Enhanced Low-Dimensional Sensing Mapless Navigation of Terrestrial Mobile Robots Using Double Deep Reinforcement Learning Techniques.
Linda Dotto de MoraesVictor Augusto KichAlisson Henrique KollingJair Augusto BottegaRicardo Bedin GrandoAnselmo Rafael CuklaDaniel Fernando Tello GamarraPublished in: LARS/SBR/WRE (2023)
Keyphrases
- mobile robot
- low dimensional
- reinforcement learning
- obstacle avoidance
- indoor environments
- unstructured environments
- autonomous navigation
- sensor fusion
- unknown environments
- high dimensional
- robot control
- navigation tasks
- outdoor environments
- real robot
- potential field
- path planning
- dimensionality reduction
- dynamic environments
- manifold learning
- principal component analysis
- high dimensional data
- autonomous robots
- input space
- function approximation
- multi robot
- collision free
- feature space
- euclidean space
- autonomous vehicles
- motion control
- topological map
- motion planning
- machine learning
- model free
- robotic systems
- mobile robotics
- multi agent
- pattern recognition
- reinforcement learning algorithms
- data points
- markov decision processes
- optimal control
- recognizing facial expressions
- linear dimensionality reduction
- collision avoidance
- robot navigation
- optimal policy
- sensor networks
- training set
- learning algorithm
- temporal difference
- nonlinear dimensionality reduction
- action selection
- graph embedding
- kernel function
- multiple robots
- neural network
- data sets