Modeling of injured position during transportation based on Bayesian belief networks

Abstract

Development of rescue robotics for injured transportation connected with problem of selection of injured position based on trauma type. The paper presents a model of a position of the injured during transportation based on Bayesian belief networks. The developed Bayesian belief network structure is represented by the signs of trauma, trauma itself, positions for transportation of the sufferer corresponding to trauma, and the relationships between them. Conditional probabilities tables is determined based on the expert information; available medical research focused on the identification of the similar relationships between the elements of the diagnostic process; historical statistics. The simulation results show that the developed Bayesian belief network enables one to solve probabilistic forecasting tasks based on subjective and incomplete data. The former are obtained during questioning the sufferer; the latter are based on the computer vision systems (examination) and sensors for various purposes (manipulation) that are installed on specialized robots. The developed tools are focused on the rescue robotics based on intelligent hardware and software for human-robot interaction.