ARIMA Model and Few-Shot Learning for Vehicle Speed Time Series Analysis and Prediction.

Yingzi Wang,Ce Yu,Yue Zhu,Yongjia Zhang,Jue Hou,Sisi Chu,Xin Ning

doi:10.1155/2022/2526821

Abstract

In the fields of traffic management, traffic health, and vehicle safety, vehicle speed prediction is an important research topic. The greater the difference between vehicle speed and average vehicle speed, or the more discrete the vehicle speed distribution, the higher the accident rate. This paper proposes a vehicle speed prediction method based on adaptive KF (Kalman filtering) in the ARMA (Autoregressive Moving Average) environment to address the problem of high-speed moving vehicle speed prediction. The ARMA theory is used to model the prediction of speed time series. The contribution rate of each coefficient representing the original time series is different after fitting the original time series with the ARMA model, so each coefficient must be given a certain weight. Multisource traffic data fusion and interval speed prediction are carried out on the basis of few-shot data preprocessing and traffic state division, according to different traffic states. The speed prediction accuracy is very high, according to the algorithm verification results.

Highlights

IntroductionIntersections account for roughly 30% of all traffic accidents [1]
According to statistics, intersections account for roughly 30% of all traffic accidents [1]
Time series differ from ordinary series in that the data are organized in chronological order, and each numerical point has a corresponding time point [7]. ese data are generated as part of the day-today operations of businesses, hospitals, schools, and other institutions, and they gradually accumulate into a large-scale time series database. e two types of data can be mutually supplemented and verified using data fusion. e accuracy of vehicle speed prediction across corresponding regions will be improved as well

Summary

Introduction

Intersections account for roughly 30% of all traffic accidents [1]. Drivers must pay attention to influencing factors such as vehicles traveling in different directions, pedestrians, and nonmotor vehicles, as well as the relative positions of their own vehicles, and make driving behavior decisions in a short amount of time [2, 3]. If the driver can be given some information about his own vehicle’s operation ahead of time, it will greatly reduce the driver’s driving difficulty at the intersection. When driving on the road, drivers set their own speed based on the terrain and road conditions [4]. When the road’s horizontal, vertical, and horizontal geometric elements exceed the minimum requirements for safe driving of automobiles on this grade of road, and external conditions such as traffic density, terrain, and climate are favorable, the actual driving speed of automobiles often approaches or exceeds the design speed [5, 6]. Time series differ from ordinary series in that the data are organized in chronological order, and each numerical point has a corresponding time point [7]. ese data are generated as part of the day-today operations of businesses, hospitals, schools, and other institutions, and they gradually accumulate into a large-scale time series database. e two types of data can be mutually supplemented and verified using data fusion. e accuracy of vehicle speed prediction across corresponding regions will be improved as well

Methods

Results

Conclusion