Supply Chain > ADAS/AV Stack > Sensors overview


EV Sensors Overview


ADAS (Advanced Driver-Assistance Systems) and AV (Autonomous Vehicle) sensors provide the raw observations used for perception, sensor fusion, and safety decisions. Most modern architectures use a multi-sensor suite: cameras for rich scene understanding, radar for robust ranging and velocity, and (in some systems) LiDAR for high-precision 3D depth. Additional sensors such as ultrasonics, IMUs (Inertial Measurement Units), and vehicle motion sensors support close-range awareness, localization, and control stability.


Sensor suite at a glance

Most ADAS systems rely on complementary sensing modalities to reduce blind spots and handle diverse lighting and weather conditions.

Sensor type Primary contribution Strengths Constraints Typical use
Cameras Visual scene understanding High detail, classification, lane and object recognition Lighting and glare sensitivity; occlusion L2 ADAS baseline; essential for many stacks
Radar Range and velocity Works in many weather/lighting conditions; strong for relative speed Lower spatial resolution than vision; multipath artifacts ACC/AEB; long-range perception; redundancy
LiDAR (where used) 3D depth and geometry High-precision depth; strong spatial structure Cost/packaging; cleaning; performance varies in adverse conditions Higher autonomy stacks; premium ADAS in some OEMs
Ultrasonics Short-range proximity Low cost; good for parking-range distances Short range; environmental sensitivity Parking assist; low-speed maneuvers
IMU (Inertial Measurement Unit) Acceleration and rotation High-rate motion estimation; supports stability and localization Bias drift; requires fusion with other sensors Localization support; stability; sensor fusion
Vehicle motion sensors Wheel speed, steering angle, yaw rate Direct motion-state inputs for control Not environment perception All ADAS control loops
GNSS (Global Navigation Satellite System) Absolute positioning and timing Wide-area positioning; time base Urban canyon multipath; outages Navigation; fleet location; localization augmentation

Camera systems

Cameras are typically the highest-volume, highest-bandwidth ADAS sensor. A modern ADAS vehicle may use multiple cameras for forward, rear, and surround coverage.

  • Common roles: forward perception, lane keeping, traffic sign recognition, surround view, driver monitoring (DMS) where equipped
  • Key hardware blocks: image sensor, lens, ISP (Image Signal Processor), serializer/PHY, enclosure and heater (where used)
  • Supply-chain drivers: sensor resolution, HDR performance, low-light capability, and automotive qualification

Radar

Radar provides range and velocity information and is robust across many lighting conditions.

  • Common roles: adaptive cruise control, automatic emergency braking, blind-spot monitoring
  • Key hardware blocks: radar transceiver, antenna array, radar processing silicon, enclosure and radome
  • Integration note: radar may connect via Ethernet or CAN depending on bandwidth and architecture

LiDAR (where used)

LiDAR provides high-fidelity depth and 3D structure. It appears in some higher autonomy stacks and some premium ADAS implementations.

  • Common roles: precise depth perception, redundancy and structure for localization
  • Key hardware blocks: emitter/receiver, optics, scanning mechanism (if used), processing, cleaning/heating provisions
  • Integration note: typically high-bandwidth; Ethernet is common

Ultrasonics

Ultrasonics are low-cost short-range sensors used for low-speed maneuvers.

  • Common roles: parking assist, obstacle proximity sensing
  • Constraints: limited range and resolution; sensitive to surface and environmental conditions

Localization and motion sensors

ADAS/AV stacks rely on localization support beyond environment perception.

  • IMU: high-rate acceleration and rotation inputs for stability and localization support
  • Wheel speed and steering angle: core signals for motion control and state estimation
  • GNSS: absolute position and time base; often fused with inertial and wheel sensors

Sensor interfaces to compute and the IVN

Sensor data must move to compute with adequate bandwidth and predictable latency. Interface choices strongly shape wiring, cost, and scalability.

Interface Typical sensor fit Why used Notes
Automotive Ethernet Cameras, LiDAR, high-content radar, zonal aggregation High bandwidth; scalable topologies Requires PHYs/switches and attention to signal integrity
CAN / CAN-FD Radar modules (in some designs), status/control messaging Deterministic control and diagnostics Not suitable for high-rate camera streaming
Direct point-to-point (varies) Certain camera/radar topologies Simplify dedicated links Often evolves toward Ethernet as systems scale

Synchronization and time-stamping

Sensor fusion quality depends on timing alignment between sensors and compute.

  • Time synchronization: common time base across sensors and compute nodes
  • Accurate time-stamping: required for associating measurements across modalities
  • Hardware implication: clocks and synchronization support in sensors, switches, and compute

Cleaning, heating, and packaging support

Sensors must operate in real-world environments. Hardware support systems often determine reliability as much as the sensors themselves.

  • Heaters and defogging: keep lenses and radomes clear
  • Washers and air jets (where used): remove debris from critical sensors
  • Mounting and calibration stability: vibration, thermal expansion, and alignment retention
  • Environmental sealing: water ingress and corrosion control

Sensor redundancy (hardware perspective)

Redundancy is achieved by overlapping modalities and fields of view.

  • Modality redundancy: cameras plus radar; LiDAR where used
  • Coverage redundancy: multiple sensors covering critical forward and side zones
  • Power and data path considerations: independent power rails and network paths where required

Supply-chain notes

Sensor content is a major driver of ADAS BOM cost and supplier differentiation.

  • Cameras: image sensors, lenses, ISPs, and packaging quality drive capability and reliability
  • Radar: RF silicon, antenna design, and radome packaging drive performance
  • LiDAR: optics, emitter/receiver technology, cleaning/heating needs drive cost and packaging complexity
  • Interconnect: Ethernet PHYs, switches, connectors, and harness design become critical at scale