International Journal of Transportation Engineering
Volume:11 Issue: 2, Autumn 2023

The rapid growth of urbanization and the global population have resulted in climate change, air contamination, and various human health problems. Thus, estimating air pollution indices has become important to environmental science studies. With relevant data increasingly available, machine learning frameworks have been proposed as a particularly useful method to predict air pollution. Based on four years of Tehran’s neighborhood air pollution data analysis, this paper proposes three machine learning approaches to predict NO2 and CO concentration: Autoregressive Integrated Moving Average (ARIMA), Long Short-Term Memory Networks (LSTM), and Multiple Linear Regression (MLR). This paper compared the ability of the ARIMA, LSTM, and MLR machine learning methods to forecast the daily concentrations of NO2 and CO at Punak air quality monitoring station, from 2017 to 2020. By applying four performance measurements, the ARIMA model displays the worst performance among the three models in all datasets with RMSE values of 47.39 and 1.29, and 0.012 and 0.01 for NO2 and CO respectively. The LSTM and MLR models achieve the best forecasting result with RMSE = 17.6 and 6.41, MAE = 10.59 and 4.33, = 0.458 and 0.46, and RRSE =1.06 and 1.10 for NO2 forecasting and RMSE = 0.42 and 0.32, MAE = 0.24 and 0.25, 0.96 and 0.98, and RRSE = 0.43 and 0.44 for CO forecasting.

Road accidents reduce traffic safety due to injuries and fatalities. Investigating and prioritizing factors contributing to road accidents have been based on deficient traditional ways as they do not consider the probability density of factors contributing to road accidents. Accordingly, an examination of accident road factors based on the probability density seems necessary. Thus, this paper first aimed at using principle components analysis (PCA) as a statistical prioritization tool for identifying the main and sub-main factors that contribute to injury severity on Borujerd-Khorramabad as a four-lane rural highway during the years 2015 to 2017. Secondly, the multivariate Gaussian probability model was used as a probabilistic density approach to estimate the probability density based on the relationship between factors that contributing to injury severity and the Pearson correlation. The results obtained through PCA indicated that factors contributing to injury severity were ranked in terms of Eigen values and rotated component matrix. Findings from the PCA model showed that OS, PSL, AADT, SL, R, and S, as 6 important factors affecting the accident occurrence relevant to injury severity. The results of the probability density also showed that the relation between operating speed with posted speed limits, and the relation among segment length, operating speed and radius are considerable due to increasing the probability density of accident occurrence. Moreover, AADT with operating speed and operating speed with slope and radius have significant effects on the probability density of occurring accidents. The results of the present study show that applying a multivariate Gaussian probability model helps to estimate the probability density of the accident occurrence of factors contributing to road accidents based on their values.

The most common cause of death and injury worldwide is traffic accidents. Several studies have shown that these accidents occur primarily on rural roads, especially those with horizontal curves. Studies have indicated that inconsistency in geometric design is one of the main reasons for this problem. Road geometric consistency refers to the alignment of the road geometry with the expectations of the driver. In evaluating road design consistency, operating speed is the most common criterion, which can be estimated by using operating speed models. Buses account for the majority of passenger movement, so it is important to study their operating speed. In this work, the random effect model is used for 150 buses passing through 690 horizontal curves of the Tehran-Qaem shahr highway to estimate the performance speed in horizontal curves. Among all the highways connecting the northern area of the country to the capital, that expressway was chosen for the study because of many reasons such as the minimum changes in the geometry, high volume of daily bus traffic, various topographic conditions, and the availability of data needed for the study. Besides, the variables of driving time, the geometric design of the curve, the driver's data, and the characteristics of the bus are examined in this study. Finally, the developed model revealed that the average radius of the curve, the volume of traffic, the reverse curves, the percentage of heavy vehicles, the location of fixed speed cameras, the posted speed limit, and the darkness were all significant factors.

A dynamic, efficient transportation network is an important index of country development and currently, major parts of infrastructure projects-related credits and budgets are allocated to road construction projects. Since several risks divert such projects, during their construction, from their main goals, proper risk identification and management are necessary to better implement infrastructure and road construction projects. As risks are changeable factors that differ from country/region to country/region, this research has reviewed the literature and used the experts' opinions to identify the most influential ones in Iran to eliminate or reduce their effects on the time and cost of road construction projects. To this end, a questionnaire was designed to identify the risks and prioritize them using the failure modes and effects analysis method in a fuzzy environment; defuzzification was done by MATLAB Software. Scores of the risks revealed that: 1) inflation (increased material price), 2) late financial provisions, 3) deficiency, failure, or defect of equipment/machinery, and 4) Maps and specification changes of more than 25% with respect to the values specified in the general conditions of contract due to the employer’s incorrect studies/estimation of the project, were identified as the most important risks of road construction projects.

The type of land use in each traffic area zone (TAZ) is the most important factor determining the number of vehicles, geometric and traffic characteristics in that zone. Any factor in the urban environment that causes congestion and attraction of vehicles at certain times increases the probability of a crash in that area. The purpose of this study is to investigate the effect of the share of different types of uses in various traffic areas of Shiraz city on the probability of a crash. A two-step method, including identifying the types of uses influencing the occurrence of crashes and spatial effects between independent variables and crash data in space Kernel density estimate (KDE) methods, has also been used to find the suitable bandwidth for searching observations. In order to investigate the spatial effects of land use types on crash incidence, geographically weighted regressions (GWRs) and geographically weighted Poisson regressions (GWPRs) were used. Based on the validation criteria, the local GWPR model performs better than the global Poisson model and the local GWR model among the mentioned models .Additionally, the presence of residential, commercial, barren, and abandoned uses, as well as the mixing of residential and non-residential uses, significantly impact crashes. Examining the spatial effects of land use types in this study's traffic areas can be very important in carrying out safety measures.

Given the importance of reducing crashes, as well as the implementation of principled and practical road safety measures, it is necessary to calculate the Crash Modification Factors (CMFs), which is the main factor in road safety effectiveness analysis, with high accuracy. Therefore, reliability assessment is of great importance and necessity for calculating CMFs. In this study, a method for evaluating the reliability of CMFs using metaheuristic Genetic Algorithm (GA) is presented. The proposed model is defined based on the before-after study method with the comparison group and based on the Full Bayesian (FB) method for calculating CMFs. The Monte Carlo Markov Chain (MCMC) has been applied for calculating posterior distributions as a sampling method that allows the simulation of posterior samples from complex distributions. Crash data, categorized into total crashes and fatal-injury crashes, were collected from the city of Karaj and had a period of 5 years (2016-2020). The remedial action of signalizing the intersections along with the installation of the counters was considered as the treatment considered in the study. Results show that the CMF value for remedial action of signalizing intersections with counters, did not have a significant impact on reducing total crashes (CMF=1.07). On the other hand, by evaluating the CMF values calculated for fatal-injury crashes, it is determined that the calculated CMF is approximately equal to 0.75, which indicates the positive effect of the remedial action and reduction of fatal-injury crashes. In addition, according to the proposed GA-based rating system, CMF values for fatal-injury crashes have the highest rank, which indicates a very high reliability for the calculated CMF values. Therefore, it is possible to confidently take the remedial action of signalizing intersections with the installation of a counter as an effective measure to reduce the number of fatal-injury crashes.