Using Spatial Analysis and Machine Learning Techniques to Develop a Comprehensive Highway-Rail Grade Crossing Consolidation Model

Current collision prediction models fail to account for the full spectrum of relevant factors affecting the number of collisions at specific highway– railway grade crossings. A number of reasons contribute to this failure, including biases in model parameters resulting from collinearity in the model inputs, absence of important variables in the prediction model caused by lack of statistical significance, the inability of models to consider higher-order interactions, and the presence of unexplained variation in the prediction estimates. These problems have compromised the use of collision prediction models in decisions concerning the development and evaluation of cost-effective safety treatments or countermeasures for application at specific crossings. This paper introduces a stratified collision prediction model for highway–railway grade crossings. The development of this model involves three steps: ( a) crossing inventory variables are expressed in terms of a limited number of orthogonal (nonlinear) underlying attributes or factors; ( b) factor scores are estimated for each crossing and factor, and these scores are used as “seed points” in a subsequent clustering exercise to yield groups or clusters of crossings with similar underlying attributes; and ( c) for each cluster, separate collision prediction models are developed and include important treatment input variables of interest to decision makers and planners. The paper describes an application of a stratified collision prediction model to Canadian highway–railway grade crossing inventory and collision occurrence data for the period 1993 to 2001. The usefulness of the model in estimating collision reduction benefits of selected treatments is illustrated with reference to two countermeasure strategies: upgrades in the type of warning device and the removal of whistle prohibition.

Various countermeasures can be introduced to reduce collisions at highway–railway grade crossings. These countermeasures may take different forms, such as passive and (or) active driver warning devices, supplementary traffic controls (four quadrant barriers, wayside horn, closed circuit television (CCTV) monitoring, etc.), illumination, signage and highway speed limit, etc. In this research, we present a structured model that makes use of data mining techniques to estimate the effect of changes in countermeasures on the expected number of collisions at a given crossing. This model serves as a decision-support tool for the evaluation and development of cost-effective and practicable safety program at highway–railway grade crossings. The use of data mining techniques helps to resolve many of the problems associated with conventional statistical models used to predict the expected number of collisions for a given type of crossing. Statistical models introduce biases that limit their ability to fully represent the relationship between selected countermeasures and resultant collisions for a mix of crossing attributes. This paper makes use of Canadian inventory and collision data to illustrate the potential merits of the proposed model to provide decision support.Key words: highway–railway grade crossing, collision prediction model, countermeasures, Poisson regression.

In the United States, highway-railroad grade crossings are easily congested, which not only causes significant traffic delays to travelers but also brings potential threats to the first responders for emergencies. Unfortunately, very limited research efforts have been dedicated to developing practical systems that can assess traffic conditions at overcrowded grade crossings. The main challenge in evaluating the congestion conditions at the crossings is the different instance classes (i.e., vehicle, train, and pedestrian) that need to be accurately detected, especially when densely packaged. In this study, a novel convolutional neural network (CNN) named dense traffic detection net (DTDNet) is developed. DTDNet proposes to integrate the Transformer Attention (TA) module for better modeling of global context information and the learning-to-match detection head for optimizing object detection and localization using a likelihood probability fashion. To train and test DTDNet, a unique grade crossing traffic image dataset including congested and normal traffic during both daytime and nighttime is established. Experimental results on the dataset show that the proposed DTDNet achieves the maximum mean average precision (mAP) value, 0.832, outperforming the other state-of-the-art (SOTA) models. Field test results with low mean average error (MAE), mean relative error (MRE), and root mean squared error (RMSE) which are 2.200, 1.890, and 0.280, respectively suggest the proposed model has a satisfying and robust performance in the field application under different environments.”.

The ability to accurately forecast train arrival times is essential for the safe and efficient operation of highway–railroad grade crossings (HRGCs). Trains in the United States are required to give a minimum of 20 s of warning time before arriving at an HRGC. With the recent development of new detection-equipment technology, detectors potentially could be employed further upstream of the HRGC, which would result in earlier detection times. This information would be particularly useful for preemption strategies at signalized intersections located near the HRGC (IHRGCs). For example, earlier warning times could be used to reduce or eliminate the risk of unsafe pedestrian movements at IHRGCs. In this study, a modular artificial neural network (ANN) was used to forecast the train arrival time at an HRGC. An ANN was adopted because there is a nonlinear relationship between the independent variables such as train speed profile and the dependent variable arrival time at an HRGC. A modular approach was used because the trains often have different characteristics depending on their cargo and the operational rules in effect at the time they are detected. Because the train detection is continuous, different models were developed for each separate data input. In this case, the prediction interval update was assumed to be 10 s and 24 models were developed. Approximately 499 trains were used for training the ANN and 183 trains were used for testing. It was found that a modular architecture gave superior results to that of a simple ANN model, standard regression techniques, and current forecasting methods for the entire detection time period. It was found that, with an increase in detection time, the forecast accuracy increases for all methods and the prediction interval tends to decrease.

The overall goal of this study was to develop warrants or guidelines for the installation of active warning devices without speed selection (fixed-distance systems) and active warning devices with speed selection (constant warning time systems) at railroad highway grade crossings. This objective was achieved by the use of the FRA grade crossing inventory and accident data bases to develop accident rates, effectiveness values, statistical models to determine which variables are significantly related to accidents, and a benefit cost analysis, all of which were used in warrant development. Overall results of this study tended to confirm the hypothesis that constant warning time systems have greater credibility with motorists than do fixed-distance systems. The effectiveness values for changes from fixed-distance to constant warning time systems confirmed this. Accident rates for both fixed-distance and constant warning time systems were developed and showed that on the average, constant warning time systems tended to have lower rates. A separate regression technique showed that the only two variables that were significantly related to accident rates were warning device types and exposure. In general, constant warning time systems were warranted at a higher exposure level than fixed-distance systems. Keeping the cost ratio fixed, flashing lights of fixed distance are warranted at the least exposure level for crossings formerly equipped with passive devices, followed by flashing lights of constant warning time, gates of fixed distance and last, gates of constant warning time. For crossings formerly equipped with gates of fixed distance, gates of constant warning time will rarely be warranted.

The safety benefit of stop-sign treatment employed at passive highway–rail crossings has been a subject of research for many years. The objective of this study is to assess the effectiveness of the stop-sign treatment on crossing safety. Using the Federal Railroad Administration database, the research focused on 26 years of vehicle–train accident history in the United States from 1980 through 2005. A before-and-after and cross-sectional statistical analysis was conducted for 7,394 public highway–railroad grade crossings that were upgraded from cross buck only to stop signs without involvement in other traffic-control devices (TCDs) or automatic countermeasures. The study found that accident rates based on annual accident frequency per 1,000 crossings were significantly higher during the period when crossings were controlled by cross bucks only than when they were controlled by stop signs. Further, this study developed negative binomial accident prediction models, respectively, for paved and unpaved highway–rail grade-crossings that include effect for stop-sign treatment. Based on specific attributes of the current crossbuck-only-controlled crossings, decision makers and traffic engineers can use the models to examine the accident risks at crossings and assess the potential effectiveness of stop-sign treatment.

A latent class modeling approach for identifying vehicle driver injury severity factors at highway-railway crossings

The purpose of this research investigation was to analyze the effects of environment, tonography, geometry, and highway and rail traffic patterns with respect to rail-highway grade crossing accidents in rural areas.

This study applied a holistic method of healthcare program assessment, known as systems approach, to evaluate the efficacy of the East Baton Rouge Parish (EBRP) EMS Community-integrated Health Program (CIHP). The author developed assessment metrics following the Quadruple Aim Framework: IHI (Institute for Healthcare Improvement) Triple Aim + Provider Experience. Results showed that the CP program in this study delivers high ratings on patient experience and provided a 51% return on investment (ROI) through reduced emergency service utilization, ED visits and inpatient hospital admissions. However, findings did not demonstrate meaningful improvement in patients’ health-related quality of life (HRQoL) and self-efficacy (SE). Nevertheless, provider (specially-trained EMS paramedics) experience from the mixed-methods assessment offers suggestions to improving the program, as well as insights into the barriers to healthcare access that are often encountered by the frequent ED user population.

Abstract: Research on the land use of the coastal zone in the sea–land direction will not only reveal its land use distribution, but may also indicate the interactions between inland land use and the ocean through associations between inland land use and seaward land use indirectly. However, in the existing research, few have paid attention to the land use in sea–land direction, let alone the sequential relationship between land-use types. The sequential relationship would be useful in land use planning and rehabilitation of the landscape in the sea–land direction, and the association between land-use types, particularly the inland land use and seaward land use, is not discussed. Therefore, This study presents a model named ARCLUSSM (Association Rules-based Coastal Land use Spatial Sequence Model) to mine the sequential pattern of land use with interesting associations in the sea–land direction of the coastal zone. As a case study, the typical coastal zone of Bohai Bay and the Yellow River delta in China was used. The results are as follows: firstly, 27 interesting association patterns of land use in the sea–land direction of the coastal zone were mined easily. Both sequential relationship and distance between land-use types for 27 patterns among six land-use types were mined definitely, and the sequence of the six land-use types tended to be tidal flat > shrimp pond > reservoir/artificial pond > settlement > river > dry land in sea–land direction. These patterns would offer specific support for land-use planning and rehabilitation of the coastal zone. There were 19 association patterns between seaward and landward land-use types. These patterns showed strong associations between seaward and landward land-use types. It indicated that the landward land use might have some impacts on the seaward land use, or in the other direction, which may help to reveal the interactions between inland land use and the ocean. Thus, the ARCLUSSM was an efficient tool to mine the sequential relationship and distance between land-use types with interesting association rules in the sea–land direction, which would offer practicable advice to appropriate coastal zone management and planning, and might reveal the interactions between inland land use and the ocean.

Although several factors are known to contribute to collisions at highway–railway grade crossings, the mixed effects of the control factors such as a highway class and other countermeasures on collision occurrence are less well explored. This study evaluates the relationship between countermeasures and collision occurrence by using a sequential analytic strategy that combines the tree-based data stratification method with the generalized linear regression technique. After the control factor effects were controlled with the use of the tree-based data partitioning method, stratified collision prediction models were developed, and the adjusted effect of the selected countermeasures was identified. The conventional statistical models were also provided for comparison with stratified models. The study used Canadian inventory and collision data to demonstrate the expected collision reductions resulting from changes in selected countermeasures, such as warning devices and posted speed limit.

Land use change can be predicted by spatial modeling. One of spatial method to predict land use change is CLUES (Conversion of Land Use Changes and its Effect at small regional extent). The objective of the research is to determine factors that influence land use change in Kabupaten Majalengka and to predict its land use in the future. Data applied in this research project was biogeophysic, socio economic and existing land use data of Kabupaten Majalengka. Eight land use classes applied were shrub, forest, pond, dry land agriculture, pasture, settlement, and plantation and rice field. Land use classes were dependent variables; and biogeophysical and socio economical data were independent variables. Binary logistic regression was applied to determine influenced factors to land use change. Factors that influence of type of land use’s occurrence were different for each type of land use. The value of the factors was used as an input in land use modeling using CLUES. Scenario that used were scenario with land use change rate same as previous change. The simulation was applied for 12 years with the baseline data in the year of 2009. The simulation result shows that the land use types of shrub, forest, and plantation are decrease. The area that increased are land use types of dry land agriculture, settlement and rice field; however land use types of pond and pasture are remain constant.

An algorithm of regional assessment of the size and structure of the actual (corresponding to the modern state of ecosystems and land use patterns) and potential (for hypothetic natural ecosystems analogous to modern native ecosystems) pools of carbon has been developed and tested. A comparison between actual and potential values of carbon pools makes it possible to assess the integral result of land use in the studied region with multiple changes in the types of land use during the historical period. The calculations are made using a unified cartographic base and take into account the taxonomic position and texture of soil units, the types of modern land use, and the type and age structure of the reconstructed and actual vegetation. The results obtained for the southern taiga and forest-steppe zones of European Russian indicate that the modern actual carbon pool is 24% less than the potential carbon pool in Kostroma oblast (southern taiga zone) and 32% less than the potential carbon pool in Kursk oblast (forest-steppe zone). The actual phytomass reserves in these two regions have decreased by 40 and 75%, respectively, relative the potential phytomass reserves, so the portion of the soil carbon pool in the total carbon pool has increased. It is argues that the use of the territory for forestry and agriculture increases the role of the soil cover in sustaining the carbon budget of the region.

To reveal the effect of the interactions between soil depth and different land use types on soil nutrients and soil bacterial communities in a karst area, fifty soil samples from five different karst land use types in Huajiang town, Guizhou province, Southwest China were collected, and the soil bacteria were analyzed using high-throughput absolute quantification sequencing. Our results showed that land use types (LUT) and soil depth (SD) significantly influenced the content of soil organic carbon (SOC), total nitrogen (TN), total phosphorus (TP), nitrate nitrogen (NN), ammonium nitrogen (AN) and available soil phosphorus (AP), and pH; further, the interaction of LUT and SD also significantly influenced SOC, NN, NA, AP, and pH. In addition, LUT clearly impacted the Chao1 and Shannon indexes, but, SD and LUT * SD markedly affect Chao1 and Shannon index, respectively. All the soil bacterial communities were significantly different in the five different five land use types according to PERMANOVA. Importantly, Acidobacteria and Proteobacteria were the predominant phyla at soil depths of 0–20 cm and 20–40 cm among all the LUTs. At 0–20 cm, TN, AN, and SOC exerted a strong positive influence on Acidobacteria, but NN exerted a strong negative influence on Acidobacteria; at 20–40 cm soil, TN and AN exerted a strong positive influence on Acidobacteria; TP exerted no marked influence on any of the phyla at these two soil depths. At 0–20 cm of soil depth, we also found that Chao1 index changes were closely related to the TN, SOC, AN, and NN; similarly, Shannon index changes were significantly correlated to the AN, TN, and SOC; the PCoA was clearly related to the TN, SOC, and AN. Interestingly, at soil depth of 20–40 cm, Chao 1 was markedly related to the TN and pH; Shannon was markedly correlated with the SOC, TP, AN, and AP; and the PCoA was significantly correlated with the TN and pH. Our findings imply that soil nutrients and soil bacteria communities are strongly influenced by land use types and soil depth in karst areas.

Relations between rare earth elements accumulation in Taraxacum officinale L. and land use in an urban area – A preliminary study