AI-Driven Ultrasound Fencing

Human-wildlife conflict remains as an important challenge in agriculture, often resulting in significant crop losses and endangering both farmers and animals. This research presents an AI-driven, eco-friendly ultrasonic deterrent system designed to mitigate such conflicts effectively. The proposed system integrates an ESP8266 microcontroller, an external camera, ultrasonic sensors, and servo motors to detect and track wildlife in real-time. A fine-tuned YOLOv11 object detection model ensures precise identification of intruding animals, while a pan-tilt mechanism facilitates dynamic tracking. Distance sensors measure the proximity of detected animals, activating an adaptive ultrasonic frequency when they exceed a predefined threshold, thereby deterring further approach. All system events are logged in a PostgreSQL database to analyze movement patterns and deterrence efficiency. Additionally, instant alerts are sent to local authorities through SMS if boundary violations occur. The system is designed to be cost-effective, scalable, and suitable for remote deployment, enabling a promising solution for minimizing human-wildlife encounters while promoting biodiversity conservation. This approach aligns with smart farming initiatives and highlights the potential of artificial intelligence and IoT in sustainable agricultural practices.