real_world_instruction.txt

On Local:
"ssht" - ssh to turtlebot and enters the password automatically (wait for a few more minutes if "No route to host" shown)

Unplug the black ethernet cable & Plug in the orange ethernet cable and make sure this local computer is connected to Robotics Lab 5G!!!

*****************************************************************************
	      *** UNPLUG lidar wire before turning on robot ***
On Turtlebot:
0: Unplug the Lidar wire, turn on the top battery, turn on the base switch

1: "init"
# launch base
ssh tb2xavier@192.168.1.51
# password: turtlebot
# Unplug the lidar wire!!!!!!
source catkin_ws/devel/setup.bash && roslaunch turtlebot2i_bringup minimal.launch

2: "lidar" (sometimes "lidar" doesn't work, if so, use the full commands below)
# launch lidar
ssh tb2xavier@192.168.1.51
source catkin_ws/devel/setup.bash && sudo chmod 666 /dev/ttyUSB0 && sudo chmod 666 /dev/ttyUSB1 && sudo chmod 666 /dev/ttyUSB2 && roslaunch rplidar.launch

# password: turtlebot

3. (Not needed for now) Launch zed2 camera
cd ~/zed_catkin_ws && source devel/setup.bash && roslaunch zed_wrapper zed2.launch

4. (After finishing all steps below, before running test.py)
launch T265 camera (Need to re-launch it before every navigation trial!)
# Don't plug T265 into the USB hub, directly plug it into the jetson to avoid unstable connection!!
source ~/tb2_shuijing.bash && roslaunch low_res.launch

To kill everything:
"turtkill" - kills base, lidar, Realsense Camera

*****************************************************************************
On local:

t1. "laser_amcl"

# launch amcl navigation
# When "Enter map file",
# for workstation: leave as default
# for kitchen: /home/shuijing/csl_cafe_2024.yaml
# map_hallway can be changed by another constructed map
# wait until "odom received"
# if error, try:
# 1. kill the T265 if it's running
# 2. Restart everything
conda deactivate && source ~/tb2.bash && source ~/catkin_ws/devel/setup.bash && roslaunch turtlebot_navigation laser_amcl_demo.launch map_file:=$HOME/csl_entrance.yaml

# launch rviz
# calibrate the initial pose by "2d pose estimate"
# for convenience, move the robot to a "familar" area to better localize itself
# note to change the "fixed frame" to "map"
conda deactivate && source ~/tb2.bash && source ~/catkin_ws/devel/setup.bash && roslaunch turbot_rviz nav.launch


t2:
# localize the robot using AMCL
# Change the path of map in line 21 if needed
# wait until ros message is shown
conda deactivate && source ~/tb2.bash && source ~/catkin_ws/devel/setup.bash && cd ~/catkin_ws/src/2D_lidar_person_detection && python find_location.py

t3:
# Change the filename in line 29 if needed
# wait until "robot localized" is shown
conda deactivate && source ~/tb2.bash && source ~/catkin_ws/devel/setup.bash && cd ~/catkin_ws/src/2D_lidar_person_detection && python bg_removal_map.py

t4:
# launch people detector
# note to edit dr_spaam_ros/config/topics.yaml to make sure
# it subscirbes the /person_pts topic
conda deactivate && source ~/tb2.bash && source ~/catkin_ws/devel/setup.bash && source ~/virtual_envs/tb2/bin/activate && roslaunch dr_spaam_ros dr_spaam_ros.launch

######################## Only for debugging purpose ########################
Go back to t4, add /compare_map in rviz, change the turtlebot's pose so that the red & black scans are not too far from each other;
Click "2D pose estimation" to correct the robot pose,
then click "2D Nav Goal" (can be anything), the robo
t will move to align red & black scans


In rviz,
1. Change "Fixed Frame" to 'map'
2. Add 'by topic --> compare map --> image --> ok'
3. Click 2D Pose Estimate and have the error towards the location to be front.
4. click 2D navigation Goal and move around the turtlebot to match the localization again
5. Move to the initial position.

################################################

t5:
# Change the config parameters in /trained_policy_model/configs/config.py:
# Make sure env.env_name = 'rosTurtlebot2iEnv-v0', and sim2real configs are set as desired
# launch the main function
conda deactivate && source ~/tb2.bash && source ~/virtual_envs/tb2/bin/activate && source ~/Desktop/zed/zed_ws/devel/setup.bash && cd ~/Desktop/CrowdNav_sim2real/CrowdNav_sim2real && python3 test.py

*****************************************************************************
Misc:

# teleoperation
source ~/tb2.bash && source ~/catkin_ws/devel/setup.bash && roslaunch turtlebot_teleop keyboard_teleop.launch

if telop doesn't work in csl but works on xavier, run "sudo ufw disable" in csl

# change the subscribed topic of dr_spaam:
open ~/catkin_ws/src/2D_lidar_person_detection/dr_spaam_ros/config/topics.yaml
change the subscriber topic between: /person_pts (background map removed), /scan (original lidar scan)