-
Notifications
You must be signed in to change notification settings - Fork 0
My dissertation. Reinforcement learning program in Python used to discover optimal kicking motions for the Aldebaran Nao robot. Also includes a kicking module to be used by team Edinferno in the Robocup 2013.
spbwilson/edinferno-kicking-module
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
Author: Sean Wilson (s0831408) [email protected] =========================KICKING ENGINE======================= This folder hold all the code necessary to run the Python built kicking engine. See dissertation for further explination. INPUTS: distance - must be a number representing desired distance in cm displacment - the distance (cm) displacement (laterally) from the centre of the robot. (-left +right) angle - the angle in degrees for the desired kick (-left +right) ==================REINFORCEMENT LEARING CODE================== This hold all the required code to build a state-action environment for the nao robot. The user must specify the distance to optimise. There are two RL algorithms implemented: - Qlearning - Monte Carlo There is also a limit module to incrementally move nao's joints. Similarly there is a balance module to test for nao's balance point. A Plotting algorithm is included for making heat maps of the average state-values resulting from the RL algorithms We also include sample kicking motions for the benchmark used in the dissertation, and the Edinferno basic straight kick
About
My dissertation. Reinforcement learning program in Python used to discover optimal kicking motions for the Aldebaran Nao robot. Also includes a kicking module to be used by team Edinferno in the Robocup 2013.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published