Machine Learning using qlearning to simulate lidars on a car

Q-Learning Project

Project video: https://www.youtube.com/watch?v=ticezSoAzxQ

Dependencies

https://github.com/RobotLocomotion/director

Recommended setup

Add to ~/.bash_profile or ~/.bash_rc:

alias ddConsoleApp=$HOME/path-to/build/bin/ddConsoleApp

Recommended example usage

ddConsoleApp runPolicySearchTEST.py

How to run with iPython notebook

Find full path to ipython

which ipython

Run ddConsoleApp with full path to ipython, with notebook as argument

(for example:)

ddConsoleApp /usr/local/bin/ipython notebook

Run the cells to set properties of a simulation, and then sim.run()

Currently need to restart the kernel after each time starting a new simulation.

How to run simple demo

cd project/code/
ddConsoleApp movingIntersection.py

Alternatively, can press play in runcar.ipynb

Once visualizer opens:

• F8 to open the console
• timer.start()