Fusing IMU and Odometry

This is a demo fusing IMU data and Odometry data (wheel odom or Lidar odom) or GPS data to obtain better odometry.

Currently, I implement Extended Kalman Filter (EKF), batch optimization and isam2 to fuse IMU and Odometry data.

Dependencies

This demo has the following dependencies:

• ROS

  • I test it on ROS melodic and kinetic version

• GTSAM

  • use tag 4.28a now

• Kalman Filter Library

Build

Clone the repository and catkin_make:

cd ~/catkin_ws/src
git clone https://github.com/Guo-ziwei/fusion
cd ../
catkin_make
source ~/catkin_ws/devel/setup.bash

run EKF:

rosrun ekf_fusion ekf_fusion_node bag_filename.bag

run batch:

rosrun gtsam_imu imu_preintegrated bag_filename.bag

it will take a long time to get the result and I will fix it in the future

run isam2:

rosrun gtsam_imu imu_preintegrated_sm bag_filename.bag

Performance

Test on MIt dataset

<img src="https://github.com/Guo-ziwei/fusion/blob/master/image/result_ekf.png" width = 50% height = 50% div align=center />

TODO

Now the demo is running off-line (reading data from bag) and I will turn it to online.