A New Vehicle Arrival Prediction Model for Adaptive Signal Control in a Connected Vehicle Environment

Liu Yang,Yibing Wang,Zhihong Yao

doi:10.1109/access.2020.3002943

Abstract

With the development of connected vehicles (CVs) technology, it has become a new research topic to capture the dynamic traffic system by using CVs data. The traditional vehicle arrival prediction models are limited to the fixed detectors, which can only recognize the passing information of vehicles, but cannot identify the state of vehicles. This paper proposed a new vehicle arrival prediction model of traffic signal control in a connected vehicle environment. First, the vehicle's identification number (ID), position, velocity, and acceleration were obtained in a CV environment. Second, a new vehicle arrival prediction model was developed by the joint probability distribution calibrated based on CV data. Then, the simulation experiment was designed for analyzing influence factors on model performance. The results show that the average relative error of each factor is less than 10%. Finally, the reliability of the proposed model was verified based on an adaptive signal control algorithm. Therefore, the proposed model can be used in the adaptive signal control system.

Highlights

The vehicle arrival prediction model is a vital part of the adaptive signal control system, and its performance directly affects the control effect of the adaptive control system [1]–[3]
For the platoon-dispersion-based model, its data comes from fixed loop detectors, which only provides instantaneous vehicle position when the vehicle passes through the detector; it is impossible to directly measure other vehicle information
WORK Based on the joint probability distribution function of vehicle location and speed, a vehicle arrival prediction model was proposed in the connected vehicles (CVs) environment

Summary

INTRODUCTION

The vehicle arrival prediction model is a vital part of the adaptive signal control system, and its performance directly affects the control effect of the adaptive control system [1]–[3]. For the platoon-dispersion-based model, its data comes from fixed loop detectors, which only provides instantaneous vehicle position when the vehicle passes through the detector; it is impossible to directly measure other vehicle information (i.e. vehicle ID, position, velocity, and acceleration). If one or more detectors malfunction, the performance of the vehicle arrival prediction model can degrade significantly Another is that many studies only focus on predicting travel time, which cannot meet the requirements of an adaptive control system. A CV simulation experiment is developed to demonstrate that the proposed vehicle arrival model can be implemented in adaptive control algorithms or systems.

LITERATURE REVIEW

MODEL PERFORMANCE EVALUATION INDEX

MODEL APPLICATION

Findings

CONCLUSION AND FUTURE WORK