Automotive

Velocity, Not Only Position

A conventional lidar outputs the position of each point. It estimates velocity from the difference between consecutive frames. This estimate becomes less stable as targets get farther away and points get sparser. But the brake decision depends on velocity. Is the point ahead a stationary guardrail, or a car that slows down?

FMCW gives radial velocity in the same measurement as range. Radial velocity is the rate at which a target moves toward or away from the sensor along the line of sight. A time-to-collision estimate needs this value most.

Where Radial Velocity Helps, and Where It Does Not

Doppler gives only the velocity component along the line of sight. A target that crosses the path has a small radial component. For this motion, the system still needs information from more than one frame. The main value of radial velocity is in longitudinal motion:

• A car ahead brakes: the change in closing rate is visible in the same frame.
• Stationary obstacles: the distinction from slow vehicles does not depend on point count. This also applies to small targets at long range.
• Dense traffic: each point carries its own velocity, so target separation does not rely on clustering alone.

Coherent Detection Resists Interference

A coherent receiver responds only to echoes that match the frequency of its local oscillator. Direct sunlight, oncoming headlights and the lidars of other vehicles do not enter the mixing path. As more vehicles carry lidar, mutual interference changes from an edge case to a common problem.

Work with us