WildGuard is an edge AI system designed to combat wildlife poaching by detecting suspicious activities in real-time using the PSoC™ 6 AI Evaluation Kit. It leverages the DEEPCRAFT™ Siren Detection Ready Model to identify poaching-related sounds (e.g., gunshots, vehicle engines) via the onboard MEMS microphone, supplemented by IMU-based movement detection and barometric pressure monitoring for environmental changes. Deployed in remote wildlife reserves, WildGuard processes data on-device and sends instant Wi-Fi/BLE alerts to rangers, enabling rapid response to protect endangered species.
Why did I decide to make it?I was inspired to create WildGuard by the global poaching crisis, which saw over 100,000 elephants killed in Africa between 2010 and 2012, threatening biodiversity. Traditional monitoring systems like satellites or drones are costly and impractical for vast reserves. I aimed to build an affordable, low-power, edge-based solution using the PSoC™ 6 AI Kit to empower conservationists with real-time, actionable insights, addressing a critical environmental challenge with cutting-edge AI technology.
How does it work?WildGuard uses the PSoC™ 6 AI Kit’s MEMS microphone to capture ambient sounds, processed by the DEEPCRAFT™ Siren Detection Ready Model to detect gunshots or vehicle engines. The IMU (accelerometer, gyroscope, magnetometer) monitors movement patterns to identify intruders, while the barometric pressure sensor tracks environmental changes. All data is processed on the PSoC™ 6 MCU for low-latency edge inference. Upon detecting suspicious activity, the system sends alerts via Wi-Fi/BLE to rangers’ devices using ModusToolbox™ code, adapted from the sound recognition example. A 3.7V LiPo battery powers the system for remote deployment. Testing involved royalty-free audio (CC0-licensed gunshots/vehicles from freesound.org) and simulated movements, confirming reliable detection and alerts.
Comments