cheetah_inekf_lcm

This project is the wrapper of invariant-ekf for MiniCheetah. It takes input from LCM/UDP, and performs EKF update on IMU-contact odometry.

Configuring yaml settings files

1. Change lcm_enable_debug_output param in config/settings.yaml to false if no debug output is wanted for receiving lcm messages
2. Change project_root_dir param in config/settings.yaml to filepath to your installation directory for this repo
3. Change estimator_enable_debug param in config/settings.yaml to true if you want to view the state of the inekf in the terminal while it is running
4. Change estimator_publish_visualization_markers param in config/settings.yaml to true if you want to publish the robot pose over LCM channel "LCM_POSE_CHANNEL"
5. Change estimator_lcm_pose_channel param in config/settings.yaml to true if you want to change the name of the LCM channel that the robot pose is published over
6. Change estimator_static_bias_initialization param in config/settings.yaml to true if you want to initialize static bias for the IMU
7. Change system_enable_pose_publisher param in config/settings.yaml to true if you want to save the robot pose to file and publish the robot pose over ROS
8. Change system_inekf_pose_filename and system_inekf_tum_pose_filename params in config/settings.yaml to a different filepath to specify which files you would like the robot poses to be saved to (the second is a tum syntax)

Helpful Commands:

Generating LCM Types:

1. cd cheetah_inekf_lcm_root_directory/scripts
2. bash ./make_types.sh

Running Cheetah Estimator

1. cd ~/pathto/catkin_ws
2. In a new terminal in the catkin_ws, do catkin_make (perhaps multiple times)
3. Run source ~/devel/setup.bash
4. In the same terminal, run rosrun cheetah_inekf_lcm cheetah_estimator
5. Run lcm file lcm-logplayer --speed=1.0 --lcm-url=udpm://239.255.76.67:7667?ttl=2 NAME_OF_LCM_LOG_FILE_HERE
6. The terminal should begin printing out the robot state if the settings.yaml output variables are enabled

Debugging Inekf Output

1. Start running the cheetah estimator using the instructions above
2. Enter rviz in the terminal
3. Select Add by topic setting and select path
4. Changed fixed frame to the same value as map_frame_id in config/settings.yaml
5. The robot pose should begin being drawn in rviz