Motion Detection with IoT-Based Home Security System

Abstract

This paper presents a development of motion detection technique with an IoT-based Home Security System. Two low-cost motion detectors, namely, pyroelectric infrared radial (PIR) and microwave sensors are considered in this work, and a Raspberry Pi is selected as the main platform to receive the sensor signals and send a push notification to the user whenever an intruder is detected. The performance of each sensor was firstly analysed in terms of its detection accuracy and false activation. A series of experiments were designed in a car porch area with various walking trajectories to evaluate the accuracy of each sensor, and three environmental scenario with different wind speeds were considered for the evaluation on false triggers. We also introduce a simple algorithm for each sensor to minimize the effects from the sensors’ noise. Based on the experimental results, it is shown that the microwave sensor outperforms the PIR in terms of both detection accuracy and false activation. It is also observed that, the implementation of the algorithm offers a great reduction in terms of false activation, but with a slight trade-off in motion detection accuracy. Nevertheless, the overall performance of the microwave sensor is enhanced as the false trigger can be totally eliminated at the cost of only 5% accuracy reduction, compared to that of the PIR sensor which still shows a non-zero false triggers with 6.25% accuracy reduction after the implementation of the algorithm.