Automotive
Trusted AI for radar systems
Radar is becoming a richer perception modality, but safety-critical use cases still demand explainability, robustness, and hardware-efficient inference from the full stack.
Published on 25/09/2025
Making perception systems dependable
The challenge is no longer just detecting objects. Engineers need radar stacks that remain stable across weather conditions, edge cases, and embedded compute limits while still being interpretable enough for validation and safety arguments.
This fits imec’s system view well because sensing hardware, signal processing, and AI acceleration all need to be shaped together instead of handed off sequentially.