LoRaDRL: A DRL-Based Approach for LoRa Optimization
Overview
A deep reinforcement-learning algorithm (DDQN) that selects PHY-layer transmission parameters in LoRaWAN networks. Existing rule-based algorithms cause packet collisions in dense LPWAN deployments; this DRL approach reduces collisions and improves Packet Delivery Ratio by up to 500% in some scenarios. Published at IEEE LCN 2020 with a follow-up paper.
Outcome
Demonstrated up to 500% PDR improvement over the best existing LoRaWAN parameter selection methods, with peer-reviewed publications in IEEE venues.
Architecture & Pipeline
-
01
LoRaWAN Sim Env
Custom Python simulator
- Python
-
02
Network State
PHY-layer conditions
-
03
DDQN Agent
TensorFlow
- DDQN
- TensorFlow
-
04
Action
PHY parameters
-
05
PDR Reward
Up to 500% improvement
-
06
Updated Policy
Production parameter selection
End-to-end flow, in order. Scroll sideways on a narrow screen.
Key Features
- Double Deep Q-Network (DDQN) algorithm
- Custom Python simulation environment for LoRaWAN
- Validated against the best published baselines
- Two peer-reviewed IEEE publications