A Lightweight Deep Recurrent Q-Learning Technique for Autonomous Wildfire Surveillance.
Jeremy CantorO. Patrick KreidlJohn NuszkowskiAlan HarrisAyan DuttaPublished in: ICMLA (2023)
Keyphrases
- lightweight
- cooperative
- unmanned aerial vehicles
- reinforcement learning
- learning algorithm
- function approximation
- multi agent
- state space
- recurrent neural networks
- surveillance system
- smoke detection
- search and rescue
- stochastic approximation
- model free
- optimal policy
- wireless sensor networks
- video surveillance
- feed forward
- reinforcement learning algorithms
- learning rate
- communication infrastructure
- neural network
- dos attacks
- low cost
- development environments