INSTALLATION GUIDE

This document is for recording the actions done to integrate all the required systems (Husky, Outster LiDAR, FAST-LIO 2.0, AEDE, and TARE Planner) together on ROS Noetic.

📘 All packages should be added to the catkin_ws/src directory

1.1 Husky

Source code: https://github.com/husky/husky/tree/noetic-devel

Brief explanation of packages

• husky_bringup: installation and integration package
• husky_base: driver
• husky_gazebo: simulator bringup using Gazebo GUI
• husky-viz: vizualization configuration
• husky_control: controller configuration
• husky_description: URDF (Unified Robotics Description Format) description
• husky_msgs: messages (HuskyStatus exposes the status info)
• husky_navigation: autonomous mapping & navigation demos

Bolded packages are the important ones used for operating the robot (TBV)

1.2 Ouster

Installations:

git submodule add https://github.com/ouster-lidar/ouster-ros.git
source /opt/ros/noetic/setup.bash
cd ../.. # back to /catkin_ws
catkin_make --cmake-args -DCMAKE_BUILD_TYPE=Release

Running:

• First run source devel/setup.bash (ensure that the it's the devel in the catkin_ws directory)
• LiDAR must be connected then run:

  roslaunch ouster_ros sensor.launch  \
  sensor_hostname:=os1-122011000244.local
  metadata:=<json file name>          #<path to rosbag file>  optional
  

• If bag file is being used instead

  roslaunch ouster_ros replay.launch
  bag_file:=<path to rosbag file>
  metadata:=<json file name>          # optional if bag file already contains metadata topic
  

• to see the topics being published, open another terminal, source the same bash file then run rostopic list

1.3 Fast-LIO2.0

Installations:

1. sudo apt install libpcl-dev (this is for Point Cloud Library)
2. sudo apt install libeigen3-dev
3. sudo apt install libeigen3-dev
   • Alternatively, download tar.gz file from Eigen source
   • Untar using tar -xzf <tarfile name>
   • See instructions in INSTALL.txt in the untarred folder
4. Intall livox_ros driver (TO-DO: See if this part can be removed from the source code)
   • First install Livox SDK and follow installation instructions in README.md
     • This does not have to be in the main repo but needs to be done on the OS used
   • Next, add livox driver to our catkin_ws/src directory using git submodule add https://github.com/Livox-SDK/livox_ros_driver.git ws_livox/src
   • Run the following

     cd ws_livox
     catkin_make
     

5. Go back to catkin_ws/src
   • Add Fast-LIO package using git submodule add git clone https://github.com/hku-mars/FAST_LIO.git
   • do the following:

     cd FAST_LIO
     git submodule update --init
     

   • If running FAST-LIO separately only with LiDAR, continue with the following:

     cd ../..
     source ../ws_livox/devel/setup.sh
     catkin_make
     

   • Else, running source devel/setup.bash in our catkin workspace, followed by catkin_make will automatically build the required files accordingly too so the previous step will be unnecessary.

Running

source devel/setup.bash
roslaunch fast_lio mapping_ouster64.launch

1.4 Autonomous Exploration Development Environment (AEDE)

Clone source code:

git clone https://github.com/HongbiaoZ/autonomous_exploration_development_environment.git
git checkout distribution

Change topic names in loam_interface.launch to match the topic names published by the LiDAR's Odometry data and Fast-LIO's PointCloud data:

• set stateEstimationTopic to /Odometry
• set registeredScanTopic to /cloud_registered

Running

When running on the Husky

source devel/setup.sh
roslaunch vehicle_simulator system_real_robot.launch

1.5 TARE Planner

Clone source code:

git clone https://github.com/caochao39/tare_planner.git

Change topic names in loam_interface.launch to match the topic names published by the LiDAR's Odometry data and Fast-LIO's PointCloud data:

• set stateEstimationTopic to /Odometry
• set registeredScanTopic to /cloud_registered

Running

When running on the Husky

source devel/setup.sh
roslaunch tare_planner explore_husky.launch

2 Integrated Launcher

• Run catkin_create_pkg startup roscpp to make a new startup package
• In the /launch folder, add autonomous_husky_startup.launch
  • This launch file will call all the other launch files required, namely: husky_base (which also calls husky_control and husky_teleop internally), ouster-ros's sensor.launch, fast-lio's mapping_ouster64.launch, AEDE's system_real_robot.launch
• In the /src folder, add twist_unstamp.cpp
  • This is to extract the Twist data and publish it to the /cmd_vel topic to be used for controlling the husky

3 Operating the Autonomous Husky

Semi-automated setup for deployment

1. One time setup (refer to bash_aliases.txt)

   • Add aliases to ~/.bashrc in PC:

   alias cd_pc='cd ~/Documents/autonomous_husky/catkin_ws'
   alias source_husky_pc='source ~/Documents/autonomous_husky/catkin_ws/devel/setup.bash'
   

   • Add aliases to ~/.bashrc in NUC:

   alias cd_nuc='cd ~/kaijuntay/autonomous_husky/catkin_ws'
   alias source_husky_nuc='source ~/kaijuntay/autonomous_husky/catkin_ws/devel/setup.bash'
   

   • source ~/.bashrc in either console to use the alias commands for easier life when testing

2. On the NUC:

   source setup_nuc.sh
   python3 preflight.py
   cd catkin_ws
   source devel/setup.bash   #or source_husky_nuc
   

   After running the above, preflight should show all tests passed and also display the IP address of the NUC to be used in the next step.

3. On the PC:

   source setup_this.sh <NUC_IP> # this sets up the ROS_IP and ROS_MASTER_URI
   cd catkin_ws
   source devel/setup.bash
   

4. To run everything:

   • In the ssh terminal:

     roslaunch startup autonomous_husky_startup.launch
     

     All the nodes needed should be launched

     • If TARE boundary is to be used, add argument useBoundary:=true
     • If a specific boundary is to be used for TARE, add argument tareBoundary:=<BOUNDARY_FILE_NAME>

       # Example
       roslaunch startup autonomous_husky_startup.launch useBoundary:=true tareBoundary:=<BOUNDARY_FILE_NAME>
       

     Note that either the e_stop should be activated or the bumper of the controller should be held on to otherwise robot will start moving automatically.

   • In the main PC:

     roslaunch startup visualize.launch
     

Note: Since /cmd_vel has a lower priority than the joy_teleop/cmd_vel input from the controller, simply pressing the left bumper on the controller when AEDE is manoeuvring the robot will override the AEDE's control