1. What is LiDAR?
LiDAR (Light Detection and Ranging) is an active sensing technology that acquires precise dimensions and spatial configurations of targets by emitting laser beams and analyzing the echo signals (e.g., Time of Flight (ToF) or frequency difference (FMCW)). Compared to passive imaging solutions that rely on ambient light (such as cameras), LiDAR possesses strong resistance to lighting interference. It can stably output 3D point cloud data with centimeter-level accuracy even in extremely dark or strong backlight conditions, making it a key means for the digital reconstruction of physical environments.
Currently, this technology has deeply empowered fields such as autonomous driving, UAV mapping, industrial robot navigation, and smart cities, becoming the core hardware support in complex perception tasks.

Autonomous driving LiDAR[1]	Mobile robot LiDAR[2]	UAV LiDAR[3]

2. Working Principle of LiDAR
A LiDAR system mainly consists of three parts: a transmitter that emits light waves, a receiver that captures the reflected light waves, and a processor that interprets the data.

Schematic diagram of LiDAR ToF working principle[4]

Transmitter (Source): Responsible for electro-optical conversion, directionally emitting stimulated radiation optical signals. Depending on the modulation method, it can emit extremely narrow pulses (pulsed waves) or continuously frequency-modulated light waves (continuous waves).
Receiver (Sensor): Utilizes high-sensitivity photodetectors (such as APD, SPAD, or SiPM) to capture echoes in real time, converting them back into electrical signals while suppressing background light noise.
Processor (Brain): Responsible for converting physical signals into structured data. Its core calculation logic is divided into two mainstream approaches:

• Time of Flight (ToF): Calculates distance by measuring the absolute time difference \(\Delta t\) of the pulse round trip, using the formula $d = \frac{c \cdot \Delta t}{2}$. This is the mainstream approach for mass-produced LiDARs today, offering the advantages of fast response speed and relatively simple system structure.
• Frequency Modulated Continuous Wave (FMCW): Acquires information by measuring the frequency shift (Doppler Effect) between the emitted light and the echo light. It can not only calculate the distance but also directly obtain the instantaneous velocity of the target through a single detection. With extremely strong anti-interference capability, it is the core evolutionary direction for next-generation high-performance LiDARs.
• Point Cloud Generation: The processor ultimately fuses the ranging information with the real-time scanning angle to generate high-density 3D point clouds containing 3D coordinates $(x, y, z)$, reflection intensity, and velocity (for FMCW).

3. Types of LiDAR
1. Classification by Scanning Method
Mechanical LiDAR

Principle: The transmitting and receiving modules are driven by a motor to perform 360-degree rotational scanning.
Features: The most mature technology, capable of achieving full horizontal coverage. However, it is relatively bulky and contains a large number of moving parts, which limits hardware lifespan and results in higher costs.    

Semi-solid-state LiDAR

Principle: The transmitting and receiving modules remain stationary, while the beam direction is altered by tiny internal moving parts (such as MEMS micromirrors, rotating mirrors, or prisms).
Features: More compact in structure than mechanical LiDAR. It is the mainstream solution for mass-produced L2+/L3 autonomous driving vehicles, striking a balance between cost and performance.

Solid-state LiDAR

OPA (Optical Phased Array): Controls the beam direction using the phase difference of multiple waveguide units, with no mechanical movement whatsoever.
Flash: Similar to a digital camera, a single pulse illuminates the entire scene and captures the image.
Features: Extremely high reliability and the smallest form factor. However, it has a shorter detection range (Flash) or extremely complex manufacturing processes (OPA). It represents the ultimate development goal for the future.

2. Classification by Ranging and Perception Technology
The ranging principle directly affects the LiDAR's accuracy, anti-interference capability, and ability to perceive moving targets:
Time of Flight (ToF)

Principle: Calculates distance by recording the time difference between the emission of the laser pulse and the reception of the reflected light.
Features: Fast response speed, high technological maturity, and wide ranging distance (up to 200 meters or more). It is currently the most widely applied technology in the market.

Frequency Modulated Continuous Wave (FMCW)

Principle: Emits continuously frequency-modulated light waves and calculates distance and instantaneous radial velocity by measuring the frequency difference between the emitted and reflected waves.
Features: Can simultaneously acquire position and velocity information (4D sensing), with extremely strong anti-interference capability and high eye safety. However, the system is complex and costly.

Triangulation Ranging

Principle: Calculates distance using the geometric triangular relationship between the emission point, reflection point, and receiver.
Features: Suitable for short-range scenarios, commonly found in robotic vacuum cleaners or consumer electronic devices (such as smartphone facial recognition).

Image Sources:
[1]: https://en.wikipedia.org/wiki/Lidar#/media/File:Cruise_Automation_Bolt_EV_third_generation_in_San_Francisco.jpg
[2]: https://en.wikipedia.org/wiki/Lidar#/media/File:LIDAR_equipped_mobile_robot.jpg
[3]: https://en.wikipedia.org/wiki/Lidar#/media/File:Yellowscan_LIDAR_on_OnyxStar_FOX-C8_HD.jpg
[4]: https://www.yellowscan.com/knowledge/how-does-lidar-work/