Human Presence Detection with Thermal Sensor using Multilayer Perceptron Algorithm

The thermal sensor employing MEMS technology has enabled high accuracy, smaller footprint, low-noise, compact size, and contactless measurement in many embedded computing applications. In the Industrial Internet of Things and home automation system, the automation of different devices requires detecting human presence to initiate proper functioning. We have presented a low-profile IoT platform for realizing the human presence by using the Omron D6T-44L06 thermal sensor. The entire prototype has been built over Aurdino Nano. The experimentation has been conducted mainly in two phases—first, implementing the hardware for the dataset preparation and second, creating a model for Artificial Neural Network(ANN). The prototype built was first used for logging data in the presence and absence of humans using the thermal sensor, and then the model was trained using Keras library in Python. Finally, the realized and trained model has an efficiency of 99.6% and was further validated by experimental outlook.