Identification of Remote IoT Users Using Sensor Data Analytics

Abstract

The immense progress of sensor technology and Internet of Things (IoT) has contributed well for the provision of various smart services through smart applications. These services include remote sensing, monitoring, control and operations in the fields of health care, transportation and weather forecast etc. alongside these great benefits users and device security prevails as a great challenge. Recently existing biometric identification methods are incorporated with other identification techniques of remote user recognition to improve the performance. In this research paper we have introduced a novel user identification framework using sensor data of walk activity. Accelerometer and heart rate sensors are used in combination for this purpose. As we know that both of these sensor readings are biologically more correlated during the walk activity. Heart rate is a unique biometric parameter for user identification whereas accelerometer sensor is known for its effective usage for activity recognition. The fusion method is adopted to make the proposed identification technique more customized to remove the overlapping probabilities of existing classification methods. The actual data set of 15 subjects is used for the experiments. The results are elaborated to prove the validity of the proposed approach. Accuracy for user identification is improved and a certain level of overlapping is reduced despite the low level of accuracy of heart rate sensors currently embedded in smart IoT devices.