Low-power machine learning architecture for ecological application on tinyMCU

Number of Students : 1
Guides : Chetan Singh Thakur

We propose a low powered (mW range) edge device capable machine learning system to detect the anomaly sound detection specially in the ecological application. We will design our system based on the data obtained from a microphone and create a sensor network using LoRA connectivity. The necessary data will be trained on contemporary Neural Network/Machine Learning algorithm and deployed on microcontroller. We will build a complete hardware prototype, which will be deployed on the forest for real-time detection of various poaching activities. The skillsets required in this project are embedded system, basic machine learning, hardware-software codesign.