Stephen Osafo-Gyamfi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Stephen Osafo-Gyamfi has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Amirarsalan Mehrara Molan
Univ. of Mississippi
6 shared publications - Ali Hajbabaie
North Carolina State University
4 shared publications - Gaurav Aryal
Univ. of Mississippi
3 shared publications Hayden Edwards
North Carolina State Univ
3 shared publications- Hayden Edwards
Department of Civil, Construction, and Environmental Engineering, North Carolina State University, Raleigh, NC
3 shared publications William Rasdorf
North Carolina State Univ
3 shared publications
Latest Publications
- 1 November 2026
- Published by American Society of Civil Engineers (ASCE) in Journal of Transportation Engineering, Part A: Systems
- Vol. 152 (11)
Travel Time and Cost-Benefit Evaluation of Alternative Intersections with Three-Critical-Phase Traffic Signals in North Carolina
This paper explores the potential benefits of three-phase intersections as a compromise between two-phase and conventional four-phase designs, particularly in terms of improving travel time efficiency and safety. While two-phase intersections offer advantages in reducing travel time and enhancing safety under near-capacity traffic conditions, they are often unpopular due to unfamiliarity and their impact on accessibility, especially when there is high redirected traffic demand from the minor road. Three-phase intersections may address these limitations by combining some of the benefits of two-phase intersections, while they only remove one of the signal phases. This study evaluates six case study sites in North Carolina using traffic microsimulation to assess the performance of nine proposed three-phase alternative intersection designs, including two versions of partial median U-turn (MUT), partial displaced left turn (DLT), reverse reduced crossing U-turn intersection (reverse RCUT), a partial DLT/MUT combination, thru-cut, seven-phase, and redirect left and through (redirect L&T) designs. A benefit-to-cost (B/C) analysis was conducted to quantify the costs and travel time savings associated with each design. Results indicate that three-phase designs generally offer noteworthy improvements in travel times savings over conventional four-phase designs, especially in locations with specific traffic patterns and demand levels. Three-phase designs could also result in lower travel times than intersection designs with two-phase signals at case study sites with significantly high traffic demands. The findings emphasize the importance of tailoring intersection design to site-specific conditions and suggest that three-phase intersections may serve as a cost-effective solution for enhancing operational performance from the cost-benefit analysis. The findings of this study are intended to assist state DOTs in selecting and designing optimal intersection configurations.
- 1 November 2026
- Published by American Society of Civil Engineers (ASCE) in Journal of Transportation Engineering, Part A: Systems
- Vol. 152 (11)
Travel Time and Cost-Benefit Evaluation of Alternative Intersections with Three-Critical-Phase Traffic Signals in North Carolina
This paper explores the potential benefits of three-phase intersections as a compromise between two-phase and conventional four-phase designs, particularly in terms of improving travel time efficiency and safety. While two-phase intersections offer advantages in reducing travel time and enhancing safety under near-capacity traffic conditions, they are often unpopular due to unfamiliarity and their impact on accessibility, especially when there is high redirected traffic demand from the minor road. Three-phase intersections may address these limitations by combining some of the benefits of two-phase intersections, while they only remove one of the signal phases. This study evaluates six case study sites in North Carolina using traffic microsimulation to assess the performance of nine proposed three-phase alternative intersection designs, including two versions of partial median U-turn (MUT), partial displaced left turn (DLT), reverse reduced crossing U-turn intersection (reverse RCUT), a partial DLT/MUT combination, thru-cut, seven-phase, and redirect left and through (redirect L&T) designs. A benefit-to-cost (B/C) analysis was conducted to quantify the costs and travel time savings associated with each design. Results indicate that three-phase designs generally offer noteworthy improvements in travel times savings over conventional four-phase designs, especially in locations with specific traffic patterns and demand levels. Three-phase designs could also result in lower travel times than intersection designs with two-phase signals at case study sites with significantly high traffic demands. The findings emphasize the importance of tailoring intersection design to site-specific conditions and suggest that three-phase intersections may serve as a cost-effective solution for enhancing operational performance from the cost-benefit analysis. The findings of this study are intended to assist state DOTs in selecting and designing optimal intersection configurations.
- 13 July 2026
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
Evaluating Pedestrian Safety and Service at Alternative Intersections with Three-Critical-Phase and Two-Critical-Phase Traffic Signals
Intersections constitute an important part of transportation infrastructure, with conventional intersection designs being the most dominant. However, these conventional designs might face operational and safety issues, emphasizing the need for innovative and newer intersection designs to address these challenges. A key concern is the inadequate pedestrian accommodation in conventional intersections, primarily as a result of limited attention to pedestrian needs in older designs. On the other hand, alternative intersection designs may have the potential to improve pedestrian performance by reducing vehicle–pedestrian conflicts, shortening signal cycle lengths, and possibly decreasing pedestrian crossing distances. This paper focuses on the pedestrian safety and service of 10 types of alternative intersections with three-critical-phase traffic signals and two types with two-critical-phase traffic signals in a comparison with a signalized conventional intersection design (with four signal phases). The evaluation was conducted using a wide range of hypothetical scenarios in simulation modeling, the 20-design flag method from the National Cooperative Highway Research Program (NCHRP) Research Report 948, and the conflict point analysis to assess pedestrian performance. The findings suggest that three-phase intersection designs such as the partial median U-turn designs may offer better pedestrian operational and safety performance compared with conventional four-phase designs.
- 13 July 2026
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
Evaluating Pedestrian Safety and Service at Alternative Intersections with Three-Critical-Phase and Two-Critical-Phase Traffic Signals
Intersections constitute an important part of transportation infrastructure, with conventional intersection designs being the most dominant. However, these conventional designs might face operational and safety issues, emphasizing the need for innovative and newer intersection designs to address these challenges. A key concern is the inadequate pedestrian accommodation in conventional intersections, primarily as a result of limited attention to pedestrian needs in older designs. On the other hand, alternative intersection designs may have the potential to improve pedestrian performance by reducing vehicle–pedestrian conflicts, shortening signal cycle lengths, and possibly decreasing pedestrian crossing distances. This paper focuses on the pedestrian safety and service of 10 types of alternative intersections with three-critical-phase traffic signals and two types with two-critical-phase traffic signals in a comparison with a signalized conventional intersection design (with four signal phases). The evaluation was conducted using a wide range of hypothetical scenarios in simulation modeling, the 20-design flag method from the National Cooperative Highway Research Program (NCHRP) Research Report 948, and the conflict point analysis to assess pedestrian performance. The findings suggest that three-phase intersection designs such as the partial median U-turn designs may offer better pedestrian operational and safety performance compared with conventional four-phase designs.
- 7 April 2026
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2680 (12) , 401-418
Safety Effectiveness of Unsignalized Restricted Crossing U-Turn Intersections in Mississippi using an Empirical Bayes Before–After Study
At-grade intersections on rural high-speed highways can be hazardous, particularly where lower-speed vehicles from minor roads must cross high-speed traffic on major roads. To enhance safety at these junctions, an alternative design known as restricted crossing U-turn (RCUT) has gained considerable interest in recent years. RCUTs can help reduce both the number and severity of crashes by rerouting minor road traffic to downstream U-turns, thereby minimizing crossing conflict points between high-speed and low-speed vehicles. This research aimed to estimate the safety performance following the implementation of eight unsignalized RCUT intersections on rural high-speed highways in Mississippi. The study considered two different definitions of the intersection influence area and employed two distinct methodologies: naïve average annual number of crashes analysis, and a rigorous Empirical Bayes (EB) method. To facilitate the crash modification factor (CMF) development, new local safety performance functions (SPFs) were developed, and the results were compared with those derived from the HSM (Highway Safety Manual) SPFs and state-specific SPFs. Results showed significant crash reductions following RCUT implementation: a 63% reduction in total crashes and a 79% reduction in fatal and injury crashes. Annual angle crashes dropped by over 96%, and left-turn crashes by 40%, although sideswipe and rear-end crashes increased slightly as a result of redirected traffic. Despite methodological differences, all analyses confirmed RCUTs’ effectiveness in improving safety, making them a promising solution for high-speed rural intersections in Mississippi.
- 7 April 2026
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2680 (12) , 401-418
Safety Effectiveness of Unsignalized Restricted Crossing U-Turn Intersections in Mississippi using an Empirical Bayes Before–After Study
At-grade intersections on rural high-speed highways can be hazardous, particularly where lower-speed vehicles from minor roads must cross high-speed traffic on major roads. To enhance safety at these junctions, an alternative design known as restricted crossing U-turn (RCUT) has gained considerable interest in recent years. RCUTs can help reduce both the number and severity of crashes by rerouting minor road traffic to downstream U-turns, thereby minimizing crossing conflict points between high-speed and low-speed vehicles. This research aimed to estimate the safety performance following the implementation of eight unsignalized RCUT intersections on rural high-speed highways in Mississippi. The study considered two different definitions of the intersection influence area and employed two distinct methodologies: naïve average annual number of crashes analysis, and a rigorous Empirical Bayes (EB) method. To facilitate the crash modification factor (CMF) development, new local safety performance functions (SPFs) were developed, and the results were compared with those derived from the HSM (Highway Safety Manual) SPFs and state-specific SPFs. Results showed significant crash reductions following RCUT implementation: a 63% reduction in total crashes and a 79% reduction in fatal and injury crashes. Annual angle crashes dropped by over 96%, and left-turn crashes by 40%, although sideswipe and rear-end crashes increased slightly as a result of redirected traffic. Despite methodological differences, all analyses confirmed RCUTs’ effectiveness in improving safety, making them a promising solution for high-speed rural intersections in Mississippi.
- 13 December 2025
- Published by SAGE Publications in Public Works Management & Policy
- Vol. 31 (2) , 140-171
Public Acceptance Scoring System for Alternative Intersections
Intersections are key components of road infrastructure, with conventional designs being the most common. However, these designs might face capacity and safety issues, prompting the need for alternatives. Alternative intersections, such as reduced conflict intersections (RCIs) have shown improvements in travel time and safety for both vehicles and pedestrians, but public acceptance remains a challenge. This paper introduces the Public Acceptance Scoring System (PASS), a tool designed to evaluate the public acceptance of alternative intersection designs. PASS prioritizes variables such as driver confusion, business impacts, and pedestrian discomfort. Based on a review of existing studies and input from experts at the North Carolina Department of Transportation (NCDOT), 16 key variables were identified. The results show that newer designs, like the seven-phase and partial median U-Turn (MUT) with a three-phase traffic signal, are likely to receive higher public acceptance, while partial Continuous Flow Intersections (CFI) may struggle due to traffic redirection and business access limitations.
- 13 December 2025
- Published by SAGE Publications in Public Works Management & Policy
- Vol. 31 (2) , 140-171
Public Acceptance Scoring System for Alternative Intersections
Intersections are key components of road infrastructure, with conventional designs being the most common. However, these designs might face capacity and safety issues, prompting the need for alternatives. Alternative intersections, such as reduced conflict intersections (RCIs) have shown improvements in travel time and safety for both vehicles and pedestrians, but public acceptance remains a challenge. This paper introduces the Public Acceptance Scoring System (PASS), a tool designed to evaluate the public acceptance of alternative intersection designs. PASS prioritizes variables such as driver confusion, business impacts, and pedestrian discomfort. Based on a review of existing studies and input from experts at the North Carolina Department of Transportation (NCDOT), 16 key variables were identified. The results show that newer designs, like the seven-phase and partial median U-Turn (MUT) with a three-phase traffic signal, are likely to receive higher public acceptance, while partial Continuous Flow Intersections (CFI) may struggle due to traffic redirection and business access limitations.
- 30 July 2025
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2679 (10) , 61-78
Performance Effects of the New Displaced Left Turn and Median U-Turn Combination Design
During the last two decades, transportation professionals have turned to alternative intersection designs with two critical signal phases, such as median U-turn (MUT) and reduced conflict intersections (RCIs), to improve safety and mobility. Intersections with two critical signal phases often have substantial safety and performance benefits. Nonetheless, two-critical-phase designs may lack public acceptance and may have noticeable impacts at some intersection sites owing to the substantial changes required to convert a four-phase traffic signal to a two-phase traffic signal. Conversely, three-critical-phase designs might mitigate some of these issues because only one of the signal phases will be removed (compared with a four-phase signal), but they may still present some of the benefits of two-phase designs. This study utilized microsimulation to analyze the performance effects of a new three-phase design, the partial displaced left turn and median U-turn combination (DLT/MUT combination), a partial DLT, and an existing (conventional) intersection at two intersection sites in North Carolina. The results showed that the partial DLT/MUT combination can outperform the conventional intersections in high-traffic volume scenarios. On average, this new design reduced vehicle travel time by 63% and 21% in comparison to the conventional intersection in case study sites #1 and #2, respectively. It also performed similarly with the partial DLT in operations, but it should be safer and more friendly to people walking and biking, and more cost-effective.
- 30 July 2025
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2679 (10) , 61-78
Performance Effects of the New Displaced Left Turn and Median U-Turn Combination Design
During the last two decades, transportation professionals have turned to alternative intersection designs with two critical signal phases, such as median U-turn (MUT) and reduced conflict intersections (RCIs), to improve safety and mobility. Intersections with two critical signal phases often have substantial safety and performance benefits. Nonetheless, two-critical-phase designs may lack public acceptance and may have noticeable impacts at some intersection sites owing to the substantial changes required to convert a four-phase traffic signal to a two-phase traffic signal. Conversely, three-critical-phase designs might mitigate some of these issues because only one of the signal phases will be removed (compared with a four-phase signal), but they may still present some of the benefits of two-phase designs. This study utilized microsimulation to analyze the performance effects of a new three-phase design, the partial displaced left turn and median U-turn combination (DLT/MUT combination), a partial DLT, and an existing (conventional) intersection at two intersection sites in North Carolina. The results showed that the partial DLT/MUT combination can outperform the conventional intersections in high-traffic volume scenarios. On average, this new design reduced vehicle travel time by 63% and 21% in comparison to the conventional intersection in case study sites #1 and #2, respectively. It also performed similarly with the partial DLT in operations, but it should be safer and more friendly to people walking and biking, and more cost-effective.
- 29 January 2025
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2679 (6) , 120-137
Traffic Operations Analysis of Seven New Alternative Intersections with Three-Phase Traffic Signals
This paper presents a traffic operational comparison of seven new alternative signalized intersection designs with three critical phases: 1) offset thru-cut, 2) thru-cut, 3) seven-phase, 4) redirect left and through from one minor road (redirect L&T), 5) reverse reduced conflict intersection (reverse RCI), and 6) and 7) two new versions of partial median U-turn (MUT) intersection. These designs will also be compared with the four-phase conventional intersection and two alternatives with two-phase traffic signals: full MUT and reduced conflict intersection (RCI) designs under various traffic conditions at near-capacity levels. Overall, 1,800 simulation runs were conducted to estimate traffic operations under different traffic conditions with varying traffic volumes, traffic distributions, and turning traffic ratios. Traffic signals were optimized for all the tests, and the optimized signal data was imported into microsimulation to evaluate the performance of the 10 intersections included. According to the results, four of the three-phase intersection designs (offset thru-cut, thru-cut, MUT #1, and reverse RCI) could result in significantly shorter travel times than the conventional intersection. The traffic operations of these four intersection designs were found to be similar to the two-phase designs and even better than the full MUT in scenarios with high turning traffic ratios. The other three intersection designs (seven-phase, redirect L&T, and MUT #2) also showed potential as possible substitutes for existing four-phase intersections under conditions with higher turning traffic demands.
- 29 January 2025
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2679 (6) , 120-137
Traffic Operations Analysis of Seven New Alternative Intersections with Three-Phase Traffic Signals
This paper presents a traffic operational comparison of seven new alternative signalized intersection designs with three critical phases: 1) offset thru-cut, 2) thru-cut, 3) seven-phase, 4) redirect left and through from one minor road (redirect L&T), 5) reverse reduced conflict intersection (reverse RCI), and 6) and 7) two new versions of partial median U-turn (MUT) intersection. These designs will also be compared with the four-phase conventional intersection and two alternatives with two-phase traffic signals: full MUT and reduced conflict intersection (RCI) designs under various traffic conditions at near-capacity levels. Overall, 1,800 simulation runs were conducted to estimate traffic operations under different traffic conditions with varying traffic volumes, traffic distributions, and turning traffic ratios. Traffic signals were optimized for all the tests, and the optimized signal data was imported into microsimulation to evaluate the performance of the 10 intersections included. According to the results, four of the three-phase intersection designs (offset thru-cut, thru-cut, MUT #1, and reverse RCI) could result in significantly shorter travel times than the conventional intersection. The traffic operations of these four intersection designs were found to be similar to the two-phase designs and even better than the full MUT in scenarios with high turning traffic ratios. The other three intersection designs (seven-phase, redirect L&T, and MUT #2) also showed potential as possible substitutes for existing four-phase intersections under conditions with higher turning traffic demands.
- 20 May 2024
- Published by Seismological Society of America (SSA) in Seismological Research Letters
- Vol. 95 (5) , 2886-2893
Predicting Peak Ground Acceleration of Strong-Motion Earthquakes Using Variable Snapshots of P-Wave Data with Long Short-Term Memory Neural Network
Conventional earthquake early warning systems (EEWS) rely on mathematical functions that utilize P-wave parameters extracted over a 3 s window to estimate peak ground acceleration (PGA). Advancements in the capabilities of deep neural networks to approximate universal functions, coupled with the availability of strong seismic event data, offer an unprecedented opportunity to evaluate the relationship between variable snapshots of P-wave data and the PGA of strong-motion earthquakes. This convergence of technology and data opens new avenues for research into utilizing smaller snapshots of P-wave in EEWS. Our study centers on the utilization of a long short-term memory (LSTM) neural network to model long dependencies within the P-wave of 1839 earthquakes recorded by the Kiyonshin Network (K-NET) for the prediction of PGA of S waves. Our methodology involves experiments that ultimately evaluate the network’s performance on 4, 3, and 2 s of P-wave snapshots. Our findings indicate that there is sufficient information in 2 s of temporal accelerometer readings after the onset of P waves to predict PGA accurately with an LSTM network.
- 20 May 2024
- Published by Seismological Society of America (SSA) in Seismological Research Letters
- Vol. 95 (5) , 2886-2893
Predicting Peak Ground Acceleration of Strong-Motion Earthquakes Using Variable Snapshots of P-Wave Data with Long Short-Term Memory Neural Network
Conventional earthquake early warning systems (EEWS) rely on mathematical functions that utilize P-wave parameters extracted over a 3 s window to estimate peak ground acceleration (PGA). Advancements in the capabilities of deep neural networks to approximate universal functions, coupled with the availability of strong seismic event data, offer an unprecedented opportunity to evaluate the relationship between variable snapshots of P-wave data and the PGA of strong-motion earthquakes. This convergence of technology and data opens new avenues for research into utilizing smaller snapshots of P-wave in EEWS. Our study centers on the utilization of a long short-term memory (LSTM) neural network to model long dependencies within the P-wave of 1839 earthquakes recorded by the Kiyonshin Network (K-NET) for the prediction of PGA of S waves. Our methodology involves experiments that ultimately evaluate the network’s performance on 4, 3, and 2 s of P-wave snapshots. Our findings indicate that there is sufficient information in 2 s of temporal accelerometer readings after the onset of P waves to predict PGA accurately with an LSTM network.
- Open Access
- 29 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 60628-60640
DivNEDS: Diverse Naturalistic Edge Driving Scene Dataset for Autonomous Vehicle Scene Understanding
The safe implementation and adoption of Autonomous Vehicle (AV) vision models on public roads requires not only an understanding of the natural environment comprising pedestrians and other vehicles but also the ability to reason about edge situations such as unpredictable maneuvers by other drivers, impending accidents, erratic movement of pedestrians, cyclists, and motorcyclists, animal crossings, and cyclists using hand signals. Despite advances in complex tasks such as object tracking, human behavior modeling, activity recognition, and trajectory planning, the fundamental challenge of interpretable scene understanding, especially in out-of-distribution environments, remains evident. This is highlighted by the 84% of AV disengagements attributed to scene understanding errors in real-world AV tests. To address this limitation, we introduce the Diverse Naturalistic Edge Driving Scene Dataset (DivNEDS), a novel dataset comprising 11,084 edge scenes and 203,000 descriptive captions sourced from 12 distinct locations worldwide, captured under varying weather conditions and at different times of the day. Our approach includes a novel embedded hierarchical dense captioning strategy aimed at enabling few-shot learning and mitigating overfitting by excluding irrelevant scene elements. Additionally, we propose a Generative Region-to-Text Transformer, with a baseline embedded hierarchical dense captioning performance of 60.3mAP, a new benchmark for AV scene understanding models trained on dense captioned data sets. This work represents a significant step toward improving AVs’ ability to comprehend diverse, real-world edge and complex driving scenarios, thereby enhancing their safety and adaptability in dynamic environments. The dataset and instructions are available at https://github.com/johnowusuduah/DivNEDS .
- 29 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 60628-60640
Open AccessDivNEDS: Diverse Naturalistic Edge Driving Scene Dataset for Autonomous Vehicle Scene Understanding
The safe implementation and adoption of Autonomous Vehicle (AV) vision models on public roads requires not only an understanding of the natural environment comprising pedestrians and other vehicles but also the ability to reason about edge situations such as unpredictable maneuvers by other drivers, impending accidents, erratic movement of pedestrians, cyclists, and motorcyclists, animal crossings, and cyclists using hand signals. Despite advances in complex tasks such as object tracking, human behavior modeling, activity recognition, and trajectory planning, the fundamental challenge of interpretable scene understanding, especially in out-of-distribution environments, remains evident. This is highlighted by the 84% of AV disengagements attributed to scene understanding errors in real-world AV tests. To address this limitation, we introduce the Diverse Naturalistic Edge Driving Scene Dataset (DivNEDS), a novel dataset comprising 11,084 edge scenes and 203,000 descriptive captions sourced from 12 distinct locations worldwide, captured under varying weather conditions and at different times of the day. Our approach includes a novel embedded hierarchical dense captioning strategy aimed at enabling few-shot learning and mitigating overfitting by excluding irrelevant scene elements. Additionally, we propose a Generative Region-to-Text Transformer, with a baseline embedded hierarchical dense captioning performance of 60.3mAP, a new benchmark for AV scene understanding models trained on dense captioned data sets. This work represents a significant step toward improving AVs’ ability to comprehend diverse, real-world edge and complex driving scenarios, thereby enhancing their safety and adaptability in dynamic environments. The dataset and instructions are available at https://github.com/johnowusuduah/DivNEDS .