GSTAgri's project aims to address challenges in efficient agriculture asset management by developing an on-premises device utilizing edge computing and AI. In collaboration with GlobalStar Inc., the system focuses on monitoring crop risks, specifically predicting drought indices using metrics from Arduino sensors. The agricultural industry's need for informed decisions based on key metrics drives this initiative. The project involves capturing Bluetooth sensor data, processing it through edge devices with predictive algorithms, and presenting concise outcomes. The design process encompasses understanding microcontroller architectures, inter-device communication, and neural network exploration. The embedded design processes sensor data locally, reducing transmitted data and costs. The machine learning model, trained on a diverse dataset, employs LSTM layers for time series forecasting. The model predicts drought likelihood and severity. The user interface visualizes drought indices and relevant metrics. When all these components come together the system achieves efficient on-premises monitoring of crop risks.

Team Members

Nidhi Dubagunta
Mister Gardener
Evan Rosenthal
Ayush RoyChowdhury
Connie Wang