Building the NEVOD DIY Acoustic Drone Alert Network
The NEVOD project offers users the opportunity to build their own decentralized network of acoustic nodes for early detection of unmanned aerial vehicles (UAVs), commonly known as drones. This initiative aims to establish a system of collective security, enabling specific areas to “hear” the sky and respond promptly to potential threats. Unlike isolated sensors, nodes integrated into a network provide district-wide alerts, significantly enhancing monitoring effectiveness.
NEVOD Architecture and Functionality
Each NEVOD node is an autonomous device assembled by the user. Once installed on a pole or rooftop, it begins listening to a defined sector. Detected events are then transmitted to the user’s Home Assistant system. Key features of the system include:
- Local Processing: Nodes independently detect three classes of UAVs, determine their direction, and send data (event, azimuth, spectrogram) to the collective network. Raw audio and precise node coordinates are not transmitted externally, ensuring privacy.
- No Cloud Services: The system operates without the need for centralized cloud platforms.
- WebUI and MQTT: Nodes are equipped with a web interface and support the MQTT protocol for seamless integration with smart home systems.
Technical Specifications and Assembly
The core of a NEVOD node is built around a Seeed XIAO ESP32-S3 microcontroller and a ReSpeaker XVF3800 four-microphone array. The estimated cost for the core components is approximately 10,000 rubles. STL files for the node’s enclosure are provided in the project’s documentation (ENCLOSURE.md). A detailed list of all necessary components can be found in BOM.md.
The project offers two primary firmware options:
- Open Mic (MIT): This open-source firmware (version v0.3.0), available in the repository, provides Opus audio streaming via RTSP and features a WebUI. It is intended for board testing, general listening, and can be integrated with services like BirdNET.
- NEVOD DIY: A specialized firmware for the sensor, which includes a signed binary file, on-board detection capabilities, MQTT/HA support, and over-the-air (OTA) updates via the diy-ota channel.
The assembly process, detailed in FLASH_NEVOD_DIY.md, involves ordering components, printing the enclosure, flashing the “Open Mic” firmware for initial testing, then installing the NEVOD DIY firmware, and configuring MQTT in Home Assistant. Once local operation is stable, the node can be connected to the collective network, sharing event and azimuth information. The effectiveness of acoustic detection is influenced by external factors such as wind, noise levels, and installation height, meaning a single node cannot completely cover an entire neighborhood.
The NEVOD project presents an intriguing decentralized approach to UAV detection, leveraging acoustic signatures. The choice of the Seeed XIAO ESP32-S3 with a ReSpeaker XVF3800 four-microphone array is a cost-effective yet capable hardware foundation for directional acoustic sensing. While local processing and privacy are commendable, the inherent limitations of acoustic detection, particularly in urban environments with high ambient noise or against stealthier drone designs, will likely necessitate integration with RF-based detection for comprehensive coverage. The collective network model, however, offers scalability and redundancy that single-point solutions lack, potentially enhancing overall SNR for threat identification.