category: literaturenote citekey: valdesdrivingsimulatoranalysisfirst2025 title: Driving simulator analysis of the first time driving on a diverging diamond interchange authors: "Valdés, Didier M.; Figueroa-Medina, Alberto; Perelló, Carol A." year: 2025 date: 2025-00-00 2025 doi: 10.1016/j.aap.2025.108084 publication: "Accident Analysis & Prevention" url: "https://doi.org/10.1016/j.aap.2025.108084" zotero_key: 9B8CPCK6 zotero_storage: RQ2QA8BZ collections: transpordi_planeerimine folder: 001_artiklid firstAuthor: "Valdés, Didier M."

status: converted

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

Driving simulator analysis of the first time driving on a diverging diamond interchange☆

Didier M. Vald´es * , Alberto Figueroa-Medina , Carol A. Perello´

Department of Civil Engineering and Surveying, University of Puerto Rico at Mayagüez, Call Box 9000, Mayagüez, PR 00681-9000, USA

ARTICLE INFO

Keywords: Diverging diamond intersection Driving simulation Driver behavior Design assessment

ABSTRACT

The Diverging Diamond Interchange (DDI) is an innovative design alternative that enhances safety and operations, but its unconventional crossover maneuvers might confuse first-time drivers. The first DDI in Puerto Rico opened in 2023 at the interchange of highways PR-30 and PR-189 to improve the level of service of a conventional diamond interchange. Due to the lack of experience with a DDI in Puerto Rico, a driving simulation experiment was conducted in three phases to analyze future user behavior and performance. The simulation was conducted with 96 subjects who completed eight driving scenarios representing different maneuvers at the DDI and provided feedback about the effectiveness of the traffic control devices (TCDs) in assisting the driving task. Over 2,600 individuals also participated in training sessions with the simulator to experience the DDI. Speed profiles and vehicle trajectories were analyzed to reveal first-time drivers' challenges at the DDI, such as confusion with the crossover maneuvers, speeding, wrong route selection, and incorrect maneuvers. The results from the simulation and the survey responses assisted in identifying recommendations for modifications to TCDs and the DDI characteristics. Subsequent simulation phases served to evaluate the impacts of the implemented changes, resulting in a 47% reduction in the number of critical driver errors and a 26% reduction in total driver errors at the DDI. This study demonstrates the significant role of a driving simulator during the design process of innovative facilities like a DDI, aiding in assessing proposed design features and identifying effective improvements to TCDs.

1. Introduction

Aiming to address traffic congestion and safety concerns, the Puerto Rico Highways and Transportation Authority (PRHTA) replaced the Conventional Diamond Interchange (CDI) at the intersection of highways PR-30 and PR-189 in the Municipality of Gurabo with the first Diverging Diamond Interchange (DDI) in Puerto Rico. The road project, completed in the summer of 2023, involved the reconstruction of the two bridges of the PR-30 Expressway and the realignment of the underneath travel lanes on highway PR-189 to incorporate the DDI. The PR-30 is a six-lane access-controlled expressway, with an AADT of 95,452 vehicles per day (vpd) in 2014 and a 55-mph speed limit. PR-189 is a four-lane divided suburban highway with an AADT of 45,679 vpd in 2014 and a 40-mph speed limit. A particular characteristic of PR-189 is a sharp horizontal curve in the north access prior to the interchange. Fig. 1 shows the intersection site and the general roadway characteristics prior to the construction of the DDI.

Round 2 of the Every Day Counts (EDC) initiative of the Federal Highway Administration (FHWA) promoted the use of the DDI, also known as a Double Cross Diamond (DCD). The difference between the DDI and a conventional intersection lies in its navigational requirements, which require users to drive temporarily on the left side. This geometric innovation eases left turns onto freeway on-ramps, eliminating the need for conflicting left turns across opposing traffic lanes (MODOT, 2010). EDC advocated for alternative intersection designs, like the DDI, to improve safety and reduce delays at intersections,

Abbreviations: AADT, Annual Average Daily Traffic; CDI, Conventional Diamond Interchange; DDI, Diverging Diamond Interchange; DCD, Double Cross Diamond; EDC, Every Day Counts; EB, Eastbound; FHWA, Federal Highway Administration; IRB, Institutional Review Board; MUTCD, Manual on Uniform Traffic Control Devices; NB, Northbound; NCHRP, National Cooperative Highway Research Program; PRHTA, Puerto Rico Highways and Transportation Authority; SB, Southbound; SCL, Speed Change Lanes; TCD, Traffic Control Devices; UPRM, University of Puerto Rico at Mayagüez; U.S., United States; vpd, Vehicles per day; WB, Westbound.

E-mail addresses: didier.valdes@upr.edu (D.M. Vald´es), alberto.figueroa3@upr.edu (A. Figueroa-Medina), carol.perello@upr.edu (C.A. Perello). ´

^☆ This article is part of a special issue entitled: 'RSS 2024' published in Accident Analysis and Prevention.

^* Corresponding author.

achieving these goals at lower costs and with less environmental impacts than traditional solutions (FHWA, 2014). The National Cooperative Highway Research Program (NCHRP) Report 959 indicates that the DDI, as a design alternative to a Conventional Diamond Interchange (CDI), improves traffic flow and reduces the number of conflict points from 26 to 14 (NASEM, 2021).

The first DDI in the United States (U.S.) opened in 2009 at the Kansas Expressway (MO-13) intersection and Interstate 44 in Springfield, Missouri. Before the DDI, the southbound peak traffic on MO-13 experienced significant congestion, with queues stretching up to 2 miles. The DDI addressed this challenge, demonstrably improving traffic flow. Following its implementation, the DDI concept has been adopted at other highway locations across the U.S. Further recognition came in 2009 when Popular Science Magazine lauded the DDI design as one of the year's best engineering innovations (MODOT, 2010). Claros et al. (2015) evaluated the Springfield DDI one year after its opening and found a significant 46 % reduction in crashes. Additionally, the study revealed that over 80 % of road users perceived a decrease in traffic delays with the new DDI configuration.

While research has proven the effectiveness of the DDI in enhancing safety at the interchange and its ramp terminals, the impact on nearby intersections and Speed Change Lanes (SCLs) remains to be decided. Claros et al. (2017) examined crash data from twelve intersections next to DDI ramps in Missouri. The analysis, which included 32 SCL zones, 16 highway entrances, and 16 exits, revealed no statistically significant evidence that the DDI design either increased or decreased crash frequency at these locations. The study confirmed that DDI ramp terminals showed improved safety compared to traditional diamond interchanges.

Human factors research is crucial in assessing the interaction between road users and novel intersection designs like the DDI. FHWA completed a driving simulation study of the first DDI built in the U.S. to assess driver performance when encountering the unfamiliar intersection design. While the study found an initial tendency for drivers to disregard signage and markings, persisting in right-side driving behavior at the crossover section, it also yielded promising findings. In conjunction with speed reduction strategies, the DDI design appeared to contribute to lower severity crashes when compared to traditional diamond intersections (FHWA, 2007). Such findings highlight the significance of human factors research, particularly the use of driving simulation, in proactively showing potential safety concerns and evaluating the effectiveness of design features for innovative intersection designs such as the DDI.

Educating novice learners about novel concepts requires meticulous attention (Khalil et al., 2010). Pham (1998) advocates for a multifaceted approach, employing a combination of text, video, audio, images, multimedia, and even simulations. The multifaceted approach reduces the burden of ineffective cognitive load, promoting deeper understanding and knowledge retention. Supporting this notion, Khalil et al. (2010) showed that innovative computer-based instructional strategies,

aligned with cognitive load theory, are more effective than traditional paper-based methods in reducing cognitive overload. Incorporating educational and outreach activities that use simulation tools like driving simulations can enhance a driver's effective cognitive load. The experience gained through simulation translates into practical knowledge, better equipping drivers to navigate unfamiliar intersection designs when encountering them for the first time.

Driving simulation offers a valuable and cost-effective tool for analyzing novel safety and operational measures in transportation systems. It enables researchers to assess driver decision-making processes and measure performance metrics such as vehicle acceleration, lateral positioning, turning maneuvers, and chosen speeds (Fisher et al., 2011). The simulation data offers crucial insights into the interplay between human factors and road safety. By incorporating geometric features and operational models within simulated scenarios, driving simulation allows for replicating real-world driver behavior, offering a comprehensive evaluation environment.

To evaluate driver behavior and potential safety improvements associated with the proposed DDI design for the PR-30/PR-189 intersection, the PRHTA commissioned the University of Puerto Rico at Mayagüez (UPRM) to conduct a driving simulation study with local participants focusing on the proposed DDI configuration. The term "first drivers" is a general reference for participants who were unfamiliar with DDI intersections, as this was their first experience driving on this type of roadway. This paper discusses the method followed in the driving simulation experiment and presents major findings about the modifications made to the Gurabo DDI.

2. Methodology

Considering the introduction of novel designs with maneuvers significantly different from those encountered in conventional designs, the initial phase of the study involved adapting the general road project design and implementation methodology. While following the established sequence of planning, conceptual design, detailed design, and construction of a road project development process, this study incorporated the iterative testing procedure using a driving simulator proposed by Valdes et al. (2019), as shown in Fig. 2. The iterative process entailed assessing the final design of the DDI through a driving simulator experiment with a sample composed of local drivers. Driver behavior at the DDI was observed, and performance data was collected, including adherence to designated routes, speed variations, and acceleration patterns within the simulated intersection.

The observed behavior and the data collected with the simulator were analyzed to identify recommendations for possible modifications to the DDI design and traffic control devices to the PRHTA designers. The modifications incorporated into the DDI design were retested in the driving simulation scenarios in a subsequent round, allowing researchers to observe potential changes in driver performance. This

Fig. 1. Location of the expressway PR-30 and highway PR-189 intersection (Modified from Google Earth, 2020).

Fig. 2. Proposed method for integrating a driving simulator into the project development process. Taken from Valdes et al. (2019)

iterative process was repeated, enabling the progressive refinement of the DDI design until a balanced and optimized outcome was achieved. Three simulation phases were conducted for this project using a total of 96 participants. The sample of participants for each experiment phase was composed of different individuals.

Phase 1 employed various approaches to assess driver behavior and perceptions regarding the DDI design. This phase incorporated a driving simulation experiment with 48 subjects to observe their interactions with the DDI configuration. Also, a survey was conducted to measure the participants' perceptions of the driving experience and the effectiveness of the traffic control devices. Educational activities were also conducted to familiarize participants with the new intersection configuration. These activities included a short video, a brochure, and training sessions with the UPRM Driving Simulator. The simulator was strategically located at the university adjacent to the construction site to provide the opportunity to gather feedback from drivers accustomed to navigating the existing intersection. The results of the experiment's first phase provided potential trends in driver performance and behavior. The UPRM research team provided recommendations to PRHTA designers based on these results. The PRHTA designers examined the suggestions for adding or modifying the proposed pavement markings and roadway signs to enhance operations and safety at the DDI.

Phase 2 consisted of a follow-up driving simulation study with 24 subjects to evaluate the revised DDI design at the PR-30 and PR-189 intersection. The revised design included modifications to traffic control devices (TCDs), such as reducing the amount of pavement markings used for the lane extensions at the crossovers, installing overhead route directional signs on the bridge and before the DDI, and adjusting the pavement markings in the channelizing islands to address the driving

issues identified in Phase 1. This data helped the comparative analysis of the driver responses between Phases 1 and 2, enabling an assessment of the impacts on driver behavior and performance from the modifications made to the DDI. Phase 2 also included a comprehensive educational and training program for the community. This program focused on significant aspects of the DDI, such as the essential design elements, the operational differences compared to traditional CDIs in Puerto Rico, and the new driving rules associated with the implemented DDI. Participants practiced navigating the DDI using the UPRM Driving Simulator. Phase 2 of the driving simulation study was conducted in the summer of 2022 at Las Catalinas shopping mall, located approximately 3 miles from the PR-30/PR-189 site, to ensure accessibility to the local community of drivers.

Phase 2 concluded with a final report summarizing the results of the driving simulation study and the community engagement activities. Based on the findings and recommendations, PRHTA designers refined again the selection and location of road signs and pavement markings in the final design and implementation of the DDI. These modifications to the TCDs focused on making them more conspicuous and adding redundancy to ensure that the required operational and directional information was presented to the drivers. Additional signs were implemented to prevent drivers from going the wrong way and to inform people about the new traffic pattern ahead at the intersection.

Phase 3 focused on raising awareness of the DDI and analyzing the behavior of drivers based on the final design. The UPRM driving simulator was again used with scenarios of the DDI characteristics, including all the final changes to traffic control devices made by PRHTA designers. Phase 3 was conducted with 24 subjects during the final construction phase before opening the intersection to the public. The UPRM driving

simulator was used again as an educational tool in public campaigns conducted by PRHTA. The base scenario of the finished interchange was presented in various venues to raise awareness of the operational changes introduced at the intersection with the implementation of the DDI. The educational campaigns included showing the DDI scenarios with the UPRM driving simulator at four events: a meeting with the press at the Traffic Management Center located on the nearby PR-22 Freeway, on a national radio program, in a popular TV program, and at Plaza las Am´ericas, the largest shopping mall in Puerto Rico, located in the Municipality of San Juan.

3. Experimental procedure

The study developed a driving simulation experiment to assess the behavior and performance of local drivers when navigating the proposed DDI design. The experiment was conducted during construction and before opening the DDI to the public. Four research questions were addressed in the experiment:

1. How did drivers navigate the DDI during their first driving experience?
2. Did drivers experience confusion while driving in this unfamiliar interchange geometry?
3. Do the proposed pavement markings and roadway signs convey a clear message?
4. Are additional pavement markings and signs needed to improve the driving performance at the DDI?

The study employed a ten-step process to address the research questions. This process began with a thorough review of the existing literature and information about DDIs. Next, a base scenario was produced based on the proposed DDI geometric design and the local context, as shown in Fig. 3. The simulated scenarios incorporated the roadway geometric design elements, road signage, pavement markings, and other relevant physical features. The DDI on PR-189 was built underneath the two arc bridges on PR-30. The two roadways are separated vertically by 6 m (19.7 ft), with the travel lanes having a width of 3.65 m (12 ft). The DDI has an entering radius of 74.25 m (243.54 ft), and the radius in the interior curve between the crossovers is 76.65 m (251.41 ft). The length of the weaving area is approximately 145 m (475.72 ft), and the posted speed limit within the crossover zone is 25 mph. The signs and pavement markings strictly adhered to the MUTCD (FHWA, 2012), and their locations were based on the design drawings. The simulation incorporated land use characteristics and the surrounding context to enhance realism, such as the university campus and

residential, commercial, and industrial uses.

Following the scenario development, an experimental design was established. Data collection instruments were created, including questionnaires and the development of scenarios encompassing eight out of twelve possible maneuvers at the DDI. Each subject had to complete the eight simulation scenarios shown in Table 1: each one corresponding to a driving maneuver at the DDI. The order of the eight simulation runs was randomized for each subject to reduce a potential learning effect. The simulation runs had the subjects operating in free-flow conditions at the DDI. Each subject was given a training session with the driving simulator prior to the experimental procedure to become familiar with the controls and operations of the simulated vehicle. An IRB protocol was submitted and approved prior to conducting the research procedure.

Data was collected and analyzed simultaneously with the presentation of the intersection through the simulator. In addition to the driver behavior and performance data obtained with the simulator, participants were asked questions about driving comfort using the interchange. The vehicle trajectories, speeds selected, and driving decisions were analyzed to find possible conditions that may affect the operation and safety at the DDI. Following each phase, a report was prepared and sent to the PRHTA for evaluation and implementation. Additionally, community training and educational activities were conducted. The final stages involved data analysis to extract meaningful insights, followed by the formulation of conclusions and recommendations based on the findings.

4. Analysis and results

This section presents the analysis and key findings from the DDI driving simulation study across all phases. The performance measures examined included vehicle trajectories, speed profiles, and survey responses. The analysis focused on identifying the principal differences observed between phases and the resulting improvements in driver behavior.

4.1. Vehicles trajectories

The study analyzed the vehicle trajectories obtained from each simulation run. This analysis focused on whether participants followed each scenario's intended paths (trajectories), as detailed in Table 1. Building upon the work of Khattak et al. (2021), the researchers categorized any deviations from the intended path as either critical or noncritical. Critical errors were considered dangerous actions that could cause severe crashes, like going the wrong way at the crossover. Non-

Fig. 3. Plan view of the original design and final construction layout of the PR-30/PR-189 DDI.

Table 1 Driving errors made by the subjects per study phase.

| Scenarios | | Critical Errors
Phase | | | Non-Critical Errors
Phase | | | |-------------------------------|--|--------------------------|----|-----|------------------------------|----|-----| | | | | | | | | | | | | I | II | III | I | II | III | | 1. SB Through (yellow) | | 2 | 4 | 0 | 3 | 1 | 1 | | 2. NB Through (red) | | 4 | 1 | 1 | 1 | 2 | 1 | | 3. NB Left (green) | | 3 | 0 | 5 | 3 | 2 | 1 | | 4. SB Left (blue) | | 3 | 2 | 1 | 4 | 2 | 3 | | 5. WB Right (orange) | | 0 | 0 | 1 | 2 | 4 | 2 | | 6. EB Left (purple) | | 2 | 2 | 1 | 2 | 3 | 3 | | 7. SB Right (black) | | 5 | 2 | 1 | 1 | 3 | 2 | | 8. EB U-turn (purple + green) | | 0 | 1 | 0 | 7 | 11 | 8 | | Total | | 19 | 12 | 10 | 23 | 28 | 21 |

critical errors were considered less severe mistakes, like taking a different route (on the correct traveled way) or missing a required exit. Table 1 shows the number of mistakes that drivers made in each scenario during each study phase.

This section discusses the results from two of the scenarios, including the crossover maneuvers at the DDI. Fig. 4 illustrates the erratic maneuvers observed for Scenario 1 of the simulation, where participants navigated the DDI in the southbound (SB) through movement going through the two crossovers. Eleven out of the 96 participants exhibited erratic maneuvers across all study phases. These maneuvers were classified into non-critical and critical errors. Four participants committed non-critical errors by slightly deviating from the intended path. Six participants, however, made critical errors. Four participants drove on the wrong side of the road (right side), one veered off the path but corrected to the left side, and another reversed in the center lane to examine the signage. It is worth noting that during Phase 3, only one participant made a non-critical error by slightly deviating from the intended path.

Fig. 5 shows the erratic maneuvers observed in Scenario 2, where participants navigated the DDI in the northbound (NB) through direction going through the two crossover maneuvers. Across all phases, ten subjects made erratic maneuvers, with six considered critical and four considered non-critical. Four of the critical errors involved driving on the wrong side of the road (right side). One driver continued straight over the traffic island instead of following the designated crossover path. The fifth driver prematurely entered the PR-30 on-ramp heading eastbound (EB) before the first crossover, stopped completely, and then reversed to rejoin the PR-189 path. The sixth critical error involved a driver initially trying to drive on the wrong side but then stopped, reversed, and corrected to the left side. Non-critical errors included two drivers making left turns after the first crossover and two drivers making right turns before the first crossover.

4.2. Speed profiles

This section presents the results from the analysis of the speed profiles of participants who had non-erratic trajectories during the DDI navigation for Scenarios 1 and 2. The speed profile presents the DDI layout with designated maneuvers alongside graphs for the 15th percentile (orange line), average speed (light blue line), and the 85th percentile (dark blue line). At the top of the profile, the approximate locations of specific TCDs is shown as reference points of the DDI. The profiles from the different simulation phases illustrate how the introduction or modification of signs affected speeds at various points.

Fig. 6 shows the speed profiles for Scenario 1, in which drivers performed the southbound through movement and navigated both crossovers. Initially, drivers kept average speeds of around 38–39 mph at the initial reference points before experiencing a large speed reduction as they approached the first traffic signal at the DDI crossover. Phase 1 shows a reduction from 31 mph to 19 mph, while Phase 2 shows a drop from 33 mph to 22 mph, and Phase 3 shows a reduction from 38 mph to 28 mph. After navigating the first crossover, the speed of drivers gradually increased between the two signals and overhead signs, reaching around 28–30 mph in the latter segments. The speed reductions at the crossovers reflected the complexity of the maneuver and the need for driver adaptation.

Fig. 7 illustrates the speed profiles for Scenario 2 across all phases when drivers completed the northbound through path, encountering various traffic control devices. In Phase 1, despite the 15-mph school zone limit, average speeds were 35 mph and 27 mph at the first and second school signs, respectively, with only the drivers going at the 15th percentile speed complying with the posted limit. Average speeds

Fig. 4. Driving errors committed in scenario 1 for the southbound through movement.

Fig. 5. Driving errors committed in scenario 2 for the northbound through movement.

Fig. 6. Speed profiles for scenario 1.

Fig. 7. Speed profiles for scenario 2.

decreased to 30 mph before the DDI crossover and further reduced to 24 mph and 27 mph at the two traffic signals. The speed dispersion was considered high, indicating non-compliance and potential safety concerns. In Phase 2, similar non-compliance was observed, with speeds of 40 mph and 27 mph at the school signs and reduced speeds of 30 mph, 23 mph, and 28 mph at the subsequent traffic signals and guide signs. Phase 3 showed drivers exhibited higher speeds of 39 mph and 31 mph at the school signs, with reductions to 37 mph and 34 mph at the traffic signals and guide signs before the DDI, and finally increasing to 35 mph at the end of the scenario. This consistent pattern of high initial speeds with gradual reductions approaching the DDI and moderate compliance with the school zone speed limit underscores the variability in driver behavior and the potential risks posed by this speed variability.

4.3. Survey responses

The experiment recruited 96 participants, who were divided across the three simulation phases. Phase 1 involved 48 participants, while Phases 2 and 3 each included 24 participants. After each simulation run, subjects were given a questionnaire to obtain feedback on the

effectiveness of the TCDs in assisting in the guidance and navigation complexity to complete a maneuver at the DDI. This section provides relevant findings from the responses obtained from each phase.

Phase 1 focused on identifying areas for improvement in the DDI based on the experience of the subjects with the signage and pavement markings. The maneuvers within the DDI that more subjects reported having issues with the signage were scenarios involving the NB Through (27 %), SB Left turn (27 %), U-turn movements (21 %), and SB Through (15 %). The same four scenarios had the highest number of subjects reporting that the pavement markings presented challenges. More subjects reported difficulty with the pavement markings for the SB Through (27 %), SB Left turn (27 %), NB Through (15 %), and U-turn movements (12 %).

The questionnaire was also used to assess the subjects' overall driving experience at the DDI in terms of their perception of safety at the intersection, the quantity of road signs and pavement markings, and their comprehension of the DDI operations. The primary outcomes from Phase I included:

1. Safety concerns: 46 % of participants reported feeling insecure while navigating the intersection. Since the simulation was conducted under free-flow conditions, participants expressed concerns about how other drivers would behave in the intersection.
2. Quantity of road signs: 49 % of participants agreed there was a lack of road signs. The most frequent issue reported was the absence of signs to guide them through the DDI and aid them in making lane selection decisions based on their travel destination.
3. Quantity of pavement markings: 25 % of subjects felt there were too many markings at the DDI.
4. Desired design modifications: 48 % of subjects suggested modifications to the DDI, such as reducing the quantity of markings and installing guide signs and warning signs to alert drivers about the new intersection, and a regulatory sign for the U-turn movement on the on– and off-freeway ramps.

These responses were analyzed in conjunction with the comments on possible modifications obtained from the subjects and the performance exhibited during the simulation to examine the TCDs in those scenarios. Informed by user feedback, Phase 2 focused on evaluating the impact on driver behavior and perception from implementing the modifications in the DDI to address those concerns from Phase 1. The DDI design was reviewed by PRHTA, reducing pavement markings in the crossover area, incorporating overhead guide signs, and modifying the curb radii on the traffic islands to assist the off-ramp turning maneuvers. Some of the modifications in TCDs incorporated into the DDI can be observed in Fig. 8. Fig. 8a shows a driver's view of the SB crossover section of the DDI created with the initial design drawings as shown to subjects in Phase 1. Fig. 8b shows the same SB crossover section scenario with the modifications shown in Phase 3. The results in driver behavior and their perceptions were demonstrably positive, decreasing the concerns from participants regarding signage clarity for both the SB Through and NB Through maneuvers. Similarly, pavement markings were perceived as significantly more evident in these scenarios.

Building on these positive results, Phase 3 evaluated the effectiveness of Phase 2 modifications while incorporating community outreach activities. Participants were asked to evaluate whether the signs and pavement markings were helpful for driving correctly through the intersection. While the magnitude of the observed improvements was moderate compared to those in Phase 2, the user perception of the DDI continued to improve. Table 2 presents the results of the questions related to the perceived effectiveness of the TCDs and the improvement in perception throughout the study for the NB and SB Through

movements. These maneuvers required participants to traverse both crossovers at the DDI.

The study culminated in positive overall outcomes, as evidenced by a notable increase in driver confidence within the DDI and a corresponding reduction in the perception of insecurity. Furthermore, the number of subjects that perceived that the amount of pavement markings in the DDI was excessive decreased. The study observed a progressive increase in driver satisfaction with the quantity of TCDs across the phases. In Phase 1, only 67 % of the participants perceived the quantity of signs was adequate. This perception improved to 71 % in Phase 2 and reached 88 % in Phase 3. Similarly, driver perception of the pavement markings improved throughout the study. In Phase 1, 25 % of participants found the amount of pavement markings used was excessive. This perception decreased to 12 % in Phase 2 and further dropped to 4 % in Phase 3.

In Phase 1, 12 % of participants indicated a lack of understanding regarding the driver operations at the DDI. This concern was effectively addressed through design modifications and educational materials, as evidenced by all subjects agreeing to understand the operations at the DDI in Phases 2 and 3. The implemented changes significantly improved user perception. After reviewing the final Phase 3 results with stakeholders, further recommendations were made to enhance DDI operations, such as upgrading warning signs to regulatory speed reduction signs and adding pavement markings for speed reduction prior to the sharp horizontal curve on the north approach leg, installing a raised crosswalk to improve safety near the school zone prior to the DDI, and relocating some signage based on driver feedback to improve decisionmaking distance, and incorporating U-turn regulatory signage for the PR-30 off-ramps.

4.4. Outreach activities

Since this was the first DDI implemented in Puerto Rico, the project prioritized educating drivers about the new intersection design. A comprehensive educational program was developed, including a brochure, video, and practice sessions with the driving simulator. The research team and project sponsors organized outreach activities throughout the three phases to raise awareness. During Phase 1, three outreach events were conducted, allowing 200 participants to experience the DDI through the simulator. Due to COVID-19 health restrictions at that time, the number of participants was limited compared to subsequent phases. The educational activities were held at the university next to the DDI site, a major shopping mall in San Juan, and a convention center. During Phase 2, the driving simulator was located at a shopping mall near the DDI construction site for a month, attracting over 1,500 people. Collaborations with the Puerto Rico Police, the Puerto Rico Traffic Safety Commission, the FHWA Local Division Office, first responders of the nearby municipalities, and PRHTA personnel further amplified the impact of the outreach efforts. Three additional public outreach activities were conducted in Phase 3 to raise awareness about the DDI. The first event occurred at a traffic management center on a nearby freeway corridor, while the second event was at a popular TV show called "Pegate al Medio Día" that impacted a national audience. This TV segment featured a well-known presenter from one of the municipalities near the DDI site. The third outreach event was held during a 3-day exposition with the driving simulator at Plaza Las Am´ericas Shopping Center, reaching over 750 people.

5. Discussion

The implementation of the DDI in Puerto Rico had several challenges that needed to be addressed. One of the main concerns was the uncustomary left-side driving maneuvers required by the DDI design. To ensure the safety of road users, it was necessary to develop effective strategies to educate drivers about the new design and to evaluate the Fig. 8. Driver's view of the crossover section of the simulated DDI. effectiveness of the traffic control devices to be installed at the DDI. The

Table 2 Perceived effectiveness of the traffic control devices at the DDI per Phase.

driving simulation experiment provided valuable information that allowed the DDI designers to identify potential problems and make necessary changes to the design and the traffic control devices.

The simulation experiment allowed designers to understand the behavior and performance of drivers in different scenarios, which helped them identify potential problems with the DDI design. For example, the simulation revealed that some first-time drivers may experience confusion when making the unconventional crossover maneuvers at the DDI. The speed profiles and vehicle trajectories obtained from the simulation were also used to analyze the decisions made by drivers. Some of the inappropriate driving decisions and errors observed included speeding, selection of wrong routes, and erratic or incorrect maneuvers. These findings were used to improve the design and traffic control devices to enhance the level of information provided to drivers, thus improving their navigation through the DDI.

The simulation experiment also provided a platform for conducting training sessions that allowed local drivers to experience the new operating conditions of the DDI in a controlled and secure environment. Over 2,600 people participated in these training sessions, which helped in reducing potential confusion for first-time drivers. The training sessions also provided a platform for the DDI designers to acquire direct feedback from users, which was used to verify the impact of the modifications added to the DDI design in the subsequent phases. The changes made to the traffic control devices proved to be effective in improving the navigation of drivers through the DDI and their perception of driving complexity and safety. The use of a driving simulator as part of a safety review during the design and construction process for innovative facilities, such as the DDI, became a valuable stage that assisted designers in assessing the effectiveness of proposed design features and the identification of improvements to the proposed traffic control devices.

6. Conclusions and recommendations

This research used a driving simulator to assess and refine the design of a novel intersection in Puerto Rico. By incorporating user feedback from the three phases of the study, the project ensured a safer and more user-friendly design for drivers. This iterative approach enhanced safety and provided cost-effective solutions for the PRHTA before operations at the new intersection began.

The analysis of the driver speeds throughout the three study phases revealed consistent speed selection behavior within the DDI's crossover area, ranging from 21 to 30 mph. This aligns with the average speeds observed for DDIs in the U.S. as reported in FHWA (2014). The modifications in signage and pavement markings at the DDI did not appear to significantly impact driver speeds. This outcome is consistent with the primary purpose of the modifications to improve driver navigation within the DDI, and not necessarily to control speeds.

The driving simulation revealed instances of driver behavior requiring improvement, including speeding, selecting incorrect routes, and performing erratic maneuvers that could pose safety hazards. These errors were categorized as either erratic (non-critical) or critical based on their severity. The results revealed that the modifications made in the DDI stimulated a 26 % reduction in the total number of errors across all phases and a 47 % reduction in the number of critical errors, suggesting that improvements to the DDI design were effective.

Throughout all three phases of the study, a significant portion of participants (75 %, 79 %, and 92 % in Phases 1, 2, and 3, respectively) reported no prior experience driving on a DDI. This lack of experience with this innovative intersection design reinforces the critical need for extensive and ongoing public education and training efforts. These efforts should not be limited to the pre-opening phase but should be sustained even after the DDI is open to traffic to ensure continued awareness, especially among unfamiliar drivers.

The study observed a significant improvement in driver perception of safety while navigating the DDI. From Phase 1 to Phase 3, the percentage of participants reporting "unsafe" perceptions decreased from 46 % to 19 %. This reduction coincides with the implementation of improved signage and a reduction in lane extension pavement markings at the crossover (the original markings are shown in Fig. 8a), suggesting that these design modifications effectively addressed user concerns and enhanced the overall feeling of safety within the DDI.

During the training and outreach activities, the participants were provided with a brochure, a video, and practice sessions of the DDI with the driving simulation. The public learned about the DDI's advantages, including better safety and reduced traffic congestion, and shared their opinions on the new intersection design. The research team found that the training program was effective in reaching its educational goals and that the participants would continue to spread the knowledge they gained to their friends and family.

The study found that driver satisfaction with the quantity of TCDs increased significantly throughout the phases. Specifically, there was a notable improvement in the perception of the quantity of signs and pavement markings, with fewer drivers finding them excessive. Furthermore, design changes and educational activities successfully addressed concerns regarding understanding DDI operations.

The UPRM driving simulator has demonstrated to be a valuable asset for training drivers on novel intersections and other innovative road designs. It provides a safe and controlled environment where users can experience realistic simulations of new geometric layouts and operational features. Additionally, the simulator offers a unique platform for public engagement, allowing researchers to gather valuable user feedback and address concerns regarding innovative designs like the DDI before implementation.

Based on the study findings, several directions for future research and development can be pursued. Monitoring and evaluating the DDI's performance in real-world traffic conditions would be beneficial in gaining further insights into its long-term sustainability and effectiveness. Additionally, exploring the integration of emerging technologies, such as intelligent transportation systems and vehicle-to-infrastructure communication, could boost the efficiency and safety of DDIs. Furthermore, continuing community outreach and education efforts are crucial to ensure sustained public support and engagement.

CRediT authorship contribution statement

Didier M. Valdes: ´ Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – review & editing. Alberto Figueroa-Medina: Conceptualization, Formal analysis, Methodology, Supervision, Visualization, Writing – review & editing. Carol A. Perello: ´ Data curation, Formal analysis, Investigation, Methodology, Writing – original draft.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: The authors report that the Puerto Rico Highways and Transportation Authority and the Safety Research Using Simulation UTC provided financial support.

Acknowledgments

This material is based upon work supported by the Puerto Rico Highways and Transportation Authority (PRHTA) and the Safety Research Using Simulation (SAFER-SIM) University Transportation Center and partially funded by the U.S. Department of Transportation (US-DOT) Office of the Assistant Secretary for Research and Technology under Grant No. 69A3551747131. Any opinions, findings, or conclusions expressed in this paper are those of the authors and do not necessarily reflect the views of the USDOT or the PRHTA.

The authors extend their gratitude to the following entities and persons: the UPRM SAFER-SIM team, the Puerto Rico Department of Transportation and Public Works, Constructora Santiago Corp., including Jos´e Torres and Jos´e Chinea, the FHWA Local Division Office, including Andr´es Alvarez, ´ Juan C. Rivera, and Grace Díaz, the UPRM Civil Infrastructure Research Center, the Puerto Rico LTAP Center, and the volunteers that participated in the simulation experiments and training activities carried out in this project.

Data availability

Data will be made available on request.

References

Claros, B., Edara, P., Sun, C., 2017. Safety effect of diverging diamond interchanges on adjacent roadway facilities. Transp. Res. Rec. 2618 (1), 78–90. https://doi.org/ 10.3141/2618-08.
Claros, B., Edara, P., Sun, C., Brown, H., 2015. Safety evaluation of diverging diamond interchanges in Missouri. Transp. Res. Rec. 2486 (1), 1–10. https://doi.org/10.3141/ 2486-01.
Pham, B., 1998. Quality evaluation of educational multimedia systems. Australasian Journal of Educational Technology 14 (2). https://doi.org/10.14742/ajet.1904.
Valdes, D., Colucci, B., Figueroa, A., Rivera Ortiz, J., Bido Lorenzo, L., Rojas, M., Colon, E.. August 3). Framework for considering a diverging diamond interchange to improve operations and safety. https://laccei.org/LACCEI2019-MontegoBay/meta/F P417.html.
FHWA, 2014. Diverging Diamond Interchange Informational Guide. Report FHWA-SA-14-067. United States Department of Transportation. Federal Highway Administration. https://rosap.ntl.bts.gov/view/dot/29474.
FHWA, 2012. Manual on Uniform Traffic Control Devices for Streets and Highways. 2009 Edition with Revisions 1 and 2. United States Department of Transportation. Federal Highway Administration. https://mutcd.fhwa.dot.gov/pdfs/2009/ mutcd2009edition.pdf.
FHWA, 2007. Drivers' Evaluation of the Diverging Diamond Interchange. Report FHWA-HRT-07-048. United States Department of Transportation. Federal Highway Administration. Turner-Fairbank Highway Research Center. https://www.fhwa.dot. gov/publications/research/safety/07048/.
Fisher, D.L., Rizzo, M., Caird, J., Lee, J.D. (Eds.), 2011. Handbook of Driving Simulation for Engineering, Medicine, and Psychology. 1st Edition. CRC Press. Taylor & Francis Group. doi: 10.1201/b10836.
Khalil, M., Mansour, M., Wilhite, D., 2010. Evaluation of cognitive loads imposed by traditional paper-based and innovative computer-based instructional strategies. Veterinary Medical Education 37, 353–357. https://doi.org/10.3138/jvme.37.4.353. PMID: 21135402.
Khattak, A., Ahmad, N., Wali, B., Dumbaugh, E., 2021. A taxonomy of driving errors and violations: evidence from the naturalistic driving study. Accid. Anal. Prev. 151, 105873. https://doi.org/10.1016/j.aap.2020.105873.
MODOT, 2010. Missouri's Experience with a Diverging Diamond Interchange Lessons Learned. Missouri Department of Transportation. https://rosap.ntl.bts.gov/view/ dot/18089/dot_18089_DS1.pdf.
NASEM, 2021. Diverging Diamond Interchange Informational Guide, Second Edition. The National Academic Press. National Academies of Sciences, Engineering, and Medicine. doi: 10.17226/26027.

valdesdrivingsimulatoranalysisfirst2025.md 48 KB 文件歷史 原始文件