Sub-system
Cameras see detail and colour and estimate distance poorly. Radar measures distance and speed directly through weather and sees shape poorly. Lidar measures precise geometry and struggles in fog and heavy rain. Sensor fusion exists because their weaknesses do not overlap.
In one sentence
Perception sensors are the devices an automated driving system uses to observe its environment — cameras measuring light, radar measuring radio reflections, and lidar measuring the time of flight of laser pulses.
A camera is a dense, cheap, passive sensor. It reads signs, lane markings and traffic lights — information encoded visually and available no other way — but must infer distance rather than measure it, and it fails in glare, darkness and heavy weather.
Radar measures range and relative velocity directly using the Doppler shift, penetrates rain and fog, and is inexpensive and mature. Its angular resolution is poor, so it knows something is ahead more precisely than it knows what. Higher-resolution imaging radar narrows that gap.
Lidar builds a precise three-dimensional point cloud by timing laser pulses. It measures geometry directly at high accuracy and works in darkness. It is affected by fog, rain and spray, and it has historically been the expensive sensor — the cost has fallen substantially with solid-state designs.
Split by what the sensor physically measures, because light, radio and laser time-of-flight are built by entirely different companies.
Fusion combines measurements with different uncertainties, timings and coordinate frames into one estimate of the world. The hard parts are precise time synchronisation, calibration between sensors that shifts as a vehicle ages, and deciding what to believe when two sensors disagree — which is exactly the situation that matters most.
A camera-only system can ship on every vehicle at negligible marginal cost and gathers data from the whole fleet. A lidar-equipped system costs more per vehicle and provides direct geometric measurement that does not have to be inferred. As lidar prices fall the argument shifts, which is why several manufacturers that had excluded it have reconsidered.
Sensors are useless when obscured. Production systems need heating, washing and drainage for every aperture, mounting that survives vibration and thermal cycling for years, and continuous self-diagnosis to detect a blocked or misaligned sensor. This unglamorous engineering is a large share of what makes a system deployable.
Technology dependencies are solved by engineering; supply dependencies are solved by building something, which takes years.
Automotive image sensors need high dynamic range and long-term reliability, and come from a small group of suppliers.
Automotive image sensorsCompound semiconductor emitters and sensitive detectors tie lidar to the photonics supply chain rather than the silicon one.
Compound semiconductorsA fused view of the world only exists if something can synchronise and process every stream inside the frame time, so the sensor set is specified against what the computer can consume.
Autonomy computeOther pages in this map that name Perception sensors as something they cannot do without.
What each company supplies at this step, and — where a public figure exists — its share of this specific market — with what that share measures, the period it covers and who published it. Some rows also show the company’s own reported revenue for the segment covering this step, which is a different thing: it says how much this business matters to that company, not how much of the market it holds. Not a ranking and not a recommendation.
Supplies complete vision-led sensing and driving systems rather than individual sensors.
Supplies radar and vision sensors together with the fused perception software above them.
Supplies the full sensor set — camera, radar and the software that combines them — at the largest volume.
Supplies camera, radar and lidar as one specified set for a vehicle programme.
Supplies integrated camera and radar sensing packages to Japanese programmes.
Supplies cameras, lidar and the cleaning systems that keep them usable.
Whether lidar cost falls far enough for mass-market inclusion rather than premium fitment.
Whether imaging radar closes enough of the resolution gap to reduce lidar's role.
Whether sensor cleaning and diagnostics become a differentiator in real-world reliability.
Each page explains one technology in plain language, states what it depends on, and names companies by what they supply at that step. Company roles are described qualitatively and deliberately carry no market shares, revenue figures or rankings — those change faster than an explainer can, and a stale number is worse than none. Ticker links point at company pages on this site and are provided for reference only.
Nothing here is investment advice, a recommendation, or a forecast. A company named on a page about a technology is not thereby a good investment, and the chokepoints described are structural facts about supply chains rather than predictions about prices. Technology moves; where a page describes something as unresolved or in development, that was true when it was written.
Plutux no es un asesor de inversiones. Los datos de mercado y el análisis generado por IA son solo informativos y educativos, no asesoramiento de inversión. Aviso legal