Skip to content

Use QLearning to train a robot to avoid black areas in the arena

Notifications You must be signed in to change notification settings

AndreaRettaroli/irs-q-learning-project

Repository files navigation

irs-q-learning-project

To run this project you need to use ARGoS To install ARGoS follow this steps

Ensure that your machine or virtual machine has all the necessary packages specified here.

In my case i used Ubuntu 20.04 LTS 64bit virtual machine.

To run train you need to create Qtable

    lua ./create_Q-table.lua Qtable-circuit.csv 256 3

This 3 params are specific for this code, 253 3 are Qtable dimensions, Qtable-circuit.csv is the filename, if you change it you need to change it also in your code.

    ./train-script.sh -f ./circuit-learning.argos -e 100

params e is epochs number

    ./test-script.sh -a ./circuit-testing.argos -s test-results.txt -e 100

test-results.txt is the name of the output file of results. params e is epochs number.

About

Use QLearning to train a robot to avoid black areas in the arena

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published