Decentralized Deep Reinforcement Learning for a Distributed and Adaptive Locomotion Controller of a Hexapod Robot.
Malte SchillingKai KonenFrank W. OhlTimo KorthalsPublished in: CoRR (2020)
Keyphrases
- legged robots
- reinforcement learning
- multi agent
- distributed systems
- central pattern generator
- real robot
- mobile robot
- robot control
- cooperative
- quadruped robot
- peer to peer
- inverted pendulum
- motion control
- distributed agents
- robotic systems
- adaptive control
- modular robots
- legged locomotion
- decentralized control
- structured peer to peer
- control architecture
- temporal difference
- multi robot coordination
- markov decision processes
- state space
- autonomous robots
- multi robot
- optimal control
- rough terrain
- perceptual aliasing
- control system
- action space
- real time
- human robot interaction
- contact force
- control scheme
- biped walking
- dynamic programming
- learning algorithm