Serialized Recommendation Prediction for Steering Point Behavior of Intelligent Transportation Vehicles Based on Deep Learning

Abstract

This work aims to use Deep Learning (DL) techniques to predict vehicle steering and trajectory behavior in urban road traffic. First, the background of digital logistics and intelligent vehicle research is analyzed. Then, a vehicle trajectory tracking system is constructed through a coordinated control strategy to predict the vehicle trajectory. In addition, a behavioral serialization recommendation strategy for the steering point of the intelligent vehicle is proposed by modeling the dynamic state of the intelligent vehicle during driving. The research results demonstrate that the Root Mean Square Error (RMSE) of trajectory prediction shows a certain downward trend with the increase of network neurons. At the same time, the average value of systematic errors continues to increase. When the number of network neurons is 20, the predicted RMSE of the Unknown Input Observer (UIO)-1 group of vehicles is 2.0; the predicted RMSE of the longitudinal force UIO-3 group of vehicles is 1.1. In addition, the research topic is the behavior of intelligent transportation vehicle turning point. To this end, the vehicle steering behavior is measured at the turning point using the tire tilt angle of the vehicle at the turning point. When the vehicle starts, the inclination angle of the vehicles in the UIO-1 group is 5.5 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$^{\circ}$</tex-math> </inline-formula> , and the inclination angle of the vehicles in the UIO-2 group is 4.2 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$^{\circ}$</tex-math> </inline-formula> . This work provides theoretical reference and technical support for the digitalization and intelligent upgrade of smart city road traffic system.