Deep reinforcement learning collision avoidance using policy gradient optimisation and Q-learning.
Shady A. MagedBishoy H. MikhailPublished in: Int. J. Comput. Vis. Robotics (2020)
Keyphrases
- collision avoidance
- policy gradient
- reinforcement learning
- reinforcement learning algorithms
- actor critic
- function approximation
- path planning
- mobile robot
- dynamic environments
- state action
- reinforcement learning methods
- state space
- single agent
- model free
- model free reinforcement learning
- policy search
- optimal policy
- rl algorithms
- markov decision processes
- path finding
- temporal difference
- multi agent
- genetic algorithm
- control problems
- machine learning
- optimal control
- partially observable markov decision processes
- function approximators
- action selection
- dynamic programming
- temporal difference learning
- neural network
- learning algorithm
- learning problems
- gradient method
- average reward
- fuzzy neural network
- markov decision process
- fuzzy logic
- input output
- decision problems