Ali Payıdar Akgüngör

An Artificial Intelligent Approach to Traffic Accident Estimation: Model Development and Application

Abstract This study proposes an Artificial Neural Network (ANN) model and a Genetic Algorithm (GA) model to estimate the number of accidents (A), fatalities (F) and injuries (I) in Ankara, Turkey, utilizing the data obtained between 1986 and 2005. For model development, the number of vehicles (N), fatalities, injuries, accidents and population (P) were selected as model parameters. In the ANN model, the sigmoid and linear functions were used as activation functions with the feed forward‐back propagation algorithm. In the GA approach, two forms of genetic algorithm models including a linear and an exponential form of mathematical expressions were developed. The results of the GA model showed that the exponential model form was suitable to estimate the number of accidents and fatalities while the linear form was the most appropriate for predicting the number of injuries. The best fit model with the lowest mean absolute errors (MAE) between the observed and estimated values is selected for future estimations...

A new delay parameter dependent on variable analysis periods at signalized intersections. part 1: Model development

Abstract Delay is an important factor in the optimization of traffic signals and the determination of the level of service of a signalized intersection. This paper proposes a methodology and a new formulation to identify the delay parameter in signalized intersection delay models. In this study, the delay parameter is modeled as a function of analysis period instead of a fixed value used by the existing delay models. Therefore, the proposed delay model including new delay parameter can produce more reasonable delay estimations at signalized intersections for variable time periods. A comparative study of the proposed time‐dependent model against the existing four different models was performed to present the improvements in this model.

A New Delay Parameter Dependent on Variable Analysis Periods at Signalized Intersections. Part 2: Validation and Application

Abstract The main objective of the present study is to investigate the performance of the proposed model in Part 1 for variable demand, time and oversaturated conditions. To accomplish this objective and test the proposed model, an experimental study was performed. The proposed delay model for oversaturated traffic conditions was calibrated and verified by the TRAF‐NETSIM microscopic simulation program. In the calibration and verification of the proposed model, the simulation study was performed to produce various traffic and time conditions using 48 different scenarios. The delays obtained from the simulations and the proposed model were statistically compared using linear regression analysis. The results indicated that there was a good relationship R2 = 0.989 at 95 % confidence level between the delays generated by the simulations and the delays estimated by the proposed model.

Road traffic accidents and safety programme in Turkey

Worldwide economic costs associated with road traffic accidents are estimated at around

Investigating urban traffic based noise pollution in the city of Kirikkale, Turkey

518 billion (US) annually. The author argues that Turkey has two to five times more traffic accident-related fatalities and injuries than some European countries primarily because Turkeys passenger and freight traffic is concentrated (95%) on its highway transportation network. Turkish traffic fatalities, injuries, and accidents are briefly evaluated in this study. The author concludes that there will continue to be a significant increase in the number of traffic injuries and accidents unless necessary precautions are made. Factors contributing to projected increases include poor health rehabilitation and emergency services, reckless driving, inadequacy of road safety policies, rapid increases in motor vehicle numbers, and an insufficient road system.

Evaluation of Ankara- Istanbul High Speed Train Project

Abstract The study presents an investigation into traffic based noise pollution in the city of Kirikkale, Turkey. For this purpose, traffic noise levels were measured at 15 intersections across the city during three peak times ‐ morning (08:00–09:00), noon (12:30–13:30) and evening (17:00–18:00) hours. The comparison of Leq values against the limit values of the Turkish Noise and Control Regulations for Settlement Zones showed that Leq values exceeded the limits at all stations. A linear regression analysis performed between the Leq and logarithm of total traffic volume (log Q) produced a coefficient of determination of 0.52. A multi regression analysis carried out between the Leq and four different vehicle types resulted in a correlation coefficient of 0.74. The correlation matrix indicated that the highest correlation was found for trucks/buses with r = 0.92. The spatial maps of traffic noise created by the Kriging method under ArcView GIS displayed that there seemed to be significant differences in the...

Estimating Traffic Accidents in Turkey Using Differential Evolution Algorithm

Abstract Many countries have been carrying out some changes in their transportation policies for the sake of environmental and economical considerations. These circumstances have also affected Turkey to change the national transportation policies. In this study, transportation policies of Turkey have been investigated and, in this context, an important railway project of Ankara ‐ Istanbul line has been analyzed. In evaluation, the present condition of the rail line, high speed train project is compared in terms of market study, technological and economical point of views. Finally, conclusions and suggestions are presented.

Comparison of artificial bee colony and flower pollination algorithms in vehicle delay models at signalized intersections

Abstract Estimating traffic accidents play a vital role to apply road safety procedures. This study proposes Differential Evolution Algorithm (DEA) models to estimate the number of accidents in Turkey. In the model development, population (P) and the number of vehicles (N) are selected as model parameters. Three model forms, linear, exponential and semi-quadratic models, are developed using DEA with the data covering from 2000 to 2014. Developed models are statistically compared to select the best fit model. The results of the DE models show that the linear model form is suitable to estimate the number of accidents. The statistics of this form is better than other forms in terms of performance criteria which are the Mean Absolute Percentage Errors (MAPE) and the Root Mean Square Errors (RMSE). To investigate the performance of linear DE model for future estimations, a ten-year period from 2015 to 2024 is considered. The results obtained from future estimations reveal the suitability of DE method for road safety applications.

Analysis and Modelling of the Relationship between Stopped and Control Delays by Differential Evolution Algorithm

Delay is a significant research topic since it includes indicators such as travel quality, lost time and fuel consumption. Furthermore, the delay is used for optimization of traffic control systems and determination of the level of service at signalized intersections. Therefore, researchers have focused on accurate estimation of delay. The objective of this study is to simply and accurately estimate the delay and evaluate the performance of the proposed approaches which are artificial bee colony (ABC) and flower pollination algorithms (FPA). In this study, ABC and FPA have been used to develop different delay models which are linear, semi-quadratic, quadratic and power forms. Analysis period ( T ), the green ratio ( g / C ; effective green to cycle length) and the degree of saturation ( x  =  v / c ; volume to capacity) are used as input parameters while developing the models. The results of present models are compared to estimations obtained from analytical models which are Highway Capacity Manual and Australian (Akçelik) delay models. Semi-quadratic form yielded to best results in terms of coefficient of determination ( R 2 ), mean square error and mean absolute error. Additionally, FPA approach showed better performance than ABC approach finding the optimal solution in the lower number of iterations.

Optimizing a fuzzy logic traffic signal controller via the differential evolution algorithm under different traffic scenarios

Abstract: Two delay definitions namely stopped delay and control delay are commonly used to determine the level of service of a signalized intersection. In this research, we first developed a number of new models to describe the relationship between stopped and control delays using Differential Evolution (DE) algorithm. In these models, we considered four cycle lengths of 60, 90, 120 and 150 seconds with the green ratio ranging from 0.35 to 0.60. Taking the saturation flow as 1800 vph, five different traffic volumes were considered ranging from 600 to 1400 vph with 200 vph increments. Three analysis periods were selected as 15, 30 and 60 minutes. The data needed for model development and verification were obtained from a four leg intersection by using CORSIM simulation. In the second part of the study, the effect and sensitivity of cycle length, traffic volumes and green ratio parameters on the conversion ratio (Ds/Dc) to relate stopped and control delays were evaluated based on the selected parameter ranges.