Level 3 highway autonomy is transitioning from limited pilots to series production, shifting the primary technical challenge from vehicle control to perception robustness at speed. At highway velocities, eyes-off operation is fundamentally limited by long-range sensing performance, particularly angular resolution, signal-to-interference ratio, and dynamic range under dense clutter conditions.
This talk analyzes the perception requirements imposed by L3 highway use cases and explains why high-resolution radar is emerging as a key enabler. At speeds up to 130 km/h, perception stacks must reliably detect and classify objects beyond 200 meters, resolve closely spaced targets with angular separations below 1 degree, and discriminate weak reflections adjacent to strong scatterers such as debris near guardrails, stalled vehicles near trucks, or vulnerable road users in mixed traffic. Conventional automotive radars, typically constrained to low channel counts and limited dynamic range, struggle with target separability, ghost suppression, and saturation effects in these scenarios.
We present how next-generation high-resolution radar architectures, combining large virtual apertures, fine Doppler resolution, and enhanced dynamic range, address these limitations. The session will discuss concrete performance metrics, including range and angular resolution, detection probability versus false alarm rate, and robustness under adverse weather, and show how these translate into more reliable free-space estimation, stationary object handling, and trajectory prediction for L3 highway autonomy.
By engineers, for engineers: A technically grounded guide to the rapidly evolving in-cabin technology industry and companies.
