Machine Learning Based Predictive Model for Data Fusion Based Intruder Alert System

Abstract

A Linear Regression and Support Vector Machine based algorithm is proposed with the two people detection sensors placed at the entry point of a building. The two sensors are ultrasonic sensor and microphone sensor. A Linear Regression model and SVM model is created for this and based on a simple top-down calculation done of the output of them both, the decision is taken for intruder identification. Arrays of ultrasonic sensors and microphone sensors are employed for aggregated output.