Prediction of cooperative platooning maneuvers using NARX neural network

To accurately predict traffic information is of great importance in a large number of applications in connection with Intelligent Transport systems (ITS), since it reduces the uncertainty of future traffic states and improves traffic mobility. The most important research is done in the domain of cooperative intelligent transport system (C-ITS). Only minor attention has been given to coordinated maneuvering, since testing with real vehicles which can drive autonomously requires a large-scale infrastructure with important security measures. In this paper, we propose hybrid automaton modelling in Matlab/Simulink/ Stateflow to emulate flexible platooning conditions, analysing how cooperation interactions can be accomplished using inter-vehicle communication and certain control of the vehicles. Such analysis reveals to be necessary in order to establish the improvement of traffic mobility in Intelligent Transportation Systems through cooperation behaviour profile prediction. This study presents an approach towards NARX neural network prediction of flexible Platooning maneuvers profile. In order to estimate prediction, MSE and R were utilized. The study results suggest that in the case of noise in test data, NARX neural network would be an efficient prediction tool, and useful for the prediction mobility in Intelligent Transport systems.