Robot kinematics exercises

Preliminary

Download MLexercises repository and build its docker image mlnotebook.

Check mlnotebook image

docker image ls mlnotebook

Build robot gym docker image

./build.bash

Check mlrobotgymenvs image

docker image ls mlrobotgymenvs

Run the image

./run.bash

Run robot simulations

Run inside the container

python run.py [options]


    usage: run.py [-h] [-env ENV] [-steps STEPS] [-seed SEED] [--render] [--log]

    optional arguments:
      -h, --help    show this help message and exit
      -env ENV      environment [r2,r3,r5] (default: r2)
      -steps STEPS  Execution steps (default: 10,000)
      -seed SEED    Random seed (default: 1000)
      --render      Enable rendering
      --log         Enable data log

Press 'ESC' in the GUI or CTRL-C in the terminal to exit the simulation

Example:

python run.py -env r2 -seed 1000 -steps 100000 --render

Environments:

r2: 2D robot with 2 joints
r3: 2D robot with 3 joints
r5: 3D robot with 5 joints

Log data

python run.py -env r2 -seed 1000 -steps 100000 --log > logfile.csv

Colab local runtime

Run the server from the container

jupyter notebook --ip=0.0.0.0 --no-browser --NotebookApp.allow_remote_access=True --allow-root --NotebookApp.allow_origin='https://colab.research.google.com' --port=8888 --NotebookApp.port_retries=0

Npte: in same cases, the connection gets stuck. Just retry!