Robots Atlas>ROBOTS ATLAS
Robotics

LiDAR

2007ActivePublished: 28 August 2026Updated: 28 August 2026Published
Key innovation
Active 3D range sensing using laser light and time-of-flight, delivering accurate distance independent of ambient illumination — unlike passive cameras.
Category
Robotics
Abstraction level
Building block
Operation level
DataRobot controlSystem
Use cases
Autonomous vehiclesMobile and humanoid roboticsSLAM (localization and mapping)Surveying and cartographyObstacle avoidance3D scanning and modelingForestry and geologyArchaeologyAtmospheric research

How it works

An emitter (laser diode or fiber laser, typically 905 nm or 1550 nm) sends a short pulse or a modulated wave. The light reflects off an object and returns to a detector (e.g. an avalanche photodiode APD or SPAD). Timing electronics compute the time of flight: directly by measuring pulse delay (dToF) or via the phase of a modulated signal (iToF); in the FMCW variant it compares the frequencies of the transmitted and received signals, obtaining range and radial velocity at once. Distance is computed as d = c·t/2. A scanning mechanism (rotating mirror, MEMS, phased array, or flash illumination) sweeps the beam across the field of view. The collected measurements (range + angles + intensity) are assembled into a 3D point cloud that downstream algorithms use for object detection, mapping, and SLAM.

Problem solved

Robots and vehicles need accurate, metric information about distance and the 3D shape of their surroundings. Passive cameras do not measure depth directly and struggle under poor or changing lighting, while radar has low angular resolution. LiDAR provides dense, accurate 3D geometry independent of ambient light.

Components

Laser emitterGenerates the measurement beam

Light source (laser diode or fiber laser), typically 905 nm or 1550 nm, generating pulses or a modulated wave.

Official

Scanning mechanismControls beam direction

Steers the beam across the field of view: rotating mirror, MEMS, optical phased array, or flash illumination (no scanning).

Mechanical (rotating 360°)
MEMS
Optical phased array (solid-state)
Flash

Official

Detector (receiver)Captures reflected light

Photodetector capturing the returning light, e.g. an avalanche photodiode (APD) or single-photon (SPAD).

Official

Timing electronicsComputes distance

Computes time of flight (dToF) or phase/frequency (iToF/FMCW) and converts it to distance d = c·t/2.

Point cloud processingBuilds the 3D representation

Assembles range, angle, and intensity measurements into a 3D point cloud used by perception, mapping, and SLAM.

Official

Implementation

Implementation pitfalls
Sensitivity to weatherHigh

Rain, fog, snow, dust, and exhaust scatter the beam, reducing range and introducing spurious points.

Fix:Noise filtering, longer wavelengths (1550 nm), fusion with radar and camera.
Low-reflectivity or specular surfacesMedium

Dark, wet, or mirror-like objects return little light to the detector, causing gaps in the point cloud.

Fix:Higher power/sensitivity, SPAD detectors, sensor fusion.
Motion distortionMedium

Sequential scanning while the sensor moves warps the point cloud unless de-skewed.

Fix:Compensation using IMU/odometry and per-point timestamps.
Mutual interferenceLow

With many lidars present, pulses from other units can be falsely detected.

Fix:Pulse coding, randomized emission timing.

Evolution

1960
Invention of the laser
Inflection point

Theodore Maiman demonstrates the first working (ruby) laser, enabling active light-based range measurement.

1961
First lidar-like system

Hughes Aircraft Company builds an early laser rangefinder/tracking system.

1963
Term "lidar" coined

The term "lidar" appears in print; the "Colidar Mark II" rangefinder is produced.

1971
Apollo 15 — laser altimeter

Apollo 15 astronauts use a laser altimeter to map the lunar surface, popularizing the technology.

2005
DARPA Grand Challenge

The competition exposes the limits of single-line lidar and cameras for autonomous driving, motivating 3D lidar.

2007
Velodyne HDL-64E — 3D lidar
Inflection point

Velodyne (David Hall) introduces a rotating 64-beam sensor producing a 360° 3D map (~1M points/s) — a breakthrough for autonomous vehicles.

Hyperparameters (configurable axes)

WavelengthHigh

Typically 905 nm (cheaper, eye-safety limited) or 1550 nm (longer range, more eye-safe).

905 nm
1550 nm
RangeCritical

Maximum measurement distance; depends on laser power, detector sensitivity, and target reflectivity.

50–250 m
Angular resolutionHigh

Angular density of points; determines the ability to resolve small and distant objects.

Points per secondMedium

Measurement throughput; modern sensors reach hundreds of thousands to millions of points/s.

~1 000 000 pkt/s (Velodyne HDL-64E)
Field of viewHigh

Angular coverage horizontally and vertically; mechanical units give 360° horizontal, solid-state usually a sector.

Scanning mechanismHigh

How the beam is swept: mechanical, MEMS, phased array, or flash.