In this demo, a Kalman filter is applied to track the pose of an automotive. IMU data and speed information1 are collected through Carla rosbridge. In the update step, the speed measurement is used to update the states. Moreover, non-holonomic constraints are applied.
- Linux and ROS (has been test on Ubuntu 20.04 with ROS Noetic).
- Carla Simulator and Carla rosbridge. Make sure your Python can find carla package, Python 3.7 is recommended.
- First, clone into you catkin workspace
cd catkin_ws/src
git clone https://github.com/LHengyi/ESKF-for-Automotive-Dead-Reckoning.git - Build
catkin_make - To run
start Carla server
CarlaUE4.shroslaunch automotive_dead_reckoning carla_localization_ad_rosbridge.launch in another terminal
roslaunch automotive_dead_reckoning wheel_ins.launch 
Footnotes
-
In practice, vehicle speed information can be obtained through on-board interface, visual odometry or even by binding an IMU on a non-steering wheel. ↩