Developing a Gait-Based Stacked Ensemble Authentication Framework for Internet of Things
Irshed Hussain, S. Gopinath, Ish Kapila, B. Rajasekar, Hitesh Kalra, Srinivasa Rao K S · 2025
It consists of the utmost importance that Internet of Things (IoT) systems have robust authorization techniques because of the sensitive nature of data that is updated constantly, particularly in healthcare settings along with additional industries that need confidentiality. With the use of cellphone sensors, this study provides a gait-based stacking ensembles authenticating framework that was developed with the intention of securing access to IoT devices. A smooth and ongoing verification of users depending upon their stepping patterns is ensured by the structure, which makes use of information collected by accelerometers and gyroscopes to gather gait attributes. Identification is handled via a stacking ensembles artificial intelligence classifier, which is comprised of two primary parts: the preprocessors for the reduction of distortion and the collection of features. Users may be properly identified by using characteristics such as quadratic speed, rotation percentage, and gravitation. The proposed structure outperforms other authenticating approaches when it comes of dependability and effectiveness, as shown by the results of experiments, which show that it obtains an accuracy rate of 98%. It is suited for use in practical uses since the system overcomes prevalent problems in Internet of Things authorization, including high mistake rates and lengthy development times. Greater safety including decentralized authorization in large-scale Internet of Things communities will be the subject of future research, which will concentrate on expanding this framework in order to support changes in circumstances and combining it with blockchain technology.