ANN APPLICATION IN ROAD SAFETY MODELING

Abstract

The purpose of road accidents forecasting is to analyze the tendency of road accidents under existing road traffic conditions, to evaluate the feasibility and practical effectiveness of traffic safety measures reasonably, to control the factors affecting road accidents, and to reduce the traffic accidents. Characteristics such as nonlinearity, uncertainty and randomness in traffic system make it difficult to forecast road accidents, the behavioral feature of traffic system. The objectives of this study are to determine the critical accident variables for accident prediction purposes, to utilize Artificial Neural Network (ANN) as a tool for accident prediction analysis, and to develop an Accident Prediction Model at NH-58 (Non-urban highway from Km 72 to Km 130) Uttarakhand, India. This study uses a series of artificial neural networks to model and estimate accidents by the software, NeuroSolutions 6.0. A sensitivity analysis was performed and the obtained results illustrate that variables like mixed use (Schools, hospitals, religious buildings) developments, Side Drain Condition, Guard Rail, Marker Post, Curve Warnings and Chevron Markings, Cross Drainage works, Fly Over/Road over bridge/ Under Pass, Access Roads/Side Roads, Overhead Structures/Hoardings and traffic volume are most significant factors that increases accident. Remedial measures are also forwarded finally.