Abstract
Automotive radar depends on precisely timed frequency ramps for reliable range andvelocity estimation. This thesis investigates whether automated code generation can
produce sequencer code for the new RISC-V-based Sequencer (RVS). In this platform,
a RISC-V processor executes software instructions to control ramp generation instead
of relying on a fixed hardware-only decoder. The transition increases flexibility, but it
changes the way scenarios are encoded. Ramp scenarios that were previously handled by a
proprietary design tool workflow must now be translated into executable C++ code for the
software-defined sequencer. Manual implementation of these artifacts is labor-intensive
and error-prone, especially for domain experts focused on radar design.
To address this problem, the thesis presents and implements a generation workflow
that extends an existing proprietary design tool with a dedicated output path for RVS.
The approach extracts relevant scenario information from the proprietary design tool,
maps it to an intermediate representation, and renders deterministic source files using
template-based code generation. The architecture keeps the familiar scenario-authoring
process unchanged and adds an automated backend for the new RVS target.
Evaluation is performed with representative scenarios of increasing complexity, includ
ing simple ramps, interleaved ramps, stepped parameter variations and combined feature
cases. Generated artifacts are integrated into the RVS Runtime Library, compiled, and
validated in a virtual prototype by comparing simulated and expected waveforms. Results
show that the prototype reliably preserves intended timing and frequency behavior while
producing readable and structured code.
The thesis also identifies key limitations. Artifact size and management effort increase
with scenario complexity, and the current workflow is one-directional because no reverse
transformation from generated C++ back to design-level scenarios is available. Overall,
the work demonstrates technical feasibility and practical value. It also provides a solid
foundation for future development toward production-ready integration of automated
ramp-scenario generation.
| Date of Award | 2026 |
|---|---|
| Original language | English (American) |
| Awarding Institution |
|
| Supervisor | Franz Leopold Wiesinger (Supervisor) & Michael Atzmüller (Supervisor) |
Studyprogram
- Embedded Systems Design
Cite this
- Standard