Assessing Safety and Infrastructure Design at Railway Level Crossings Through Microsimulation Analysis

Abstract

The European Union (EU) is paving the way toward “Vision Zero”, a future goal of eliminating road fatalities and severe injuries. Railway level crossings are critical safety hotspots where road and rail traffic intersect and present a unique challenge in balancing the safety of both rail and road users while ensuring efficient traffic flow. Collisions at these crossings account for a significant proportion of railway-related fatalities in the EU, underscoring the need for targeted safety interventions. This article explores the impact of signal preemption strategies on the safety and operational performance of railway level crossings through a microsimulation analysis. Using VISSIM, a railway level crossing and its adjacent road intersection were modeled under existing and alternative scenarios. The preemption strategy was designed to clear vehicles from the crossing area before train arrivals, reducing conflict risks and optimizing traffic flow. Key findings reveal that the proposed preemption strategy significantly reduces queue lengths within critical safety zones, mitigating vehicle spillback and enhancing operational efficiency. The analysis highlights the importance of integrating railway operations with traffic signal systems, particularly in urban areas with limited queue storage capacity.

1. Introduction

Collisions between road users (such as drivers of motor vehicles, motorcyclists, cyclists, and pedestrians) and trains at rail level crossings (RLCs) represent a major safety issue in many countries, particularly within the European Union (EU). According to the EU Agency for Railways [1], in 2021, over 32% of railway-related fatalities in the EU were caused by crashes on railway level crossings. Except for human lives and fatalities, which account for around 65% of total costs in the EU [2], an unsafe railway system can directly negatively impact the economy. Indicatively, the cost of delays, costs of material damage or rolling stock or infrastructure, the costs to the environment, and costs related to the modal shift, administration, rerouting, reputational damage, and productivity losses are some of the most important to be mentioned.

Several studies [3,4,5,6,7,8,9,10] have focused on understanding and identifying the factors that affect the safety level at railway level crossings. According to Read et al. [11], the physical and operational characteristics of RLCs and the behavior of road users are the most critical factors that influence the rate and severity of crashes.

To mitigate the frequency and severity of crashes at railway level crossings and enhance safety, effective measures such as grade separation and the installation of boom barriers can be considered [12,13,14,15,16]. However, the implementation of these solutions often proves to be prohibitively expensive [17,18,19]. The adoption of low-cost countermeasures to minimize the exposure and risk of crashes and accidents is being discussed in several studies [20,21]. A frequently cited argument suggests that the funds required to install low-cost rail level crossing warning devices (LCLCWDs) could be used instead to upgrade a larger number of passive level crossings across the rail network [20]. This approach would increase the overall safety of the railway line by addressing more crossings. Therefore, prioritization of interventions should consider the balance between cost-effectiveness and the potential for reducing risks across the whole network. Traditional warning systems, including flashing lights and gates, have demonstrated limitations in effectively preventing accidents as accidents at railway level crossings continue to occur due to a combination of human error, insufficient protective measures, and inadequate integration of traffic management systems [5,22,23]. Therefore, there is an urgent need to explore innovative technological solutions to improve safety.

In this context, implementing low-cost rail level crossing warning devices is related to adopting innovative technologies for train detection and communication systems. These devices often rely on wireless sensor networks, which use ground-based sensors to detect approaching trains and transmit signals to warning systems installed at crossings [21,24]. GPS-based train detection systems, which enable more precise and reliable activation of warning systems, can also be used.

Enhancing Intelligent Transport System (ITS) architecture for integrating railroad operating systems with traffic management systems has been discussed in many recent studies [25,26,27,28,29]. The efficiency of such methods adapts to the ability to provide motorists with advanced warnings of approaching trains through interconnected information systems that integrate traffic management and rail operation networks. By leveraging these technological advancements, there is a significant potential for safety improvement in railway level crossings [30].

Preemption of traffic signalization near railway level crossings involves the interconnection of the railway warning time system and the highway or roadway traffic signal system (Figure 1) [31]. The main purpose of this interconnection is to ensure that no motorists occupy the railway tracks at the time of a train’s arrival at the crossing and to prevent any movements towards the track.

This paper has two main objectives:

• To examine traffic signal preemption as a cost-effective countermeasure and practical alternative to grade separation to mitigate the risk of crashes;
• To develop a microsimulation traffic model with PTV VISSIM to analyze and evaluate the anticipated safety and operational benefits of an existing railway level crossing and the adjacent road intersection.

The novelty of this study lies in its ability to provide valuable insights into signal preemption strategies at railway crossings in urban areas with limited queue storage as a cost-effective and practical alternative to grade separation. Specifically, the innovative aspects of this research are reflected in both its methodological framework and the practical implications of its findings. It addresses the challenges posed by increased motorized traffic at railway level crossings near adjacent road intersections, a topic often overlooked in prior studies. This analysis advances current knowledge by evaluating preemption strategies employing a microsimulation approach and considering the interaction between railway crossings and adjacent road intersections, thereby addressing a critical gap in the existing body of research.

The remainder of this study is organized as follows: Section 2 provides a comprehensive review of the relevant existing literature and research related to railway level crossing safety. Section 3 outlines the methodological framework used to collect data from an existing railway level crossing and simulate both current and alternative scenarios. The microsimulation model facilitated the testing of traffic and train scenarios, allowing for an in-depth analysis of how the proposed countermeasures perform under varying conditions. Section 4 presents the results of these simulations, including quantitative analysis and comparative evaluations. Finally, Section 5 summarizes the key findings and their implications for improving railway level crossing safety.

2. Background

2.1. Railway Level Crossings and Safety Issues

Fatalities resulting from crashes at railway level crossings (RLCs) account for a relatively small proportion of all road-related fatalities, comprising approximately 1%. However, they represent a significant percentage of railway-related fatalities (approximately 30%) [1].

This discrepancy underscores the urgent necessity for focused safety measures at RLCs since these locations provide a distinct and significant hazard where road and rail systems cross [32]. Despite technological breakthroughs and safety measures, human error, including traffic rule violations, distractions, or intentional risky behavior, continues to be a primary cause of incidents [33,34,35]. Addressing these challenges necessitates a comprehensive strategy that includes engineering solutions, public awareness initiatives, enforcement of traffic regulations, and new technologies such as automated warning mechanisms, user-side safeguards, and rail-side signaling preemption to guarantee the safety of both road and rail users [36,37,38,39,40,41,42].

Figure 2 illustrates the trends in railway level crossing accidents, serious injuries, and fatalities in the European Union (EU) over the past 15 years (2009–2023). This analysis is based on data from the European Union Agency for Railways (ERA), utilizing the Common Safety Indicators framework [43]. Following a period of improvement in the number of level crossing accidents and associated fatalities during the later years of the previous decade and the early years of the last decade (2009–2016, apart from 2012), the subsequent period (2017–2022) exhibited a more stagnant trend. This stagnation occurred despite a noticeable decline in 2020, likely attributable to the reduced mobility and transportation activity caused by the COVID-19 pandemic.

In total, 224 fatalities, 176 serious injuries, and 399 accidents occurred during 2023 in the EU. This translates to an average of almost one fatality per day in Europe. The analysis of accident and fatality data underscores and confirms that railway level crossings are high-risk locations requiring significant attention from a safety perspective.

According to [2], the EU Member States with the lowest accident rates have generally implemented comprehensive strategies to enhance safety at level crossings. As a result, there are fewer crossings with insufficient or no protective measures in these countries. On the other hand, nations with a low population density and low railway traffic volumes, such as Greece (EL) and Croatia (HR), result in high accident rates on railway level crossings [2]. This is related to the reduced motivation for the thorough management of level crossing safety.

Railway level crossings (RLCs) can be classified into the following two categories [44]: (a) passive level crossings and (b) active level crossings. In the first case, passive level crossings are equipped with only warning signs and road markings, intending to encourage road users to reduce speed, yield, or stop. On the other hand, active level crossings are additionally equipped with barriers (half or full) or gates and warning systems, such as flashing lights for the case of visible devices and horns for the case of audible devices.

Furthermore, an active level crossing can be further classified into four main categories: (a) rail-side protected level crossing, (b) automatic level crossing with user-side protection, (c) automatic level crossing with user-side warning, and (d) manual level crossing [45]. In the case of automatic level crossings, the road users are warned or protected by the activation of devices from the approaching train. In the case of the manual active level crossing, the railway employee manually activates the warning or protection.

In 2023, European (EU) countries reported approximately 93,000 level crossings. As is presented in Figure 3, the majority (40%) belong to passive level crossings, which do not provide any active warning to road users. Level crossings that provide full user-side protection and warning and rail-side protection represent 17% of total cases. The remaining 43% are distributed among three categories: automatic level crossing with user-side protection (26%), automatic level crossings with user-side warning (10%), and manual level crossings, which represent only 7%.

The distribution of these level crossings highlights the varying levels of safety features and operational designs across the EU [46]. According to the proportion of crossings equipped with active warning or protection systems, it is evident that significant disparities exist in the adoption of advanced safety technologies, which could directly impact the safety outcomes at these critical intersections.

Delving into the distribution of accidents by level crossing type in 2023, it is evident that 71% of accidents were almost equally divided between automatic level crossings with user-side protection and passive level crossings (Figure 4). In contrast, level crossings that provide full user-side protection and warning, along with rail-side protection, accounted for only 3% of the total accidents. This notably low percentage underscores the significant safety improvements associated with these fully protected level crossings, highlighting the effectiveness of advanced safety measures in reducing accident risk.

2.2. Addressing the Safety at RLCs near Road Intersections with Traffic Signal Preemption

Despite ongoing efforts to improve safety, passive level crossings remain prevalent, posing higher risks to road users due to the absence of active warning systems. The most effective approach to increase safety is to remove the RLCs or replace them with overpasses or underpasses. However, the implementation of these solutions often proves to be prohibitively expensive, especially considering the incredibly high number of existing total railway level crossings in the EU [17].

Moreover, while automatic barriers at level crossings can mitigate the likelihood of unintentional errors by road users, the issue of deliberate violations and specific errors, such as vehicles becoming stranded on the tracks, continues to contribute significantly to the occurrence of accidents, as reported above [47]. To enhance safety at railway level crossings (RLCs), several studies advocate for the integration of advanced technologies and proactive risk management strategies [48,49,50]. Among these strategies, traffic signal preemption is extensively utilized at intersections near railway level crossings to mitigate the risk of accidents by ensuring that vehicles are cleared from the tracks before the arrival of a train.

Several methodologies have been proposed and applied to evaluate safety at railway level crossings (RLCs), considering traffic signal preemption [31,51,52,53]. Chen and Rilett [51] developed a simulation-based optimization system to improve traffic signal operations at intersections near railway level crossings. Their methodology employed sophisticated algorithms to enhance signal timing, concentrating on minimizing vehicle delays and increasing overall safety. Chen et al. [52] developed a transition preemption strategy algorithm to address safety and delay issues at intersections near railway level crossings caused by the limitations of standard preemption. They applied their methodology to a dual-track corridor, and the results showed that the algorithm improved pedestrian safety and reduced vehicle delays.

Signal preemption was also studied in [53]. The authors developed an adaptive traffic control system for tram intersections, prioritizing both trams and emergency vehicles using transit signal priority techniques and passing vehicle search algorithms. Urbanik and Tanaka [31] conducted an extensive review of traffic signal preemption at intersections near highway-rail grade crossings (HRGCs). They found that despite the availability of advanced systems, most US and Canadian agencies still relied on outdated two-wire preempt systems, with 55% using these basic systems. Case studies from Portland, Ohio, and California highlighted the need for updated guidelines and improved coordination between highway and railway agencies to optimize safety and mobility. They recommended addressing operational limitations, enhancing training, and improving inspections with performance measures.

Recent studies have also utilized advanced technologies, such as driving simulators and traffic simulation software, to assess the impact of signal preemption on safety and traffic conditions [54,55,56,57,58]. Preemption during peak hours is thoroughly examined in [59], focusing on safety and mobility issues at railway level crossings and adjacent road intersections in urban areas. The study highlighted the limitations of traditional preemption solutions and explored advanced traffic signal systems using VISSIM-based simulations. Results showed the coordinated strategy effectively reduced delays, stops, and queue lengths, demonstrating its potential to enhance safety and mobility. Tey et al. [55] used VISSIM to compare the safety performance of different railway level crossing warning systems based on field video data. Results showed that active warning systems were at least 17% safer than passive signs, with surrogate measures identifying the most efficient safety approach for varying traffic and train volumes.

3. Materials and Methods

3.1. Examined Study Area and Problem Statement

A railway level crossing located in the urban area of Larissa, Greece, was selected for this study. This railway level crossing is part of the main railway axis in Greece, which begins in Piraeus, traverses the country, and extends northward to Thessaloniki. As a result, a substantial number of trains operate on this railway line to accommodate both passenger and freight demand.

As is presented in Figure 5, the examined railway level crossing is an active manual level crossing, where user-side protection is manually activated by a railway employee. More specifically, the users are warned of a train’s approach through the closure of barriers. Moreover, a crossbuck sign is placed at the crossing and indicates that vehicles should yield to trains approaching or present on the tracks. The sign specifies the existence of more than one track.

In the study area, a road intersection is located 170 m from the railway level crossing, as is presented in Figure 6. This four-legged road intersection primarily facilitates traffic from the east suburban area to the city’s central business district (CBD). Furthermore, during the summer months, a significant volume of road users pass through this intersection to travel to and from coastal destinations, leading to significant delays in travel time for road users, queues, and an increase in pollutant emissions. Therefore, during peak hours, the limited queue space between the road intersection and the railway level crossing may result in queued vehicles blocking the railway level crossing or road intersection, hence increasing the risk of accidents.

3.2. Field Data Collection

To ensure the accurate collection of traffic data under saturated conditions, initial field observations were scheduled during three distinct time periods: morning peak (08:00–09:00), afternoon (14:30–15:30), and evening (20:00–21:00) on a typical weekday. These time slots were determined based on a preliminary analysis of traffic delays derived from Google Maps API over the course of a tested month. Based on the preliminary results, the afternoon peak hour (14:30–15:30) was identified as the period with the highest traffic saturation, especially for the traffic flow that approaches the road intersection through the railway level crossing and was consequently selected for conducting the traffic surveys.

Field data were collected during the evening peak hour over a week under dry weather conditions. A video camera was installed atop the nearest building to capture the full range of movements and record videos maintaining 3840 × 2160 resolution at 60 frames per second, providing high-definition clarity with smooth motion. This approach ensured that no personal data were collected, maintaining compliance with data privacy standards. It is noted that pedestrian movements were not considered in the present analysis as they were out of the scope of the paper.

The following data were collected from the field surveys:

• Traffic volumes (in PCU/h);
• Origin–Destination (OD) matrices;
• Travel time (min);
• Queues (meters);
• Traffic signal timing plan.

3.3. Simulation Analysis with VISSIM

3.3.1. Simulation Model

The microscopic traffic simulation model PTV VISSIM [60] was employed to simulate the traffic conditions for both the current situation and the alternative configuration. All required modes of transport were included in the model (passenger cars, trucks, buses, and trains), while pedestrian movements were not considered in the present analysis as it was out of the scope of the paper.

To create a reliable transport network, the geometric features of the examined study area were obtained from Google Earth [61] and additionally measured on-field to properly adjust any discrepancies or inaccuracies in the data. Other operational characteristics, except for road traffic volumes and turning movements, such as speed limits and train passing times, were identified on the field.

A total of 10 runs were conducted for each scenario. The total duration for each simulation scenario was set to 4800 s. In addition, a 900-s warm-up period to account for the initial under-saturation of the model and a 300-s recovery period, were included. Consequently, the simulation model’s analysis period was set to 3600 s, corresponding to the AM peak hour. The influence of train frequency on safety and mobility analysis at railway level crossings has been extensively examined in the literature [52,54,59]. Zhao et al. [54] emphasize that sensitivity analyses of train volumes can provide deeper insights into the safety and mobility impacts of railway operations. Their study incorporated six train arrivals during a one-hour analysis period. Similarly, a study by other researchers [59] expanded the scope of train frequency testing by increasing the number of observed trains from two per hour to five per hour, corresponding to a 12 min interval between arrivals. Tey et al. [55] further investigated railway level crossing safety through microsimulation analysis, assuming a train frequency of one every 30 min. In alignment with the methodological frameworks and assumptions applied in these studies, the present analysis considers a train frequency of one every 15 min for the railway level crossing under investigation.

3.3.2. Simulation Scenarios and Design

Interconnected traffic signals near railway level crossings equipped with active warning systems have been in use for many years [62,63,64]. These systems override normal traffic signal operations when a train is approaching, ensuring safety for road users. Hence, traffic signal preemption at road intersections near RLCs is implemented to clear any vehicles at risk of being struck by an approaching train.

According to the Manual of Uniform Traffic Control Devices (MUTCD) [65], preemption should be considered in cases where the railway level crossing is located within 61 m (200 feet) of an intersection. However, it is argued [66] that this distance should not be strictly confined to these limits and should instead be adjusted according to the specific circumstances of each case. More specifically, it is suggested that a queuing analysis can determine if signal preemption is required [67]. Therefore, preemption should be considered at all intersections which are affected by the queues extending from the railway level crossing.

The proposed preemption strategy was implemented in the simulation model and integrated as a module within VAP logic, an add-on component of VISSIM designed for advanced signal control [60]. The efficient determination of phases, sequences, and timing of the traffic signal heads for various road users, such as road vehicles, pedestrians, and trains, is critical for successful preemption implementation.

The main challenge is the connection and coordination of the signal controller at the RLC and the road intersection. According to the MUTCD [65], a minimum warning time of 20 s is required before the train approaches and occupies the rail crossing. Therefore, the minimum distance of the train detector installation from the RLC, considering a train average speed of 80 kph at the specific location, would be 445 m. Moreover, this distance should be longer than the distance between the RLC and the road intersection (190 m). This ensures that vehicles queued at the intersection have sufficient space to clear the RLC entirely, avoiding potential obstructions or hazardous situations on the tracks. Therefore, 450 m from the RLC was selected for the train detector installation.

Figure 7 depicts the schematic design of (Figure 7a) the existing situation and (Figure 7b) the railway preemption actions which were implemented to be examined in the alternative scenario.

The existing situation was calibrated to ensure that the developed simulation model replicates local traffic conditions and, therefore, achieves sufficient reliability. The calibration of the model involved the improvement of the geometric features of the simulated scenario, the verification of the signal timing, and fine-tuning the Wiedemann 74 car-following model to accurately reflect real-world diving behavior.

Traffic volumes and travel times were the two calibration criteria that were selected to ensure the validation of the model. More specifically, according to the FHWA [68], at least 85% of the total links (turning movements) should closely match field conditions. The GEH statistic, defined by Equation (1) [69], was calculated to compare the field traffic counts with the traffic volumes estimated by the model. Following the recommendations of the FHWA [68], the GEH statistic for individual link flows was less than 5 in more than 85% of the cases.

$$\mathrm{GEH}=\sqrt{{\displaystyle \frac{{(\mathrm{E}-\mathrm{V})}^2}{(\mathrm{E}+\mathrm{V})/2}}}$$

(1)

where

• E: model-estimated volume in vehicles/hour;
• V: field-measured volume in vehicles/hour.

Additionally, the differences in total travel times per trip, as calculated in the simulation model, were within 5% of the observed field values. Hence, the simulation model for the existing situation showed that all calibration criteria were met.

Table 1. Comparison of observed and simulated values.

Trip	Traffic Volume (veh/h)			Travel Time (s)
	Observed Value	Simulated Value	GEH Statistic	Observed Value	Simulated Value	Relative Error
1–2	1560	1693	3.29	64.2	66.5	3.6%
1–3	650	662	0.45	55.9	56.7	1.6%
1–4	900	857	1.45	120.2	114.4	4.8%
2–1	1820	1769	1.20	64.2	62.4	2.8%
2–3	270	275	0.30	78.9	80.3	1.8%
2–4	1040	1168	3.85	97.2	99.4	2.3%
4–1	510	529	0.83	131.7	137.9	4.7%
4–3	120	94	2.55	132.5	131.4	0.8%
4–2	780	814	1.19	138.0	141.2	2.3%

presents the GEH statistic and compares the observed and modeled travel time for each trip. It is noted that the trip numbering is consistent with that shown in Figure 7.

Following the calibration and verification of the model in the existing situation, the alternative scenario, which included the railway preemption actions, was simulated to evaluate its effectiveness in enhancing safety.

Figure 8 presents the alternative scenario as simulated in PTV VISSIM.

4. Results

According to the Highway–Rail Crossing Handbook [70], to better understand and measure the efficiency of the concept of preemption, the Clear Storage Distance (CSD) and the Minimum Track Clearance Distance (MTCD) were used. More specifically, the Clear Storage Distance (CSD) refers to the segment of roadway between the RLC and the road intersection that provides sufficient space for vehicles to queue safely. It is measured 1.8 m from the rail to the intersection stop line. In contrast, the Minimum Track Clearance Distance (MTCD) denotes the critical area that must remain free of vehicles to prevent potential collisions. The MTCD is measured from the portion of the gate arm farthest from the near rail, in case flashing-light signals are used concurrently with automatic gates, and ends 1.8 m beyond the tracks.

Two performance measures, (a) queue lengths and (b) maximum queue length, were used to evaluate the operation and safety performance of various configurations and scenarios at RLCs. These metrics provide a comprehensive understanding of how traffic responds to changes in design, signalization, and control mechanisms. All performance metrics were applied to all approaches, with a particular focus on the clearance lane, to evaluate the overall operation of the road network and the safety impact. Specifically, when a train preempts the traffic signal, the queue in the clearance lane is expected to be cleared before the train reaches the railway level crossing (RLC).

Figure 9 illustrates and assigns numbers to the various locations of the performance indicators used in the analysis. The clear storage distance and the minimum track clearance distance are also identified.

More specifically, queue lengths were analyzed to determine the average accumulation of vehicles during peak hours and train passings, while maximum queue lengths assessed the worst-case scenarios for traffic congestion. These data are critical for understanding the potential for spillback into the adjacent intersection.

The number of stops within the queue reflects the interruptions experienced by drivers and highlights areas where smoother traffic flow can be achieved. Frequent stops indicate inefficiencies in signal timing or inadequate clear storage and track clearance distances.

Performance metrics were calculated per one-second interval in the simulation output to gather data for the whole duration of the signal preemption.

Table 2. Summary of the queue statistics at a one hour average for all runs.

Counter No	Existing Situation		Railway Preemption Actions
	Av. Queue Length	Max Queue Length	Av. Queue Length	Max Queue Length
1	105	170	0	0
2	95	180	150	230
3	87	130	72	110
4	270	350	230	300
5	65	102	54	95

provides a summary of the queue statistics at a one-hour average for all runs. It is noted that each counter number corresponds to the one shown in Figure 9.

As shown, the largest differences in queue lengths are observed at Counter 1. When a train preempts the traffic signal, the queue in the clearance lane is expected to dissipate before the train arrives at the railway level crossing. Under the existing conditions, during a train’s passage, the maximum observed queue length corresponds to the distance between the railway level crossing and the adjacent road intersection. This indicates that when the employee manually activates the protective barriers, the road traffic forms queues that obstruct critical turning movements at the intersection, exacerbating traffic congestion.

Conversely, under the signal preemption scenario, sufficient time is allocated for the road traffic to stop at a red light, preventing them from exacerbating traffic conditions or being placed in potentially hazardous situations as they approach the railway tracks. This highlights the effectiveness of signal preemption strategies in mitigating congestion and enhancing safety near railway level crossings.

Furthermore, at the next critical counter, Counter 4, the existing conditions revealed a significant queue extending from the road intersection to several meters beyond the railway level crossing. This indicates that road traffic is exposed to considerable risk, as the arrival of the train and the activation of the barriers could find a substantial number of vehicles occupying both the area before and after the railway crossing, thereby obstructing the entire clear storage distance.

In contrast, under the alternative scenario involving traffic signal preemption, shorter queues were observed, although the situation remains critical. The key distinction in this scenario was that when the traffic signal is preempted by an approaching train, adequate time is provided for vehicles to clear the covered storage distance, thereby enhancing safety (see Figure 10).

As a result, the queue length at Counter 2, located before the railway level crossing, was expected to be larger in the preemption scenario compared to the existing situation, as more vehicles become queued following the initiation of the preemption. For the remaining queue lengths, particularly at Counters 3 and 5, the alternative scenario demonstrated improved performance in both maximum and average queue lengths. This improvement can be primarily attributed to the implementation of traffic signal preemption, which allows the right-of-way to be reallocated from conflicting movements when active, thereby providing additional time for other traffic flows.

Figure 10 depicts the variation in clearance lane queue lengths over the simulation period for the traffic signal preemption scenario. The boxes represent the queue length measured from the stop line, with blue boxes corresponding to normal traffic conditions and the red box indicating the queue length during the train’s arrival at the railway level crossing.

As anticipated and previously described, the queue length during traffic signal preemption was minimized, remaining below 90 m. This indicates that the likelihood of conflict between road and rail traffic was effectively eliminated, as less than 50% of the clear storage distance was occupied even during congested periods. This ensures a balance in mitigating potential conflicts while minimizing the overall impact on traffic delays for other turning movements within the road network of the study area.

This concept plays a crucial role in traffic and infrastructure design, as it helps prevent vehicles from becoming trapped on railway tracks when a train approaches, thereby reducing the risk of collisions and other safety hazards. The results suggested that the preemption plan at railway level crossings yields positive outcomes in enhancing both safety and traffic flow. The queue length metric is a key safety performance indicator for grade railway level crossings concerning signal preemption strategies. Extended queues that spill over onto the road intersection are high-risk situations, as they increase the chances of vehicles encroaching on the train’s path, potentially resulting in collisions.

5. Conclusions

Railway level crossings (RLCs) continue to be an important concern for safety and operational issues. Although they constitute a minor percentage of total road fatalities, railway level crossings account for an excessively high number of railway-related accidents and fatalities, with considerable consequences on society’s economy. Despite the use of modern technology and infrastructure designs, accidents at railway level crossings continue to occur due to a combination of human error, insufficient protective measures, and inadequate integration of traffic management systems.

This study provided a detailed analysis of the safety and operational performance of railway level crossings using a microsimulation VISSIM model and focused on the implementation of signal preemption strategies as a cost-effective and practical alternative to grade separation. This research focused on the safety and operational inefficiencies at RLCs, especially those located in urban areas and near road intersections characterized by high traffic volumes and limited queue storage distances.

According to the results, the signal preemption significantly reduced the likelihood of vehicle spillback into critical zones, particularly the clear storage distance (CSD) and minimum track clearance distance (MTCD), thereby preventing hazardous situations where vehicles could become trapped on railway tracks and minimizing potential conflicts and safety risks. Concurrently, traffic signal preemption minimized the overall impact on traffic delays for most of the turning movements within the road network of the study area.

The novelty of this study lies in its ability to provide valuable insights into signal preemption strategies at railway crossings in urban areas with limited queue storage. The innovative aspects of this research are reflected in both its methodological framework and the practical implications of its findings. It addresses the challenges posed by increased motorized traffic at railway level crossings near adjacent road intersections, a topic often overlooked in prior studies. The proposed preemption strategy presents an economically appealing alternative to prohibitively costly grade separation solutions, allowing for scalability across multiple railway level crossings, especially in budget-constrained areas.

The findings are directly applicable to urban areas where limited queue storage distances and high traffic volumes pose persistent safety challenges. Moreover, this study highlights several policy implications for improving safety and efficiency at railway level crossings. Even if the Manual of Uniform Traffic Control Devices (MUTCD) [65] proposes that preemption should be considered in cases where the railway level crossing is located within 61 m (200 feet) of an intersection, it is suggested that a queuing analysis can determine if signal preemption is required [59]. Therefore, preemption should be considered at all intersections which are affected by the queues extending from the railway level crossing. Finally, passive railway level crossings with high accident rates should be targeted for upgrades to include active warning and preemption systems.

Despite the promising findings, this study has several limitations that warrant consideration. The simulation analysis was based on data collected from a single case study, which may limit the generalizability of the results to other locations with differing traffic conditions and railway configurations. Moreover, pedestrian movements were not included in the microsimulation model, even though they can play a significant role in the safety and efficiency of railway level crossings, particularly in urban areas.

Building on the findings and limitations of this study, future research could examine more cases of railway level crossings, including various types of adjacent road intersections (such as roundabouts). Moreover, an incremental approach for traffic volumes (both motorized and rail traffic) could help identify the critical threshold at which the signal preemption concept achieves optimal performance. By systematically increasing traffic volumes in the simulation, it would be possible to pinpoint the saturation point where queue lengths, delays, and safety risks begin to escalate significantly. This analysis would provide valuable insights into the capacity limits of the preemption system.

By leveraging microsimulation and real-world data, the present research contributes to the broader goal of sustainable and safer transport systems, aligning with the EU’s Vision Zero initiative to eliminate road fatalities. The findings underscore the importance of data-driven decision-making in infrastructure planning and highlight the need for continued innovation and investment in railway level crossing safety.