This repo contains the codebase and toolset used in our paper Infrastructure-free, Deep Learned Urban Noise Monitoring at ∼100mW in ICCPS '22. Please cite our work as:

@article{yun2022infrastructure,
  title={Infrastructure-free, Deep Learned Urban Noise Monitoring at $$\backslash$sim $100 mW},
  author={Yun, Jihoon and Srivastava, Sangeeta and Roy, Dhrubojyoti and Stohs, Nathan and Mydlarz, Charlie and Salman, Mahin and Steers, Bea and Bello, Juan Pablo and Arora, Anish},
  journal={arXiv preprint arXiv:2203.06220},
  year={2022}
}

Overview

• sonyc_mkii_tools: A narrowly focused piece of software that sits on the USB-serial link on the base station (Pi). Used to de-mux and decode the bitstream from the base station and hand off data to higher level services. It can also interface with the bootloader to reprogram motes. Detailed steps on installation and usage given in sonyc_mkii_tools/README.

Additional Resources

• Trained embedding models: [TBD]
• SONYC-L3 embedding model and classifier training code: https://github.com/ksangeeta2429/embedding-approx
• eMote OS: [TBD]