The overall objective of this project was to develop an automated early-warning system that uses low-frequency microphones and on-board processing of acoustic signals, that can be used to detect elephant presence and movement and send alerts to the relevant authorities and the local communities.

Our aims of the project were:

1. Design and develop the Elephant – Automated Infrasonic Device (Elephant-AID): Our device would use a low-frequency microphone with a wide frequency response (10 Hz - 15 kHz). The signal would then be processed on-board to detect elephants using a machine-learning approach. Alerts will then be sent as text messages, detailing the elephant location and probability of a threat. In areas with limited connectivity, LoRa devices will be used to connect to the nearest cellular network.

2. Test Elephant-AID’s efficiency: The efficiency of the system will be tested in both controlled and real-life on-ground situations by a) quantifying accuracy (false positive and false negative detections), and b) measuring the rate of human-elephant conflict in deployment and control areas. While reducing false negatives is our immediate goal, lowering ‘false alarms’ is crucial for the long-term success of the project.