SICK PicoScan + IMU + EKF + SLAM Toolbox

This repository documents the full bringup pipeline for running a SICK PicoScan LiDAR with its IMU, fusing data using robot_localization (EKF), and performing 2D SLAM using slam_toolbox.

The system publishes the standard TF chain:

map → odom → base_link → laser_1 └→ sick_imu

Important

RViz can only show the scan and map moving in the world if odom → base_link changes over time. If no real odometry source exists (wheel odom / VIO / scan-matching odom), and the LiDAR is moved by hand, the scan and map will appear “stuck” in map. This is expected behavior.

---

Prerequisites

• ROS 2 (Humble / Jazzy)
• fastapi
• uvicorn
• Packages installed:
  • sick_scan_xd
  • robot_localization
  • slam_toolbox
  • tf2_ros
  • rviz2

Network example (adjust as needed):

• PicoScan IP: 192.168.0.1
• ROS PC (UDP receiver): 192.168.0.20

---

Build Workspace

cd ~/li_bot
colcon build --symlink-install
source install/setup.bash

Step 1: Launch SICK PicoScan (LiDAR + IMU)

ros2 launch sick_scan_xd sick_picoscan.launch.py \
  hostname:=192.168.0.1 \
  udp_receiver_ip:=192.168.0.20 \
  publish_frame_id:=laser \
  host_FREchoFilter:=0 \
  tf_publish_rate:=0

Expected outputs

Laser scan:

    Topic: /scan_fullframe

    Frame: laser_1

IMU:

    Topic: /sick_scansegment_xd/imu

    Frame: sick_imu

Step 2: Publish Static TF (base_link → IMU)

Declare the IMU as rigidly mounted to the robot body.

ros2 run tf2_ros static_transform_publisher \
  0 0 0 0 0 0 base_link sick_imu

    Replace the zeros with the real mounting transform when known.

Step 3: Publish Static TF (base_link → LiDAR)

If sick_scan_xd does not publish this transform, publish it manually:

ros2 run tf2_ros static_transform_publisher \
  0 0 0 0 0 0 base_link laser_1

    Do not publish duplicate TFs.
    Verify existing TFs using:

    ros2 run tf2_tools view_frames

Step 4: Start EKF (robot_localization)

The EKF publishes odom → base_link.

ros2 run robot_localization ekf_node --ros-args --params-file \
  ~/li_bot/src/picoscan_launch/config/ekf_imu.yaml

EKF Notes

For 2D SLAM, two_d_mode: true is strongly recommended.

If only IMU data is available, fusing linear acceleration will cause drift.

A stable configuration fuses yaw and yaw-rate only until wheel odometry
or VIO is added.

Step 5: Launch SLAM Toolbox (Online Async)

ros2 launch slam_toolbox online_async_launch.py \
  slam_params_file:=$HOME/li_bot/src/picoscan_launch/config/slam_config.yaml

SLAM TF responsibilities

slam_toolbox: publishes map → odom

robot_localization: publishes odom → base_link

Static TFs: publish base_link → laser_1, base_link → sick_imu

Step 6: RViz2 Visualization

Start RViz:

rviz2

Fixed Frame

Set Fixed Frame = map

Displays to add

TF

LaserScan

    Topic: /scan_fullframe