FMCW lidar measures velocity directly, and that is its real argument
Time-of-flight lidar infers motion by comparing frames. Coherent detection gets Doppler velocity per point, per measurement.
Most deployed lidar measures the round-trip time of a short pulse. Frequency-modulated continuous wave lidar instead emits a continuously chirped beam and mixes the return with a local copy, deriving range from the beat frequency.
The architectures differ in ways that matter beyond the measurement principle.
What coherent detection provides
Velocity per point. The Doppler shift of the return appears directly in the beat signal, so every measurement yields range and radial velocity simultaneously. A time-of-flight system must compare successive frames and associate points between them to estimate motion, which is inference rather than measurement.
For automotive perception, knowing immediately that a point is approaching at speed, without waiting for a second frame or solving an association problem, is a substantial advantage.
Interference immunity. Coherent detection responds only to light correlated with the local oscillator. Another vehicle's lidar, sunlight and retroreflector glare do not produce a valid beat signal. As lidar deployment density rises, mutual interference is a genuine concern for pulsed systems.
Sensitivity. Coherent gain allows detection of very weak returns, which extends range at eye-safe power levels.
What it costs
Source requirements. The laser must be frequency-tunable with high linearity and narrow linewidth over the chirp. That is a harder specification than a pulsed source.
Signal processing. Extracting range and velocity requires transforming the beat signal, which is continuous computation rather than time-stamping an edge.
Maturity and cost. Pulsed lidar has volume, an established supply chain and falling prices. FMCW is behind on all three.
Where photonic integration enters
FMCW's architecture — a tunable laser, splitters, a coherent receiver — maps well onto silicon photonics, and integration is the route to the cost and volume that automotive requires.
That is the strategic bet: FMCW is technically better suited to the application and currently more expensive, and integration is expected to close the gap.
Whether it does is an open question, and it is the same question facing silicon photonics generally — the chip is not the hard part, the packaging is.
This article was produced by the LasersNews AI desk and reviewed by our editors.
Related reading
Integrated photonics narrows the gap between laboratory result and manufacturable part
Foundry access and standardised process design kits are doing for photonics roughly what they did for electro…
Metasurface optics are running into the manufacturing question
Flat lenses built from sub-wavelength structures work. Producing them at volume, at yield, with the tolerance…
Photonics found its foundry model, and the effect resembles electronics in the 1980s
Shared process design kits and multi-project wafer runs let a small team design a photonic chip without ownin…