iBUG: AI Enabled IoT Sensing Platform for Real-time Environmental Monitoring

Microcontroller-based smart devices and sensing systems have exploded in popularity in recent years, owing to the growing adoption of Internet of Things (IoT) platforms. Due to the widespread deployment of environmental sensing technology and digitization, data creation has surged to never-before-seen levels. However, it is computationally expensive to transport all these data to the cloud systems for decision making specially for battery operated sensor nodes. As a result, edge AI has arisen as a viable alternative to traditional AI, capable of making simple decisions without the assistance of any cloud system, making it more energy efficient and reducing the requirement for continuous data transfer. A single-board IoT platform capable of performing slimmer version of machine learning with multiple wireless protocols have become increasingly popular for this purpose. In this paper we demonstrate an energy efficient IoT platform (iBUG) that already has numerous built-in sensors. We also present an algorithm utilizing TensorFlow lite to accurately predict CO
2
using only gas resistance.