Traffic Amount Forecasting on a Road using Historical Data with MATLAB Programming

Abstract

MATLABSolutions demonstrate how to use the MATLAB software for simulation ofIn order to improve traffic conditions in the urban streets, traffic flow forecasting precision is one of the major tasks. How to improve traffic flow prediction precision remains an important problem of intelligent transportation systems. A lot of methods of Short- term traffic flow forecasting are designed. The methods are developed and optimized in order to obtain more accurate forecasting information. Often traffic flow forecasting accurate depends of traffic flow data collection accurate. We propose short- term traffic flow forecasting method based on improved data detection technology. Method tested with different forecasting models. Results show that proposed method is very suitable for short- term traffic flow forecasting.

Introduction

Nowadays intelligent transportation systems are being developed very much. But lots of problems such as traffic control, gridlocks, parking, road accidents aren’t solved yet. Traffic flow forecasting precision is one of the major tasks in order to improve traffic conditions in the urban streets. The way how to improve traffic flow prediction precision remains an important problem of intelligent transportation systems. There are two types of Traffic flow forecasting. They are Short- term forecasting and Long- term forecasting. Short- term traffic flow forecasting defines the traffic intensity for the next time interval, usually from 5 to 30 minutes. Long- term forecasting of the traffic intensity can be predicted hours, days or even years ahead. There are designed many methods of Short- term traffic flow forecasting. Usually Short- term forecasting methods are divided into two types. The first type includes classic methods, such as statistics. And the second type includes modern methods based on models such as neural networks or fuzzy. Methods are being developed and optimized in order to obtain more accurate forecasting information. Various methods are using the historical data base. Often traffic flow forecasting precision depends on accuracy of collected information about traffic flow. The imperfections of many methods are problems of traffic data losing or incorrect traffic data.

Traffic flow forecasting method, based on improved data detection technology

We propose short- term traffic flow forecasting method based on improved data detection technology. The main point of this forecasting method is traffic flow information collecting only in the biggest crossroads and forecasting it in the remaining smaller crossroads. Using proposed method, we crate three different forecasting models. They are linear regression forecasting model, forecasting model based on fuzzy logic and forecasting model based on neural networks. The most used traffic flow information collection technologies are inductive loop and video processing systems. Inductive loop technology has a lot of disadvantages; therefore we decided to improve traffic flow forecasting method with video detection technology. Designed short- term traffic flow forecasting method based on improved data detection technology is showed in video.