Setting Up Sensors

Why Sensors Matter

Odometry tells the robot where it thinks it is based on how far its wheels have turned. Sensors tell the robot what's actually around it. Without sensors, the robot is blind. It can't detect obstacles, can't build a map of its environment, and can't localize itself within a map.

linorobot2 uses two primary sensor categories:

• LaserScan (2D lidar): produces a 360° ring of distance measurements on a flat plane. This is the primary sensor for obstacle detection and SLAM.
• Depth camera (RGBD): produces a 3D point cloud and/or depth image. Useful for detecting obstacles that a floor-level lidar misses (like table legs above the scan plane).

Both sensor types publish data that Nav2 and SLAM Toolbox consume directly.

Physical Robot Setup

On a physical robot, sensors are configured during installation via the install.bash script's --laser and --depth arguments. The script installs the appropriate ROS2 driver packages and sets environment variables that the bringup launch files use to start the correct driver at boot.

Supported Laser Sensors

Argument	Sensor
a1	RPLIDAR A1
a2	RPLIDAR A2
a3	RPLIDAR A3
s1	RPLIDAR S1
s2	RPLIDAR S2
s3	RPLIDAR S3
c1	RPLIDAR C1
ld06	LD06 LIDAR
ld19	LD19/LD300 LIDAR
stl27l	STL27L LIDAR
ydlidar	YDLIDAR
xv11	XV11
realsense *	Intel RealSense D435, D435i
zed *	Zed
zed2 *	Zed 2
zed2i *	Zed 2i
zedm *	Zed Mini

Sensors marked with * are depth cameras used as laser sensors. When configured this way, the launch files automatically run depthimage_to_laserscan to convert the depth image into a sensor_msgs/LaserScan message on the /scan topic, making it compatible with SLAM Toolbox and Nav2 without any extra configuration.

Supported Depth Sensors

Argument	Sensor
realsense	Intel RealSense D435, D435i
zed	Zed
zed2	Zed 2
zed2i	Zed 2i
zedm	Zed Mini
oakd	OAK-D
oakdlite	OAK-D Lite
oakdpro	OAK-D Pro

Simulated Robot

In Gazebo, sensors are defined in URDF/xacro files and are already configured for each robot type. You don't need to install any additional drivers. Gazebo simulates the sensors and publishes on the same topics as real hardware.

Sensor URDF Files

Sensor definitions live in linorobot2_description/urdf/sensors/:

generic_laser.urdf.xacro: Simulates a 360° 2D lidar:

• 360 rays, full 360° sweep
• Range: 0.08 m to 12.0 m
• Update rate: 10 Hz
• Publishes to: /scan (topic name: scan, frame ID: laser)

depth_sensor.urdf.xacro: Simulates an RGBD camera:

• Resolution: 640 × 480
• Update rate: 30 Hz
• Horizontal FOV: ~86°
• Depth range: 0.3 m to 100 m
• Publishes to: /camera topics

imu.urdf.xacro: Simulates an IMU:

• Publishes to: /imu/data at 50 Hz

How Sensors Are Included

Each robot type's URDF (linorobot2_description/urdf/robots/<robot_type>.urdf.xacro) already includes the laser and depth sensor:

<xacro:generic_laser>
  <xacro:insert_block name="laser_pose" />
</xacro:generic_laser>

<xacro:depth_sensor>
  <xacro:insert_block name="depth_sensor_pose" />
</xacro:depth_sensor>

The laser_pose and depth_sensor_pose blocks define where the sensors are mounted relative to base_link. These are defined in the robot's properties file.

Configuring Sensor Positions

Sensor positions are set in linorobot2_description/urdf/<robot_type>_properties.urdf.xacro. For example, in 2wd_properties.urdf.xacro:

<xacro:property name="laser_pose">
  <origin xyz="0.12 0 0.33" rpy="0 0 0"/>
</xacro:property>

<xacro:property name="depth_sensor_pose">
  <origin xyz="0.14 0.0 0.045" rpy="0 0 0"/>
</xacro:property>

All positions are measured from base_link (the center of the robot's base). Change these values to match where your sensors are physically mounted. After editing the xacro file, rebuild the workspace:

cd <robot_computer_ws>
colcon build

More details on how these poses relate to the TF tree are covered in the Transforms section.

Published Topics

After bringup (Physical Robot) or Gazebo launch (Simulated Robot), the following sensor topics should be active:

Topic	Type	Source
/scan	sensor_msgs/LaserScan	2D lidar or depth-to-laser conversion
/camera/image_raw	sensor_msgs/Image	Depth camera RGB image
/camera/depth/image_raw	sensor_msgs/Image	Depth camera depth image
/imu/data	sensor_msgs/Imu	IMU data. Published directly by the firmware (no magnetometer) or by the Madgwick filter (magnetometer enabled). In the Simulated Robot, published by imu.urdf.xacro
/imu/data_raw	sensor_msgs/Imu	Raw IMU acceleration and gyro data (Physical Robot only, when magnetometer is enabled)
/imu/mag	sensor_msgs/MagneticField	Magnetometer data (Physical Robot only, when magnetometer is present)

You can verify a sensor is publishing correctly with:

ros2 topic hz /scan

What's Next

With sensors publishing data, the next step is to make sure the robot knows where those sensors are mounted in 3D space, so it can correctly interpret a laser hit at angle X as being at a specific position in the world. That's the job of the TF tree, covered in Setting Up Transforms.