remains difficult for the average user.

Automotive software implements different functionalities as multiple control applications sharing common platform resources. Although such applications are often developed independently, the control performance of the resulting system depends on how these applications are integrated. A key integration challenge is to efficiently schedule these applications on shared resources with minimal control performance degradation.

Autonomous vehicles use cyber-physical systems to provide comfort and safety to passengers. Design of safety mechanisms for such systems is hindered by the growing quantity and complexity of SoCs (System-on-a-Chip) and software stacks required for autonomous operation. Our study tackles two challenges: (1) fault handling in an autonomous driving system distributed across multiple processing cores and SoCs, and (2) isolation of multiple software modules consolidated in one SoC.

Over time, software tends to grow more complex, hampering understandability and further development. To reduce accidental complexity, model-based rejuvenation techniques have been proposed. These techniques combine reverse engineering (extracting models) with forward engineering (generating code). Unfortunately, model extraction can be error-prone, and validation can often only be performed at a late stage by testing the generated code.