Autonomous Docking of Mobile Robots by Reinforcement Learning Tackling the Sparse Reward Problem.
A. M. Burgueño-RomeroJ. R. Ruiz-SarmientoJavier González JiménezPublished in: IWANN (2) (2021)
Keyphrases
- reinforcement learning
- mobile robot
- autonomous navigation
- robotic systems
- autonomous learning
- function approximation
- path planning
- dynamic environments
- reinforcement learning algorithms
- obstacle avoidance
- eligibility traces
- real robot
- robot control
- state space
- mobile robotics
- markov decision processes
- agent learns
- indoor environments
- model free
- unknown environments
- learning algorithm
- reward function
- multi agent
- sparse data
- multi robot
- motion control
- optimal policy
- action selection
- partially observable environments
- machine learning
- robot behavior
- sparse representation
- temporal difference
- autonomous robots
- autonomous vehicles
- high dimensional
- search and rescue
- transfer learning
- compressive sensing
- motion planning
- initially unknown
- learning capabilities
- inverse reinforcement learning
- learning process
- state action
- average reward
- dynamic programming
- learning agent
- collision avoidance
- learning problems
- reinforcement learning methods
- sensory information
- partially observable markov decision processes
- reward shaping
- partially observable