The Impact of Hybrid Data Fusion Based on Probabilistic Detection Identification Model for Intelligent Rail Communication Highway

Abstract

Data fusion has an emerging technology that received core attraction for maneuvering and non-maneuvering application. The multiple sources in the data fusion process are distributed accuracy based systems that differ from traditional single source in many ways; these have suppressed the irrelevant data and noise, better numerical precision, reliable, stable and accurate. The end-to-end basic probabilistic models have been proposed for different architectures in data fusion to manage sensor information. This paper proposes probabilistic detection identification model based on Bayes theorem and compare its performance with traditional single source network. Subsequently, it is cross validated using various data folds of kinematics measurements available from Differential GPS, Wireless Sensor and Radio Frequency Identification. This study proves the probabilistic model offers significant detection performance accuracy level of 95% for moving locomotives across wide range of operational scenarios.