Prayer Activity Recognition Using Smartphone

Abstract

There are 1.8 billion Muslims around the world who pray five times a day. Using the Human Activity Recognition Method, we can get a variety of benefits for diagnosing prayer activities, such as automatically switching smartphones to silent mode during prayers. Using smartphones, we can get this help Smartphones currently have many accelerometers, gyroscopes, magnetometers, and other sensors built-in. These sensors can be used to detect the activity of prayers based on artificial intelligence because, during the prayers, people perform certain activities and perform the prayers. In this paper, we have tried to determine the activity of prayers using the sensors of smartphones. Although there are many sensors in the smartphone, we have used three sensors namely accelerometer, gyroscope, and magnetometer respectively. The use of an accelerometer gyroscope as well as a magnetometer sensor has helped in getting the activities of the prayer accurately. We used machine learning to accurately predict the five prayer activities in this research project, with a maximum accuracy of 99.43%. We have seen that prayer activities are performed with different body movements Our model can diagnose this very well. We have proven that it is possible to increase this accuracy by using multiple sensors without using only one sensor.