3D LiDAR

The primary advantage is volumetric perception. While 2D LiDAR only sees a "slice" of the world at a specific height, 3D LiDAR sees the entire environment, including overhanging obstacles (like tables) and negative obstacles (like loading dock drops or stairs), preventing collisions that a single-plane sensor would miss.

Yes. Because LiDAR is an active sensor (it emits its own light source), it works perfectly in total darkness. High-quality 3D LiDARs are also designed to filter out ambient sunlight, making them reliable for outdoor use where camera-based systems often struggle with glare or shadows.

Mechanical LiDARs physically spin a laser array to achieve a 360-degree view, which offers great coverage but contains moving parts prone to wear. Solid-state LiDARs steer the beam electronically; they are more durable and compact but often have a narrower Field of View (FoV), sometimes requiring multiple sensors for full coverage.

Generally, yes. Processing 3D point clouds (which can contain hundreds of thousands of points per second) is computationally intensive compared to 2D scan lines. You will likely need an industrial PC with a GPU or a dedicated AI accelerator (like NVIDIA Jetson) to handle 3D SLAM and object detection in real-time.

This remains a challenge for LiDAR technology. Laser pulses often pass through clear glass or reflect off it unpredictably (specular reflection). Most robotics stacks combine LiDAR data with ultrasonic sensors or cameras, or use software-defined "keep-out zones," to handle glass walls effectively.

Heavy rain and fog can introduce "noise" into the point cloud as lasers reflect off droplets. However, modern 3D LiDARs often feature "multi-echo" technology, which can distinguish between a raindrop (first echo) and the wall behind it (last echo), allowing the software to filter out weather interference.

Yes, standard industrial LiDARs are Class 1 laser products, meaning they are eye-safe under all normal conditions of use. They operate at wavelengths that do not damage the human retina, even if looked at directly.

For indoor warehousing, a range of 30 to 50 meters is usually sufficient. This allows the robot to see the end of aisles and localize against distant walls. Outdoor mobile robots typically require ranges of 100 meters or more to safely brake at higher speeds.

Traditionally, navigation LiDARs were not safety-rated. However, "Safety 3D LiDARs" are entering the market that are certified for functional safety (PL-d). Without a certified sensor, you must use designated 2D safety scanners for the emergency stop circuit, using the 3D LiDAR only for navigation and non-critical obstacle avoidance.

Most major LiDAR manufacturers (Velodyne, Ouster, Hesai, Robosense) provide open-source ROS and ROS2 drivers. These drivers publish data in the standard `sensor_msgs/PointCloud2` format, making it compatible with standard packages like PointCloud Library (PCL), OctoMap, and Navigation2.

Physical maintenance is minimal but critical: the lens or window must be kept clean. Dust, oil, or mud on the lens can cause measurement errors. Many outdoor units include automated cleaning systems (wipers or air jets). Mechanical units eventually require motor replacement, whereas solid-state units are maintenance-free regarding moving parts.