Python packages needed

sudo pip3 install torch torchvision

To make sure you have the latest version of colcon_core:

sudo pip3 uninstall colcon_core sudo pip3 install colcon_core

To get euler_from_quaternion:

sudo pip3 install transforms3d

To get cv_bridge:

https://github.com/ros-perception/vision_opencv/tree/humble/cv_bridge

Don't forget to source your ros2 env first e.g. . /opt/ros/humble/setup.bash

cd <YOUR_ROS2_WORKSPACE>/src
git clone https://github.com/ros-perception/vision_opencv.git -b <ROS2_DISTRO>
cd ..
colcon build --packages-select cv_bridge --symlink-install

To get JetCam:

git clone https://github.com/NVIDIA-AI-IOT/jetcam

OLD NOTES

1. Install Robomaker

contribution from Mr. Nigel Gardiner from AWS

curl -fsSL "http://bit.ly/robomaker" | sudo -E bash -

After about 10 minutes with ROS installed, you must reboot the system

sudo reboot

2. Setup aws_mouse

(only run once) ->aws management console -> cloud 9

git clone -b saveimage  https://github.com/lbaitemple/aws_mouse/ 
cd aws_mouse
bash ./updateos.sh

Only when you see any lock error, please try

sudo rm -r /var/lib/dpkg/lock*
sudo dpkg --configure -a
bash ./updateos.sh

compile ROS (only run once)

rosdep install --from-paths src --ignore-src -r -y

3. Run and program aws_mouse (when step 1 and step 2 are completed)

3a. run the following commands only when you changed the source code

colcon build
source install/setup.bash

3b.open virtual desktop

cd ~/environment/aws_mouse
export DISPLAY=:0
source install/setup.bash
roslaunch maze_demo explore_world.launch

3c.open one more terminal to control the robot move forward for 1.2 meters

cd ~/environment/aws_mouse
source install/setup.bash
rosrun maze_demo runner_template.py oforward 1.2

3c.open one more terminal to control the robot rotate left for 90 degrees (right rotation is negative degrees)

cd ~/environment/aws_mouse
source install/setup.bash
rosrun maze_demo runner_template.py turn 90

4. see what camera sees

open another terminal

source install/setup.bash
export DISPLAY=:0
rqt_image_view 

You can stop here ....

The following is for more advanced AI usages.

if load an empty world with flowers, you can use

roslaunch maze_demo explore_world.launch worldfile:=empty_flower.world x:=0 y:=0

running the robot. [node_follow_wall2.py has the cmd_vel topic to move the robot]

open another terminal

cd ~/environment/aws_mouse
source install/setup.bash
rosrun maze_demo test_runner.py forward 0.1
rosrun maze_demo test_runner.py turn 90

or choose a different position or pose

roslaunch maze_demo explore_world.launch x:=-0.75 y:=-2.25 Y:=0.00

create a folder with label (ex. label = sunflower, daisy and background)

make sure you create a background folder first

rosrun img_recognition mkdir.py -n background

Then, you can create other folders based on target image labels

rosrun img_recognition mkdir.py -n rose
rosrun img_recognition mkdir.py -n sunflower
rosrun img_recognition mkdir.py -n daisy

remove a folder

rosrun img_recognition rmdir.py --name rose

open one more terminal to save images for training

cd ~/environment/aws_mouse
source install/setup.bash
roslaunch img_recognition save_rosimage.launch

open one more terminal to make the default label

rosservice call /save_image/select_label background

or make the label for rose

rosservice call /save_image/select_label rose

add image to an labelled folder

cd ~/environment/aws_mouse
source install/setup.bash
rosservice call /save_image/save_image_action  true

You can stop the collection

rosservice call /save_image/save_image_action  false

make another target label sunflower

rosservice call /save_image/select_label sunflower

make another target label rose

rosservice call /save_image/select_label rose

add image to an labelled folder

cd ~/environment/aws_mouse
source install/setup.bash
rosservice call /save_image/save_image_action  true

You can stop the collection

rosservice call /save_image/save_image_action  false

Backup configuration

bash ./saveconfig.sh backup

Restore configuration

bash ./saveconfig.sh restore

get a deep learning trained model

cd 
cd ~/environment/aws_mouse
source install/setup.bash
roslaunch img_recognition train_rosmodel.launch

running deep learning inferencing

open another terminal

to run the classifier using deep learning neural network

cd ~/environment/aws_mouse
source install/setup.bash
roslaunch img_recognition infer.launch 

open another terminal

to run the following command to get prediction/inference

cd ~/environment/aws_mouse
source install/setup.bash
rostopic echo -n1 /prediction
rostopic echo -n1 /inference 

if you want to show laser scan (help to see if the robot hits the wall)

roslaunch maze_demo explore_world.launch laser_visualize:=true 

if you want to show camera scan (may be combined with laser visualization)

roslaunch maze_demo explore_world.launch camera_visualize:=true