LiDAR

LiDAR (Light Detection and Ranging) is an indispensable core sensor in autonomous driving systems. By emitting laser beams and receiving reflected signals, it constructs high-precision 3D environmental models for vehicles. Its primary advantage lies in delivering real-time depth and distance information far superior to cameras, unaffected by ambient lighting conditions—earning it the reputation as the vehicle’s "eyes of depth perception".

💡 Working Principle: Two Mainstream Ranging Technologies

LiDAR ranging primarily relies on two core technologies, each catering to distinct priorities in performance and cost:

Time of Flight (ToF, Currently Mainstream): The operating principle is straightforward—it fires a laser pulse and measures the time difference (△t) taken for the beam to return after hitting an obstacle. Based on the constant speed of light (c), the distance is calculated via d = c·△t/2. Highly mature and responsive, this remains the dominant mainstream solution today.
Frequency Modulated Continuous Wave (FMCW, Future Trend): This technology emits a continuous laser beam with a time-varying frequency, measuring distance through optical interference with the reflected light. A key advantage is its ability to instantly capture an object's radial velocity via the Doppler effect, which is crucial for tracking the dynamic movements of pedestrians or cross-traffic as it transitions from the laboratory to real-world deployment.

🔭 Core Technical Specifications and Classification

Key Performance Indicator: Line Count (Channels)

"Line count" dictates a LiDAR sensor's perception accuracy. It refers to the number of independent laser channels emitted simultaneously along the vertical axis. Think of it as the sensor’s "resolution"—the higher the line count, the finer the environmental scan, resulting in a denser and sharper "point cloud" image. The industry is rapidly advancing from dozens of lines to thousands of lines and beyond, ushering in the era of true "image-grade" spatial perception.

Two Main Wavelength Approaches

LiDAR systems are broadly categorised by their operating wavelengths, each offering distinct performance characteristics:

Wavelength	Core Advantages	Main Limitations
905 nm	Compatible with mature silicon-based photodetectors, keeping costs low and facilitating large-scale mass production.	Power output is limited, restricting detection range to typically under 200 metres; susceptible to glare and strong ambient light interference.
1550 nm	Significantly safer for human eyes, allowing higher power output with detection ranges reaching 300 metres or even up to 900 metres.	Requires costly detector materials such as Indium Gallium Arsenide (InGaAs), resulting in higher unit costs.

Three Major Architectural Formats

Based on their scanning mechanisms, LiDAR systems have evolved from mechanical assemblies to solid-state designs:

Mechanical: Uses physical rotating mechanisms for 360° scanning; highly mature, but costly, bulky, and difficult to meet stringent automotive-grade reliability standards.
Hybrid Solid-State (Semi-Solid-State): Employs solutions like MEMS micro-mirrors to direct the beam through micro-oscillations, significantly reducing mechanical wear. This is currently the mainstream choice for mass-production passenger cars.
Pure Solid-State: Completely eliminates moving parts using Flash or Optical Phased Array (OPA) technology, delivering maximum reliability, though the technology is still being refined for commercial maturity.

🚗 Industry Landscape and Trends: From "Optional Extra" to "Standard Fitment"

While select carmakers such as Tesla remain committed to a camera-only "pure vision" approach, the majority of mainstream manufacturers and emerging EV players are adopting a hybrid "Vision + LiDAR" sensor fusion strategy, treating LiDAR as standard fitment to enhance safety redundancy.

Rapid Cost Reduction: Early mechanical LiDAR units commanded prices as steep as RM700,000. However, driven by technological advancements and economies of scale, costs have dropped dramatically towards the RM1,000 mark, accelerating adoption from flagship luxury models down to the mid-tier vehicle market.
Market Landscape: Chinese manufacturers (such as RoboSense, Hesai, Huawei, and Seyond) have emerged as dominant global players, commanding strong international competitiveness across both technology and cost.