Deep Reinforcement Learning Control Theory Research Project at Advanced Robotics Lab in the University of Edinburgh, Scotland, UK. Thanks to advisor Dr. Zhibin (Alex) Li, mentors Chuanyu Yang and Kai Yuan. This project is an extension of Chuanyu's research on Deep Reinforcement Learning for Humanoid Robotic Control Balance.
Run bash DDPGtrials.sh
. Make sure you have shell/bash version 4. Raw data is stored in folder allData
. Consolidated data stored in folder finalData
.
Analyzation scripts in folder "analyzeChuanyuData". plottingResults.m analyzes packaged data from single force simulation, analysisForces.m analyzes packaged data from all forces.
Changed run_ddpg.py
to take in optional argument -f [force]
. Default force is 400N. Changed logger.py
to log forces.