Cloud-LoRa: Enabling Cloud Radio Access LoRa Networks Using Reinforcement Learning...
Talkat 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI '24) · 2024-04-16..04-18·
summary
Cloud-LoRa builds a practical CRAN for LoRa that compresses IQ samples with a reinforcement-learning controller (PPO) over a DWT compressor, streams from a USRP B200 over cellular to Microsoft Azure, and adapts compression to measured TCP BBR bandwidth and latency. Deployed across 2.4×3 km farmland with eight nodes, it demonstrates real-time learning, achieves 6 dB SNR gain from joint decoding and over 2x throughput versus static lossy schemes, and exposes code on GitHub.
AI-written summary · about
chapters (12)
- 0:00Introduction to Cloud-LoRa
- 1:21Scalability in LPWAN networks
- 2:24Cloud Radio Access for LPWANs
- 4:11Backhaul bandwidth limits
- 5:12Cloud-LoRa architecture
- 7:19RL for dynamic compression
- 9:05RL agent state and actions
- 10:30Reward function design
- 11:33Cloud-LoRa pipeline
- 12:38Real-time deployment
- 13:20Evaluation results
- 15:51Conclusion
related talks
13mOne Protocol to Rule Them All: Wireless Network-on-Chip using Deep Reinforcement Learning
57mMachine Learning in the Age of IoT
11mLightning Talk: Outsourcing Telco Understanding AI
32mFine-tune LLMs from Kaggle Models using (Q)LoRA
10mDeep Learning in the Wild: Embedded PyTorch for...
11mA New Experimental Setup for Collecting nRF2.4 and LoRa Devices Signatures for ....
32mMan or Machine?: Developing a "Turing Test" for Radio Intelligence
15mWelcome to the IoT Casino