Integrating Multiple Policies for Person-Following Robot Training Using Deep Reinforcement Learning.
Chandra Kusuma DewaJun MiuraPublished in: IEEE Access (2021)
Keyphrases
- integrating multiple
- reinforcement learning
- optimal policy
- mobile robot
- real robot
- robot control
- policy search
- human robot interaction
- state space
- training set
- supervised learning
- reward function
- robot navigation
- autonomous learning
- path planning
- control policies
- autonomous robots
- motor learning
- deep architectures
- perceptual aliasing
- motor skills
- action selection
- markov decision process
- human beings
- learning algorithm
- continuous state
- training examples
- learning process
- reinforcement learning algorithms
- partially observable markov decision processes
- neural network
- deep learning
- model free
- markov decision problems
- position and orientation
- humanoid robot
- training samples
- hierarchical reinforcement learning