Hierarchical Terrain-Aware Control for Quadrupedal Locomotion by Combining Deep Reinforcement Learning and Optimal Control.
Qingfeng YaoJilong WangDonglin WangShuyu YangHongyin ZhangYinuo WangZhengqing WuPublished in: IROS (2021)
Keyphrases
- optimal control
- legged robots
- reinforcement learning
- control problems
- feedback control
- inverted pendulum
- dynamic programming
- control strategy
- quadruped robot
- infinite horizon
- risk sensitive
- optimal control problems
- rough terrain
- state space
- class of nonlinear systems
- function approximation
- mobile robot
- actor critic
- control law
- markov decision processes
- robot control
- robotic systems
- real robot
- rl algorithms
- brownian motion
- lyapunov function
- linear quadratic
- real time
- continuous stirred tank reactor
- policy gradient
- reinforcement learning algorithms
- control system
- machine learning
- data mining