ME5150_omnibase

On your PC

1. Open miniconda console.
2. Pick a folder to save this repo, from now we'll call it BASE_PATH.
3. Clone this repository.

cd BASE_PATH
git clone https://github.com/uwulises/ME5150_omnibase.git

4. Create virtual environment named "omnibase".

cd client
conda create -n omnibase python=3.9
conda activate omnibase

5. Install all required packages. This may take a while.

cd client
pip install -r requirements.txt

Now you just need to go to VSCode and execute some file, for example: streamClient.py or controlClient.py

On Raspberry Pi 4

1. Open a Command Prompt, like Windows PowerShell.
2. Connect to Rpi4 via ssh.

ssh robotica@{ROBOT NAME}.local

{ROBOT NAME} can be omni1, omni2 or omni3. If it doesn't work, then use the IP address 192.168.100.XXX (ask for it!) of your mobile base.
Ask the password.

ssh robotica@{IP ADDRESS}

3. Now that you're on Raspberry Pi, use Ubuntu commands like cd or ls to navigate.
4. To initialize the stream and control the robot, execute:

cd ME5150_omnibase
./start_video.sh
./start_control.sh

And it's ready.

Server configuration (rpi4)

1. Open a Command Prompt, like Windows PowerShell.
2. Connect to Rpi4 via ssh.

ssh robotica@{ROBOT NAME}.local

{ROBOT NAME} can be omni1, omni2 or omni3. If it doesn't work, then use the IP address 192.168.100.XXX of the mobile base, use a program like Angry IP Scanner to search. Ask the password.

ssh robotica@{IP ADDRESS}

3. Configura la Raspberry Pi. Preguntar por configuracioneees.

sudo raspi-config

4. Update packages. This may take a while.

sudo apt update
sudo apt upgrade -y 

5. Install Git.

sudo apt install git

6. Clone this repository.

git clone https://github.com/uwulises/ME5150_omnibase.git

7. Execute install_server.sh to install all

cd ME5150_omnibase/server
chmod +x install_server.sh
./install_server.sh

8. It's ready!

---

---

TO DO LIST https://www.raspberrypi.com/documentation/computers/os.html#python-on-raspberry-pi

OpenCV

Following recommendation of Picamera 2 manual

sudo apt install -y python3-picamera2

sudo apt install -y python3-opencv

sudo apt install -y opencv-data

sudo apt install screen

Python environment

• Install via apt-get install python3-venv
• This should starts with python 3.9.2 $ cd ME5150_omnibase

$ python -m venv robotica

Firmware

• Edit the cmd_vel(float vx, float vy) function, there's some errors on the rpm output
• Use the correct kinematic model using cmd_vel

Image Server

• Change the implementation to aiortc WebRTC
• Improve the framerate output

Control Server

• Port all the control moves to cmd_vel standard
• Use requests GET method for call moves

Xbox generic Control

• handle the names of the controller when the batterry runs low
• change the values output of the controller when the batterry runs low
• Look for another library which works with generic controls

ImportError when running stream_rpi.py

1. ImportError: libsrtp2.so.1: cannot open shared object file: No such file or directory https://linuxhint.com/install-ffmpeg-raspberry-pi/

sudo apt update && sudo apt upgrade -y
sudo apt install ffmpeg -y

Ensure that ffmpeg is successfully installed, running:

ffmpeg -version

2. ImportError: libsrtp2.so.1: cannot open shared object file: No such file or directory

sudo apt install libnspr4 libnss3 libsrtp2-1