Title:

Development and Integration of an AI-Enabled Temperature Sensor Array for Monitoring Freeze-Thaw Cycles in New Hampshire

**Abstract:** This study presents the development and integration of an AI-enabled temperature sensor array for monitoring freeze-thaw cycles (FTC) in New Hampshire. FTC, a crucial indicator of temperature fluctuation intensity, are monitored using low-cost sensors integrated with the iBUG platform—an edge node with embedded machine learning capabilities. The sensor array includes 20 DS18B20 Digital Temperature Sensors deployed at Thompson Farm, with data transmitted wirelessly via LoRa to the Cloud. Current findings include ongoing real-time data collection, accuracy validation against NOAA records, and initial correlations between temperature and frost depth using TinyML algorithms. The project aims to develop a linear regression model using TensorFlow Lite to predict depth from temperature, sensor number, address, and location. Applications include soil erosion studies, climate change research, land use planning, and energy efficiency optimization. This work highlights the potential of AI-driven sensor arrays for cost-effective and efficient environmental monitoring. Acknowledgements: The authors would like to acknowledge the support of Professor MD Shaad Mahmud and Dr. Alexandra Contosta for their guidance and expertise. Special thanks to Faishal Yousuf for his contributions. This research is ongoing, and the poster serves as a progress update. Keywords: Freeze-thaw cycles, temperature sensor array, AI-enabled monitoring, TinyML, TensorFlow Lite, environmental monitoring, iBUG platform

Name

Comment*

Comments are viewable only by submitter

Captcha*