DA-SLAM: Deep Active SLAM based on Deep Reinforcement Learning.
Martin AlcaldeMatias FerreiraPablo GonzálezFederico AndradeGonzalo TejeraPublished in: LARS/SBR/WRE (2022)
Keyphrases
- simultaneous localization and mapping
- reinforcement learning
- mobile robot
- indoor environments
- mobile robotics
- state space
- particle filter
- real world
- visual slam
- computer vision
- dynamic environments
- object and scene recognition
- neural network
- deep learning
- real robot
- reinforcement learning algorithms
- action selection
- data association
- markov decision processes
- augmented reality
- learning algorithm
- machine learning