System identification using neuro fuzzy approach for IoT application

Abstract

The Internet of Things (IoT) has become a popular application in recent years. However, it is the wireless communication mode. In such a scenario, the user would have to send information either nonlinear or dynamic data type in the form of a signal or an image, videos depending on the application. The proposed work is focused on this model identification that tends to nonlinear dynamic system identification for IoT applications. An Autoregressive Moving Average (ARMA) model represents model for IoT application. To verify the model supremacy, an ARMA bench mark system is used. The adaptiveness is proved through variation of weights and can be universally used for the next generation. In the first attempt, the Multilayer Perceptron model (MLP) is considered as the ARMA system and observed. Further, to improve its accuracy, the Adaptive Neuro-Fuzzy system (ANFIS) model is designed for system identification. It is shown in the result section that it identifies better than the MLP as well as traditional system identification techniques.

Keywords

• Adaptive neuro fuzzy system
• Wireless sensor Network channel
• Nonlinear dynamic system Identification
• Internet of things
• Autoregressive moving average