I. Classification of LiDAR Sensors
1. 2D LiDAR sensor
2D LiDAR sensors are primarily used to measure and generate maps on a horizontal plane. They emit a laser beam perpendicular to the horizontal plane and generate a map by measuring the position of obstacles through the reflection of the laser beam.
2. 3D LiDAR sensor
3D LiDAR sensors can generate three-dimensional point cloud data, accurately measuring the position, shape, and size of target objects. 3D LiDAR sensors can be divided into rotating and solid-state types. Rotating 3D LiDAR is the most commonly used type; it can quickly scan a circle and generate a three-dimensional map using the reflected beam.
3D LiDAR sensors have been widely used in robotics, autonomous driving, and other fields. However, different LiDAR sensors also differ in their respective advantages and disadvantages.
1. Rotating 3D LiDAR sensor
Rotating 3D LiDAR sensors have advantages such as high precision, high resolution, and high reliability, but they are relatively large in size and high in cost.
2. Solid-state 3D LiDAR sensor
Solid-state 3D LiDAR sensors are small, relatively inexpensive, and more durable. However, their resolution and accuracy are not as good as rotating 3D LiDAR.
3. Multi-line lidar sensor
Multi-line LiDAR sensors are suitable for short-range applications, capable of acquiring high-density 3D point cloud data, and are less susceptible to light interference. However, compared to rotating 3D LiDAR, multi-line LiDAR is larger and cannot perform high-speed scanning.
In summary, when selecting a suitable LiDAR sensor, it is necessary to comprehensively consider factors such as the application scenario and budget.
3. Multi-line lidar sensor
Multi-line lidar sensors use multiple laser beams to simultaneously acquire data from multiple angles. Compared to other types of lidar, multi-line lidar sensors are more suitable for scenarios with rapidly changing environments or requiring high-speed data acquisition.
II. Application of LiDAR Sensors in Vehicle Systems
1. High-precision vehicle positioning
With the help of positioning systems, vehicle systems can make decisions by obtaining real-time location information. However, positioning methods are susceptible to signal interference, especially when passing through urban buildings and tunnels where signals are easily interrupted. To obtain accurate positioning, lidar compares the vehicle's initial position with high-precision map information.
(1) The initial position of the vehicle is provided by inertial navigation devices, global positioning systems, and sensors such as wheel speed;
(2) Extract features from the local point cloud information of the lidar and combine it with the initial position to obtain vector features in the global coordinate system;
(3) The vector features are matched with the high-precision map feature information to obtain the accurate vehicle location. Therefore, compared with other vehicle-mounted sensors, LiDAR has obvious advantages in terms of positioning accuracy and stability.
2. Obstacle recognition and target tracking
LiDAR can perform real-time scanning without relying on illumination, with a scanning angle of up to 360° and relatively low computational requirements.
During scanning, obstacles (vehicles, people, guardrails, etc.) are first identified to obtain their spatial location. For obstacle classification and tracking, the target is first associated with the segmented point cloud to confirm whether the upper and lower frames belong to the same object before target tracking is performed and target tracking information is output.
3. Automatic parking system
LiDAR installed on the roof or around the vehicle can detect the location of the parking space, sense surrounding vehicles and other obstacles, and input the acquired information into the vehicle control system to provide reliable environmental information for the control decisions of automatic parking.
4. Lane Keeping Assist System
By using LiDAR to detect lane lines, the lane keeping assist system can make decisions through logical operations when it detects a vehicle deviating from the expected driving trajectory, thereby controlling the vehicle to drive along the predetermined lane path. The LiDAR lane line detection method can be based on the density of radar scan points. To obtain the required lane lines, this method first acquires the coordinate information of the radar scan points, then converts it into a grid image, and finally extracts the lane lines using the density of points in the grid image. This method offers good real-time performance.
