Internet of Things data analytics for user authentication and activity recognition

Abstract

In recent years Internet of Things has been among the hot topics of research in the field of computing and information technology. It has enabled thousands of connected devices including sensors, cell phones and daily home appliances to store data for various useful purposes. The amount of data generated by IoT devices is huge, and there is a need to analyze IoT data for prospective uses. Heterogeneity and the structure of IoT data make it a challenging task. This paper presents a frame work for IoT data analytics for creating models for user identification, user authentication and activity recognition. Focus of this study is on the data set of accelerometer sensor from internet of things perspectives to add an extra layer of security. The novelty of our approach lies in the customization of the data set, and the experiments performed for the construction of models for each individual activity and user. The dataset is collected from 19 different subjects of real world conditions performing basic activities, i.e. walking, sitting and standing. This data set is used for the authentication process without requiring any additional information. In existing studies, the results are obtained using more than one accelerometer sensor reading or a combination of gyroscope sensor and accelerometer sensor. Whereas we have used single sensing tri-axial sensor reading for activity recognition and user authentication models. These models are later verified by real time data sets which were not used in the training process. The results of the experiments show accuracy up to 93%. The results obtained by the experiments are also helpful for future research directions in the field of IoT data analytics, activity recognition and user authentication. By enhancing the accuracy and adding context aware aspects in the authentication models can lead to the significant advances in the biometric authentication process using IoT data.