Reinforcement Learning for Walking in Humanoids
Abstract :
Humanoid walking serves as an important challenge in the field of Robotics. This report presents the design and learning architecture for an omnidirectional walk used by a humanoid robot soccer agent acting in the RoboCup 3D simulation environment. The work explains the implementation of Reinforcement learning for improving the speed and stability of humanoid walking. Parameter optimization to achieve the above task was done using CMA-ES algorithm which acts a black box.
Team Members :
- Kumar Krishna Agrawal : 12MA20052
- Abhinav Agarwalla : 13MA20003
- Kumar Abhinav : 13CH30028
- Arnav Kumar Jain : 13MA20011
- Nishant Nikhil : 14MA20021
- Abhinav Agarwal : 14EE10001
- Shivam Vats : 13MA20039
- Ankush Chatterjee : 13MA20008
Working Environment :
Humanoid Nao :
Simulation Environment :
