Official ROS1/ROS2 drivers for Ouster sensors

ROS1 (melodic/noetic) | ROS2 (rolling/humble/iron/jazzy) | ROS2 (galactic/foxy)

ROS Version	Build Status (Linux)
ROS1 (melodic/noetic)
ROS2 (rolling/humble/iron/jazzy)
ROS2 (galactic/foxy)

• Official ROS1/ROS2 drivers for Ouster sensors
  • Overview
  • Supported Devices
  • Requirements
  • Getting Started
  • Usage
  • Launching Nodes
    • Sensor Mode
    • Recording Mode
    • Replay Mode
    • PCAP Replay Mode
    • Multicast Mode (experimental)
  • Launch Files Arguments
  • Invoking Services
    • GetMetadata
    • GetConfig
    • SetConfig
  • License

Overview

This ROS package provide support for all Ouster sensors with FW v2.0 or later. Upon launch the driver will configure and connect to the selected sensor device, once connected the driver will handle incoming IMU and lidar packets, decode lidar frames and publish corresponding ROS messages on the topics of /ouster/imu and /ouster/points. In the case the used sensor supports dual return and it was configured to use this capability, then another topic will published named /ouster/points2 which corresponds to the second point cloud.

Supported Devices

The driver supports the following list of Ouster sensors:

• OS0
• OS1
• OS2
• OSDome

You can obtain detailed specs sheet about the sensors and obtain updated FW through the website downloads section.

Requirements

This package only supports Melodic and Noetic ROS distros. Please refer to ROS online documentation on how to setup ros on your machine before proceeding with the remainder of this guide.

In addition to the base ROS installation, the following ROS packages are required:

sudo apt install -y                     \
    ros-$ROS_DISTRO-pcl-ros             \
    ros-$ROS_DISTRO-rviz

where $ROS-DISTRO is either melodic or noetic.

Note
Installing ros-$ROS_DISTRO-rviz package is optional in case you didn't need to visualize the point cloud using rviz but remember to always set viz launch arg to false.

Additional dependenices:

sudo apt install -y         \
    build-essential         \
    libeigen3-dev           \
    libjsoncpp-dev          \
    libspdlog-dev           \
    libcurl4-openssl-dev    \
    cmake

Note
You may choose a different ssl backend for the curl library such as libcurl4-gnutls-dev or libcurl4-nss-dev

Note
To use the PCAP replay mode you need to have libpcap-dev installed

Getting Started

To build the driver using ROS you need to clone the project into the src folder of a catkin workspace as shown below:

mkdir -p catkin_ws/src && cd catkin_ws/src
git clone --recurse-submodules https://github.com/ouster-lidar/ouster-ros.git

Next to compile the driver you need to source the ROS environemt into the active termainl:

source /opt/ros/<ros-distro>/setup.bash # replace ros-distro with 'melodic' or 'noetic'

Finally, invoke catkin_make command from within the catkin workspace as shown below:

cd catkin_ws
catkin_make --cmake-args -DCMAKE_BUILD_TYPE=Release

Specifying Release as the build type is important to have a reasonable performance of the driver.

Usage

Launching Nodes

The package supports three modes of interaction, you can connect to a live sensor, replay a recorded bag or record a new bag file using the corresponding launch files. Recently, we have added a new mode that supports multicast. The commands are listed below:

Sensor Mode

The driver offers two launch files to connect to an Ouster sensor: sensor.launch and driver.launch; they differ in terms of how the processing of incoming packets is performed. sensor.launch spawns three nodelets, one to connect to the sensor and publishes raw packets to the two other nodelets which handles converting them into Imu, Image and PointCloud2 messages. Meanwhile, driver.launch file spawn a single nodelet that handles all of these tasks. You can invoke the two files in the same way. The following line shows how to run the node using driver.launch:

roslaunch ouster_ros driver.launch      \
    sensor_hostname:=<sensor host name> \
    metadata:=<json file name>          # optional

driver.launch offers better performance and reduced overhead on the ROS bus, thus it is preferred over sensor.launch. sensor.launch is mainly provided for backward compatibilty.

Note: If you observe parts of the scan missing is missing, this suggests having a lots of dropped packets It is recommended that you increase the maximum allowed size for receive memory buffers in network subsystem, you can do so by running the script network-configure.bash under util.

Recording Mode

Note As of package version 8.1, specifiying metadata file is optional since the introduction of the metadata topic

roslaunch ouster_ros record.launch      \
sensor_hostname:=<sensor host name> \
bag_file:=<optional bag file name>  \
metadata:=<json file name>          # optional

Replay Mode

Note As of package version 8.1, specifiying metadata file is optional if the bag file being replayed already contains the metadata topic

roslaunch ouster_ros replay.launch      \
    bag_file:=<path to rosbag file>     \
    metadata:=<json file name>          # optional if bag file has /metadata topic

PCAP Replay Mode

Note To use this feature you need to compile the driver with BUILD_PCAP option enabled

roslaunch ouster_ros replay_pcap.launch     \
    pcap_file:=<path to ouster pcap file>   \
    metadata:=<json file name>              # required

Multicast Mode (experimental)

The multicast launch mode supports configuring the sensor to broadcast lidar packets from the same sensor (live) to multiple active clients. You initiate this mode by using sensor_mtp.launch file to start the node. You will need to specify a valid multicast group for the udp_dest argument which the sensor is going to broadcast data to it. You will also need to set mtp_main argument to true, this is need to configure the sensor with the specified udp_dest and any other sensor settings. You can control on which ip (IP4 only) you wish to receive the data on this machine from the multicast group using the mtp_dest argument follows:

roslaunch ouster_ros sensor_mtp.launch      \
    sensor_hostname:=<sensor host name>     \
    udp_dest:=<multicast group ip (ipv4)>   \
    mtp_main:=true                          \
    mtp_dest:=<client ip to receive data>   # mtp_dest is optional

Using a different machine that belongs to the same netwok subnet, you can start another instance of the client to start receiving sensor messages through the multicast group as shown below (note that mtp_main is set to false):

roslaunch ouster_ros sensor_mtp.launch      \
    sensor_hostname:=<sensor host name>     \
    udp_dest:=<multicast group ip (ipv4)>   \
    mtp_main:=false                         \
    mtp_dest:=<client ip to receive data>   # mtp_dest is optional

Note: In both cases the mtp_dest is optional and if left unset the client will utilize the first available interface.

Launch Files Arguments

Each of the previously mentioned launch files include a variety of launch arguments that helps the user customize the driver behaivor. To view the arguments that each launch file provides and their purpose pass --ros-args along with the specific launch file that you are interested in. For example, to view launche arguments of the driver.launch use the following command:

roslaunch ouster_ros driver.launch --ros-args

The command should list all available arguments, whether they are optional or required and the description and posible values of each argument.

New launch file parameter: point_type: This parameter allows to customize the point cloud that the driver produces through its /ouster/points topics. Choose one of the following values:

• original: This uses the original point representation ouster_ros::Point of the ouster-ros driver.
• native: directly maps all fields as published by the sensor to an equivalent point cloud representation with the additon of ring and timestamp fields.
• xyz: the simplest point type, only has {x, y, z}
• xyzi: same as xyz point type but adds intensity (signal) field. this type is not compatible with the low data profile.
• xyzir: same as xyzi type but adds ring (channel) field. this type is same as Velodyne point cloud type this type is not compatible with the low data profile.

Invoking Services

To execute any of the following service, first you need to open a new terminal and source the castkin workspace again by running the command: source catkin_ws/devel/setup.bash

GetMetadata

To get metadata while connected to a live sensor or during a replay session invoke the following command:

rosservice call /ouster/get_metadata

GetConfig

To get the current config of a live sensor, invoke the command:

rosservice call /ouster/get_config

SetConfig

To change config via a file while connected to a live sensor, invoke the command:

rosservice call /ouster/set_config "config_file: '<path to sensor config>'"

Note Changing settings is not yet fully support during a reset operation (more on this)

For further detailed instructions refer to the main guide

License

License File