Traffic Safety Detection System by Digital Twins and Virtual Reality Technology

The present work studies the prediction of vehicle driving states to enhance the accuracy of traffic safety detection under new technologies. Firstly, the vehicle simulator and environment virtual system are built based on vehicle dynamics through virtual reality (VR) technology. Secondly, the vehicle Digital Twins (DTs) model is constructed based on various sensors and the Gaussian process algorithm. Besides, the vehicle simulator uses the Adams-Moulton-2 algorithm in CarSim software for numerical calculation. Finally, the background subtraction method is introduced to monitor and predict the vehicle motion state. The simulation results indicate that the engine of the vehicle DTs system constructed here changes with the rotational speed by the actual value. Besides, the maximum prediction error of the Gaussian process reported here is 2. 55, and the maximum error of the deep neural convolutional network is 4.29, indicating high prediction accuracy of the Gaussian process. Moreover, the background subtraction method selected in the present work has a high detection rate and low false alarm rate. The present work provides a reference for the development of DTs technology and VR technology in the field of transportation.