DSHMP-IOT: A distributed self healing movement prediction scheme for internet of things applications

摘要

The IOT infrastructure - IP-based mobile sensor network- makes it possible to provide two-directional communication between mobile sensors and the remote server. Mobility direction prediction is a major challenge in IP-based networks by which we can predict the next movement direction of moving objects carrying mobile node(s). In this paper, we have introduced a Distributed Self-Healing Movement Prediction scheme for IOT applications, so-called DSHMP-IOT, to predict movement direction of mobile IP-based sensors in a multi-user environment, such as a health-care system. This is the first time that an AI solution is applied to predict the direction of the mobile node(s) in an IP-based mobile network. The proposed scheme takes advantage of Hidden Semi-Markov Model (HSMM) to predict the movement direction with high accuracy and low overhead. The previous works for estimating the direction of a mobile node(s) in IP-based mobile networks are based on AOA, a hardware-specific method. The proposed scheme has several advantages. First, it eliminates the need for special hardware (directional antenna, an antenna array, etc.) which is required in AOA based methods. Second, it is not sensitive to noise, speed and sudden changing of movement direction which cause false positive movement direction prediction in AOA method. Third, in this context, it is the only work with self-healing capability whenever one or more static sensors fail(s). Fourth, it includes a recovery mechanism which prevents the mobile node from being disconnected in case of false prediction of our learning model. The simulation results show the superiority of our scheme regarding power consumption and hand-off delay, as well as packet loss, compared to similar approaches.