Detection of an Intruder and Protecting Automated Teller Machine using Seismic Sensor

Abstract

Main motive of this proposed work is to develop an accurate security system for the Automated Teller Machine (ATM). High security is achieved by detecting harmful movements at the location. The proposed system uses Fuzzy logic and KNN classifiers to detect the motion accurately and take appropriate action. Fuzzification is implemented to work on larger data set and to observe different smaller data. Furthermore data is processed through the k-NN classifier to get the nearest result like it is intrusive behavior or normal behavior. If it is intrusive behavior then the door is locked and camera starts recording simultaneously message will be sent to the concern authority to take the counter action. Sensing system is developed with the help of geophone sensors, microcontroller. The proposed system achieves the accuracy of 98 percent to detect the harmful action.