No Access
Bipedal Walking Through Reinforcement Learning
The original publication (Chew, C.-M. and Pratt, G. A., "Dynamic bipedal walking assisted by learning", Robotica, vol. 20, pp. 477–491.) is acknowledged. Copyright (2002) Cambridge University Press. Reprinted with permission.
https://doi.org/10.1142/9789812773142_0006Cited by:0 (Source: Crossref)
Abstract:
The following sections are included:
Introduction
The Bipedal Systems
Control Architecture
Key Implementation Tools
Virtual Model Control
Q-Learning
Q-Learning Algorithm Using Function Approximator for Q-Factors
Implementations
Virtual Model Control Implementation
Reinforcement Learning to Learn Key Swing Leg's Parameter
State variables
Reward function and reinforcement learning algorithm
Simulation Studies and Discussion
Effect of Local Speed Control on Learning Rate
Generality of Proposed Algorithm
Summary
Bibliography
