1. Name of applicant(s), Name of organisation(s)
Aude Vuilliomenet, Duncan Wilson (Connected Environments Lab, Centre for Advanced Spatial Analysis (CASA), University College London)
Ella Browning (UCL Centre for Biodiversity and Environment Research (CBER), Bat Conservation Trust)
2. Email address
aude.vuilliomenet.18@ucl.ac.uk
3. Application Track
New Project Track (complete questions 1-13)
4. Project Description (max. 2 sentences)
The project “Shazam for Bats” was a proprietary smart bat monitor that was developed in 2017 by Intel and UCL. It used an Intel Edison with a Dodotronic 192K microphone to record and process the soundscape using a CNN detection algorithm to count bat calls which were uploaded in real-time to a cloud database.
5. Project Goals & Project Steps
The project aims to improve the technical capability of the first prototype of “Shazam for Bats” as well as establish it as an open-source hardware device. More specifically, the project will follow the below-described steps in order to achieve its goals:
- Redesign the bat monitor with open science hardware components (RaspberryPi).
- Implement an AudioMoth USB to provide a lower cost device to capture the soundscape (in addition to the Dodotronic microphone).
- Extend data transmission capability to LoRa.
- Document the hardware components and firmware, draft the operation manual.
- Improve the CNN detection algorithm so that it not only identifies bat calls but also bird species. Test the accuracy of the deep learning algorithm when deployed on the device.
6. People
It is expected that 5 to 6 people will contribute to the project.
- Project Lead: Aude Vuilliomenet (PhD Student working on hardware design, algorithm, documentation)
- Engineering: Duncan Wilson, Steven Gray + 1 MSc student from MSc Connected Environment (hardware, enclosure design, connectivity, datastore, algorithm development and documentation)
- Field Testing & Data Sharing: Ella Browning (Bat Conservation Trust - New data will be provided by Ella via BCT to improve the Bat detection algorithm.)
- Advisors: Duncan Wilson (Professor of Connected Environments, Bartlett), Kate Jones (Professor of Ecology, Centre for Biodiversity and Environmental Research), Oisin Mac Aodha (Edinburgh University detection algorithms)
7. Organisation & Collaboration
Funds will be managed through the Connected Environments Lab at CASA UCL. The project steps will be documented and blog posts of the different phases will be published on the blog of the CE Lab as well as shared via social media. All documents (hardware design, bill of materials, algorithms) will be published on GitHub.
8. Underrepresented Background
The people involved in this project form a group that represents gender equally and covers different geographic regions.
9. Current Infrastructure & Resources
The project is supported by and has access to the following:
- Maker spaces - within the CE Lab as well as The Bartlett faculty.
- WiFi and LoRa Network coverage within the QEOP - a testbed for the current prototype.
10. Budget
| Cost Category | Details | Estimated Costs | Comments |
|---|---|---|---|
| Materials | Hardware (RPi, Audiomoth, LoRa Module), Enclosures (waterproof filaments for 3D printer), Power Sources (Test with solar panel, battery pack) | $1000 | Materials needed to develop the version 2 |
| Subcontract | Support from experts in electronics, hardware design | $600 | Get advice to make v2 as cheap and reproducible as possible. Costs vs. Benefits of developing a PCB. |
| Other | Shipping Costs, Documentation, Videos Filming | $400 |
11. Project Outcomes & Documentation
The development of the V2 will require changes in the design of the hardware and integration of a second classification algorithm to identify birds (using the open-source BirdNet algorithm). The project will release the documentation on GitHub. Examples of documentation files are hardware design and schematic, bills of materials, setup-guide, configuration file, model training and outputs.
12. Addressing Diversity & Inclusion (see. GOSH’s Values)
This project contributes to open access to environmental monitoring and automated data processing. Specific attention will be on developing a low-cost prototype that is easy to replicate and can be used by non-technical communities. It engages with the GOSH’s values in the following way:
- Will automate the processing and analysis of bats and birds acoustic recording, thus benefiting ecologists and/or environmental agencies.
- Will act as a case study for open-sourcing hardware and software in the academic context. It aims to create video and short blog posts of the steps and processes to create OSH.
- Will democratise environmental monitoring tools and allow better investigation into how urban development and landscape management influence the ecological biodiversity of studied sites.
13. Conflicts of Interest
None!