DWA-RL: Dynamically Feasible Deep Reinforcement Learning Policy for Robot Navigation among Mobile Obstacles.
Utsav PatelNithish K. Sanjeev KumarAdarsh Jagan SathyamoorthyDinesh ManochaPublished in: ICRA (2021)
Keyphrases
- robot navigation
- reinforcement learning
- policy search
- continuous state
- optimal policy
- rl algorithms
- actor critic
- markov decision process
- action selection
- autonomous mobile robot
- state space
- reinforcement learning algorithms
- control policy
- policy gradient
- partially observable
- reinforcement learning problems
- markov decision processes
- action space
- state and action spaces
- autonomous robots
- policy evaluation
- control policies
- function approximators
- markov decision problems
- function approximation
- partially observable markov decision processes
- continuous state spaces
- scene understanding
- state action
- policy iteration
- model free
- reward function
- learning algorithm
- partially observable domains
- temporal difference
- optimal control
- agent learns
- machine learning
- reinforcement learning methods
- landmark recognition
- average reward
- dynamic programming
- topological map
- infinite horizon
- long run
- d scene
- stochastic games
- exploration exploitation tradeoff
- initially unknown
- three dimensional
- inverse reinforcement learning
- mobile robot
- transfer learning
- finite state
- learning agent