Road traffic accident(RTA)casualties in Sudan are among the major causes of death in the age group of 21 to 60.with 61%fatalities.The fatality rate of 35 per 10,000 vehicles is among the highest in the world despite t...Road traffic accident(RTA)casualties in Sudan are among the major causes of death in the age group of 21 to 60.with 61%fatalities.The fatality rate of 35 per 10,000 vehicles is among the highest in the world despite the low car ownership of 1 vehicle to 100 persons.This paper presents accident characteristics and considers road safety management.Crucial issues discussed in the paper include prediction and safety measures.The paper applies Artificial Neural Network(ANN)and regression techniques to comparatively predict traffic accident casualties.Both approaches modeled accident casualties using historical data on population,number of registered cars and other related factors from 1991 to 2009.Comparison of predictions with recorded data was very favorable.Predictions during 2010-2014 were determined using projected values for the same predictor variables.ANN forecasts provided the best fit for the data with a maximum difference of 1.84%between predictions and observed data.The study demonstrated that ANNs provide a powerful tool for analysis and prediction of accident casualties.The major causes of accidents were attributed to driver behaviour,vehicle fleet and conditions,road network defects,speed-limit violation,negligence of seat-belt usage and lack of traffic-law enforcement.展开更多
文摘Road traffic accident(RTA)casualties in Sudan are among the major causes of death in the age group of 21 to 60.with 61%fatalities.The fatality rate of 35 per 10,000 vehicles is among the highest in the world despite the low car ownership of 1 vehicle to 100 persons.This paper presents accident characteristics and considers road safety management.Crucial issues discussed in the paper include prediction and safety measures.The paper applies Artificial Neural Network(ANN)and regression techniques to comparatively predict traffic accident casualties.Both approaches modeled accident casualties using historical data on population,number of registered cars and other related factors from 1991 to 2009.Comparison of predictions with recorded data was very favorable.Predictions during 2010-2014 were determined using projected values for the same predictor variables.ANN forecasts provided the best fit for the data with a maximum difference of 1.84%between predictions and observed data.The study demonstrated that ANNs provide a powerful tool for analysis and prediction of accident casualties.The major causes of accidents were attributed to driver behaviour,vehicle fleet and conditions,road network defects,speed-limit violation,negligence of seat-belt usage and lack of traffic-law enforcement.
