Supervising vehicle using pattern recognition: Detecting unusual behavior using machine learning algorithms

Abstract

Our lives are becoming busier day by day. We are consequently forced to delegate important activities to other people. In developing countries, the middle class often have paid drivers pick up their children from schools. What if the driver decides to deviate from the usual route into a seedy part of town with the child? What if it speeds and is driving recklessly? What if it gets into an accident? Supervising our vehicles when we are not present in it, and being notified if anyone else using it for any unwanted/illegal intention is of paramount importance. Alarms are annoying, and we want to improvise the system in a smarter way for smarter monitoring. The proposed system is developed by applying Linear Regression models, kth-Nearest-Neighbor and Support Vector Machine classifier to identify a pattern and detect abnormal behavior of the vehicle.