oscillation_ctrl/README.md

# oscillation_ctrl

Repo containing oscillation damping controller for tether missions + instructions how to to set up

Cesar Rodriguez 

cesar.rodriguez@spirirobotics.com

February 2022

Steps to recreate stable PX4 environment + working repo

## 1) Installing ROS Melodic

### Setup sources.list
	sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'
### Setup keys
    sudo apt install curl # if you haven't already installed curl
    curl -s https://raw.githubusercontent.com/ros/rosdistro/master/ros.asc | sudo apt-key add -
### Installation
	sudo apt update
	sudo apt install ros-melodic-desktop-full
### Environment setup
	echo "source /opt/ros/melodic/setup.bash" >> ~/.bashrc
### Dependencies
	sudo apt install python-rosdep python-rosinstall python-rosinstall-generator python-wstool build-essential

Initilize rosdep:

	sudo apt install python-rosdep
	sudo rosdep init
	rosdep update
### PX4 Dependencies
	sudo apt-get install python-catkin-tools python-rosinstall-generator -y
	wstool init ~/catkin_ws/src
### MAVLINK
	rosinstall_generator --rosdistro melodic mavlink | tee /tmp/mavros.rosinstall
### MAVROS
	rosinstall_generator --upstream mavros | tee -a /tmp/mavros.rosinstall
### Create workspace and deps
	cd ~/catkin_ws
	wstool merge -t src /tmp/mavros.rosinstall
	wstool update -t src -j4
	rosdep install --from-paths src --ignore-src -y
### Install geographic datasets
	cd ~/catkin_ws
	sudo ./src/mavros/mavros/scripts/install_geographiclib_datasets.sh
### Build source
	cd ~/catkin_ws	
	catkin build
	
## 2) PX4 Environment Development

### Download PX4 source code
	git clone https://github.com/PX4/PX4-Autopilot.git --recursive

### Run ubuntu.sh
	bash ./PX4-Autopilot/Tools/setup/ubuntu.sh
	#Restart computer after it is done

### Build ROS and Gazebo - This defaults to Gazebo9
	wget https://raw.githubusercontent.com/PX4/Devguide/master/build_scripts/ubuntu_sim_ros_melodic.sh
	bash ubuntu_sim_ros_melodic.sh

#### Download QGroundControl from:
	https://docs.qgroundcontrol.com/master/en/releases/daily_builds.html

#### Build Gazebo Sim
	cd ~/PX4-Autopilot
	make px4_sitl gazebo

#### Create px4 package
	cd ~/PX4-Autopilot
	DONT_RUN=1 make px4_sitl_default gazebo
	source Tools/setup_gazebo.bash $(pwd) $(pwd)/build/px4_sitl_default
	export ROS_PACKAGE_PATH=$ROS_PACKAGE_PATH:$(pwd)
	export ROS_PACKAGE_PATH=$ROS_PACKAGE_PATH:$(pwd)/Tools/sitl_gazebo
	roslaunch px4 posix_sitl.launch

## 3) Set up oscillation_ctrl
### Install xterm
	sudo apt-get update -y
	sudo apt-get install -y xterm

### Clone oscillation_ctrl
	cd ~/catkin_ws/src
	git clone https://git.spirirobotics.com/cesar.alejandro/oscillation_ctrl.git

### Add files to _tools/sitl_gazebo_
copy (or add) files in _oscillation_ctrl/models_ and _oscillation_ctrl/worlds_ to _PX4-Autopilot/Tools/sitl_gazebo/models_ and _PX4-Autopilot/Tools/sitl_gazebo/worlds_ respectively

	cp -R ~/catkin_ws/src/oscillation_ctrl/models/* ~/PX4-Autopilot/Tools/sitl_gazebo/models
	cp -R ~/catkin_ws/src/oscillation_ctrl/worlds/* ~/PX4-Autopilot/Tools/sitl_gazebo/worlds

### Add files to _ROMFS/px4mu_common
copy (or add) files in _oscillation_ctrl/airframes_ to _PX4-Autopilot/ROMFS/px4fmu_common/init.d-posix/airframes_

	cp -R ~/catkin_ws/src/oscillation_ctrl/airframes/* PX4-Autopilot/ROMFS/px4fmu_common/init.d-posix/airframes

add model names to _CmakeLists.txt_ in same 'airframe' folder (with number... 4000_spiri and 4001_spiri_with_tether)

add airframe name in _~/PX4-Autopilot/platforms/posix/cmake/sitl_target.cmake_ (no number!)
### Add necessary launch files
this should not be a necessary step and will be changed in the future

copy (or add) files from px4_launch directory to '~/PX4-Autopilot/launch'

	cp -R ~/catkin_ws/src/oscillation_ctrl/px4_launch/* ~/PX4-Autopilot/launch
#### Change devel/setup.bash
In catkin_ws (or any working directory) add to devel/setup.bash:

	CURRENT_DIR=$(pwd)
	cd ~/PX4-Autopilot
	source Tools/setup_gazebo.bash $(pwd) $(pwd)/build/px4_sitl_default
	export ROS_PACKAGE_PATH=$ROS_PACKAGE_PATH:$(pwd)
	export ROS_PACKAGE_PATH=$ROS_PACKAGE_PATH:$(pwd)/Tools/sitl_gazebo

	cd $CURRENT_DIR

## JINJA TETHER FILE
- _spiri_with_tether.sdf.jinja_ can be altered to create desired tether model
- changes need to be made in px4 directory and will only take effect after running: "make px4_sitl gazebo"
	- can do "DONT_RUN=1 make px4_sitl gazebo" to avoid starting px4 shell and gazebo
- First two elements can be changed to tweak tether parameters
	- number_elements: number of segments tether will be composed of
	- tl: segment length (should be no shorter than 0.3 meters)
- __IMPORTANT:__ in order for jinja file to work, the following needs to be added to the _CMakeLists.txt_ (Ln 288 - may change in future) in the _~/PX4-Autopilot/Tools/sitl_gazebo_ folder:


		COMMAND
			${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/scripts/jinja_gen.py models/spiri_with_tether/spiri_with_tether.sdf} ${CMAKE_CURRENT_SOURCE_DIR}
		DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/scripts/jinja_gen.py models/spiri_with_tether/spiri_with_tether.sdf
		VERBATIM
		)
- This should be added in the line below VERBATIM in the add_custom_command function, which should be Ln 288)


## ROS NODES
### LinkState.py 
determines payload load angles and their rates (theta and phi) using Gazebo (needs to be made more robust), as well as determines tether length and keeps track of variables needed in case of a step or square test.

__Publishes to__:

	/status/twoBody_status # no longer used but keeps track of payload and vehicle position
	/status/load_angles    # payload angles (and states) relative to vehicle
	/status/path_follow    # boolean to run trajectory test
__Subscribes to__:

	none

### wpoint_tracker.py 
Sets original waypoints to be (in meters): [x=0,y=0,z=5]. This node listens to topic to keep track of desired waypoints. If any other node wants to change the waypoints, they publish to "reference/waypoints" and wpoint_tracker creates these new waypoints.

__Publishes to__:

	none
__Subscribes to__:

	/reference/waypoints 

### ref_signalGen.py
takes in desired position (xd) and determines smooth path trajectory.

__Publishes to__:

	/reference/path         # smooth path

__Subscribes to__:

	/status/load_angles
	/mavros/local_position/pose
	/mavros/local_position/velocity_body
	/mavros/imu/data
	/mavros/state
	/reference/waypoints

### klausen_control.py
determines forces on drone needed based on smooth path and feedback to dampen oscillations. From the forces needed, it publishes attitude commands.

__Publishes to__:

	/command/att_target # attitude commands

__Subscribes to__:

	/status/load_angles
	/reference/path
	/mavros/local_position/pose
	/mavros/local_position/velocity_body
	/mavros/imu/data

### path_follow.cpp
sets the vehicle in OFFBOARD mode (PX4) and takes off to a set height for 25 seconds before starting to publish attitude and thrust commands.

__Publishes to__:

	mavros/setpoint_position/local # needed to hover @ set height	
	mavros/setpoint_raw/attitude   # attitude and thrust commands 	

__Subscribes to__:

	/command/att_target	
	/command/bodyrate_command	
	/mavros/state

## Launching simulation
To launch a simulation, run the following command:

	roslaunch oscillation_ctrl oscillation_damp.launch

This simulation is set to have a Spiri Mu hover at an alitude of 5 m.
The launch file itself has two usable arguments:

__model:__

	spiri_with_tether           # Spiri Mu with a tethered payload			
	spiri                       # Spiri Mu without tethered paylaod
	headless_spiri_with_tether  # headless mode: launches with no Gazebo GUI. This is the default model
	
__test:__

	none   # default
	step   # step input - default is 5 m
	square # square trajectory
To run the simulation with a tethered payload headless mode and perform a step test:

	roslaunch oscillation_ctrl oscillation_damp.launch model:=headless_spiri_with_tether test:=step

Whenever the oscillation_ctrl is used, the scripts are written such that the vehicle will hover for about 30 seconds in "Position Mode". This is used to take advantage of the takeoff procedure PX4 has, as this controller assumes the vehicle is already in flight when determining the necessary thrust.