Integrar um LiDAR no ROS2: guia completo
Guia passo a passo para integrar um LiDAR Ouster, Hesai, Livox ou SICK no ROS2 (Humble/Iron/Jazzy): instalação dos drivers, configuração de rede, PTP, QoS, ficheiros de lançamento, visualização RViz2, SLAM e fusão de sensores.
1. Introdução
O LiDAR tornou-se essencial em robótica móvel para navegação autónoma, cartografia SLAM e perceção. Este guia cobre a integração dos drivers Ouster (ouster-ros), Hesai (hesai_ros_driver), Livox (livox_ros_driver2) e SICK (sick_scan_xd) no ROS2 Humble, Iron e Jazzy.
2. Instalação dos drivers
Ouster: git clone --branch ros2 https://github.com/ouster-lidar/ouster-ros.git, colcon build. Topics: /ouster/points (PointCloud2), /ouster/imu. Hesai: git clone https://github.com/HesaiTechnology/hesai_ros_driver.git. Topics: /hesai/pandar. Livox: git clone https://github.com/Livox-SDK/livox_ros_driver2.git. Topics: /livox/lidar. SICK: git clone https://github.com/SICKAG/sick_scan_xd.git. Topics: /scan (2D), /cloud (3D).
3. Configuração de rede
Configure um IP estático na mesma sub-rede que o LiDAR (ex: 192.168.1.50/24). Ative o PTP (IEEE 1588) via linuxptp para timestamping preciso: sudo ptp4l -i eth0 -m -s. Utilize a QoS BEST_EFFORT no ROS2 para as nuvens de pontos, de modo a evitar a saturação DDS.
4. Visualização e processamentos avançados
Visualize a nuvem de pontos no RViz2 (Add > PointCloud2). Para SLAM 2D, utilize slam_toolbox após projeção via pointcloud_to_laserscan. Para SLAM 3D, utilize Cartographer. A fusão LiDAR/IMU é feita com robot_localization. Para clustering e deteção de obstáculos, utilize PCL (pcl_ros).
5. Checklist e conclusão
Verifique: ROS2 sourceado, driver compilado, IP estático, ping OK, PTP ativo, QoS BEST_EFFORT, RViz2 funcional, TF publicadas e SLAM validado. Chave para uma integração bem-sucedida: configuração de rede, QoS ROS2 e árvore TF correta.
💻
Need help integrating this software solution?
We can advise you on the software, SDKs and tools best suited to your project.
Talk to an expert →Mentioned in this article
Browse the detailed pages for each mentioned entity.
Software & SDK
Flasheye
Flasheye
Edge application for LiDAR intrusion detection and counting for perimeter surveillance.
Outsight
Outsight
Real-time LiDAR perception software platform for smart city, transport and security.
RTAB-Map
OpenIntRoLab (Sherbrooke)
RGB-D and LiDAR SLAM with topological mapping and place recognition.
ROS2 Perception Pipeline
OpenOpen Robotics
ROS2 perception stack (vision, LiDAR, fusion) for mobile robots and vehicles.
NVIDIA Isaac Sim
NVIDIA
NVIDIA robotics simulator with physically accurate LiDAR sensor rendering for testing and validation.
Apollo
OpenBaidu
Open-source autonomous driving platform with LiDAR perception, HD maps and planning.
Cartographer
OpenGoogle / ROS
Google's real-time 2D and 3D SLAM library, integrated with ROS.
LIO-SAM
OpenTixiao Shan / MIT
Tightly-coupled LiDAR-IMU SLAM via factor graph optimization for high accuracy.
Autoware
OpenThe Autoware Foundation
Open-source autonomous driving framework: perception, localization, planning and control.
Hesai SDK
OpenHesai Technology
Official SDK and ROS2 drivers for Hesai LiDARs (AT, Pandar, QT, FT).
FAST-LIO / FAST-LIO2
OpenHKU MaRS Lab
Ultra-fast LiDAR-inertial SLAM with high accuracy, no additional sensor required.
RViz2
OpenOpen Robotics
The official ROS2 3D visualizer for LiDAR point clouds and other robotic data.
ROS2 Navigation (Nav2)
OpenOpen Robotics
ROS2 navigation framework for mobile robots: planning, control and LiDAR SLAM.