Traffic Safety Guide

TRAFFIC SAFETY GUIDE

Despite much advancement in traffic safety, motor vehicle accidents still account for over 36,000 deaths in the United States and have increased from 1999 to 2015 (National Vital Statistics Reports, 2017). As such, there is a need to improve traffic safety, and reduce deaths due to motor vehicle collisions. 

Road design is often informed by traffic microsimulations. These simulations are usually based upon driver models and make assumptions regarding the drivers. Using these models, it is difficult to predict accidents because, on the timescale of the microsimulations, they are relatively rare (Ciro and Maurizio, 2012). Instead, accidents are prevented using standard design practices (Erskine, 2006), but these practices may not be suitable or scalable to certain regions, road conditions, or intersection geometry. Designers need a tool that can accurately and quickly analyze a road design and identify key areas of high risk for accidents.

According to a study by World Health Organization, approximately 1.25 million die each year in traffic accidents (Who, 2015). There has been previous research investigating the causes of road crashes, including factors associated with driver behavior and road design features (Thomas, 2013; Wedajo, 2017; Mohamedshah, 1994; Çepni, 2017; Lam, 1995). William Haddon developed a matrix that identifies the risk factors before the crash, during the crash and after the crash, in relation to the person, vehicle and environment (Haddon, 1968). Some accidents were found to occur due to road designs; however, police officials considered these accidents due to drivers’ fatal mistakes (Obaidat, 2012). Hence, further analysis of crash data in relation to road designs can provide a better understanding of the higher crash rates in certain locations and allow the officials to have a clearer picture of the factors involved in accidents.

Despite much advancement in traffic safety, motor vehicle accidents still account for over 36,000 deaths in the United States and have increased from 1999 to 2015 (National Vital Statistics Reports, 2017). As such, there is a need to improve traffic safety, and reduce deaths due to motor vehicle collisions. 

Road design is often informed by traffic microsimulations. These simulations are usually based upon driver models and make assumptions regarding the drivers. Using these models, it is difficult to predict accidents because, on the timescale of the microsimulations, they are relatively rare (Ciro and Maurizio, 2012). Instead, accidents are prevented using standard design practices (Erskine, 2006), but these practices may not be suitable or scalable to certain regions, road conditions, or intersection geometry. Designers need a tool that can accurately and quickly analyze a road design and identify key areas of high risk for accidents.

According to a study by World Health Organization, approximately 1.25 million die each year in traffic accidents (Who, 2015). There has been previous research investigating the causes of road crashes, including factors associated with driver behavior and road design features (Thomas, 2013; Wedajo, 2017; Mohamedshah, 1994; Çepni, 2017; Lam, 1995). William Haddon developed a matrix that identifies the risk factors before the crash, during the crash and after the crash, in relation to the person, vehicle and environment (Haddon, 1968). Some accidents were found to occur due to road designs; however, police officials considered these accidents due to drivers’ fatal mistakes (Obaidat, 2012). Hence, further analysis of crash data in relation to road designs can provide a better understanding of the higher crash rates in certain locations and allow the officials to have a clearer picture of the factors involved in accidents.

Policies on highway geometric design in the US are developed by the American Association of State Highway and Transportation Officials (AASHTO). These policies are based on design guidelines by the state highway and transportation departments and the Federal Highway Administration (FHWA). Guidelines for highway geometric design are presented in A Policy on Geometric Design of Highways and Streets (AASHTO, 2001). Previous studies investigated various factors associated with road design features (e.g., intersections) or roadsize features (e.g., the size and type of advertising signs) (Ziegler, 1951). Roadway curves were found to have five times more accident than straight roadways, where sharper roadway curves have higher accident rates than flatter roadway curves (Timothy et al., 1983; Ziegler, 1951). These findings motivated further simulation studies to explore different approaches to lower the rate of car crashes, including the installation of intelligent road systems (e.g., physical intelligent barriers (Massoudian and Eshghi, 2015). Some of the earlier work also reviewed tunnels’ traffic safety, including the impact of tunnel length, fire incidents, and zone of the tunnel (Bassan, 2016; Ma, 2009; Amundsen, 2001; Cornelia, 2007). Short tunnels had higher crash rates than long tunnels although the chance of crash occurrence is lower in a tunnel than on an open highway section. Amundsen et al. found that the crash rates are higher in the entrance of tunnels and decreased as drivers continued driving inside the tunnel (Amundsen, 2001). However, a study of Chinese freeway tunnels found an increased number of crashes inside the tunnel in contrast to previous studies (Ma, 2009). Potential reasons were thought to be higher driving speeds and insufficient sight distance, which is required for higher design speeds (Ma, 2009).

Policies on highway geometric design in the US are developed by the American Association of State Highway and Transportation Officials (AASHTO). These policies are based on design guidelines by the state highway and transportation departments and the Federal Highway Administration (FHWA). Guidelines for highway geometric design are presented in A Policy on Geometric Design of Highways and Streets (AASHTO, 2001). Previous studies investigated various factors associated with road design features (e.g., intersections) or roadsize features (e.g., the size and type of advertising signs) (Ziegler, 1951). Roadway curves were found to have five times more accident than straight roadways, where sharper roadway curves have higher accident rates than flatter roadway curves (Timothy et al., 1983; Ziegler, 1951). These findings motivated further simulation studies to explore different approaches to lower the rate of car crashes, including the installation of intelligent road systems (e.g., physical intelligent barriers (Massoudian and Eshghi, 2015). Some of the earlier work also reviewed tunnels’ traffic safety, including the impact of tunnel length, fire incidents, and zone of the tunnel (Bassan, 2016; Ma, 2009; Amundsen, 2001; Cornelia, 2007). Short tunnels had higher crash rates than long tunnels although the chance of crash occurrence is lower in a tunnel than on an open highway section. Amundsen et al. found that the crash rates are higher in the entrance of tunnels and decreased as drivers continued driving inside the tunnel (Amundsen, 2001). However, a study of Chinese freeway tunnels found an increased number of crashes inside the tunnel in contrast to previous studies (Ma, 2009). Potential reasons were thought to be higher driving speeds and insufficient sight distance, which is required for higher design speeds (Ma, 2009).

ABOUT

Traffic Safety Guide is part of an AI4Good project series, between the AI for Good Foundation and the Applied Data Science with Venture Applications Course at SCET, UC Berkeley.

THE PRINCIPAL QUESTIONS

We are concerned with assessing driving risk levels for individual road segments and predicting the severity of a traffic accident given that one has occurred. Relevant features we will investigate include: road and intersection design, time of day, traffic density, type or size of road and regional variations (possibly a result of different driver behavior). For this study, our team will focus solely on traffic data in Chicago.

GOALS

To build a product that  would aid in the prediction of potential accident-prone areas and suggest alternative approaches to mitigate accident frequency and severity.

This database approach would be, once trained, fast to predict. As such, designers can easily analyze many different potential road designs and allow designers to more likely find the safest solutions. In the end, the designers can use our product to design safer roads, which would save lives.

ABOUT

Traffic Safety Guide is part of an AI4Good project series, between the AI for Good Foundation and the Applied Data Science with Venture Applications Course at SCET, UC Berkeley.

THE PRINCIPAL QUESTIONS

We are concerned with assessing driving risk levels for individual road segments and predicting the severity of a traffic accident given that one has occurred. Relevant features we will investigate include: road and intersection design, time of day, traffic density, type or size of road and regional variations (possibly a result of different driver behavior). For this study, our team will focus solely on traffic data in Chicago.

GOALS

To build a product that  would aid in the prediction of potential accident-prone areas and suggest alternative approaches to mitigate accident frequency and severity.

This database approach would be, once trained, fast to predict. As such, designers can easily analyze many different potential road designs and allow designers to more likely find the safest solutions. In the end, the designers can use our product to design safer roads, which would save lives.

DATA CATALOGUE

EXISTING RESEARCH

  1. American Association of State Highway and Transportation Officials, A Policy on Geometric Design of Highways and Streets, AASHTO, Washington, D.C., 2001.
  2. Amundsen, F.H., Engebretsen, A. and Ranes, G., 2001. Car Fires and Breakdowns in Norwegian Road Tunnels 1990-2000. Norwegian Public Roads Administration, Oslo.
  3. Bassan, S., 2016. Overview of traffic safety aspects and design in road tunnels. IATSS research, 40(1), pp.35-46.
  4. Cornelia, N., 2007. Comparative Analysis of Safety in Tunnels. In Young Researchers Seminar 2007.
  5. Çepni, M.S. and Arslan, O., 2017. A GIS Approach to Evaluate Infrastructure Variables Influencing the Occurrence of Traffic Accidents in Urban Roads. International Journal of Environment and Geoinformatics, 4(1), pp.17-24.
  6. Haddon Jr, W., 1968. The changing approach to the epidemiology, prevention, and amelioration of trauma: the transition to approaches etiologically rather than descriptively based. American journal of public health and the Nations health, 58(8), pp.1431 1438.
  7. Lamm, R., Psarianos, B., Choueiri, E.M. and Soilemezoglou, G., 1995, August. A Practical safety approach to highway geometric design. International case studies: Germany, Greece, Lebanon, and the United States. In International Symposium on Highway Geometric Design Practices (Vol. 30).
  8. Ma, Z.L., Shao, C.F. and Zhang, S.R., 2009. Characteristics of traffic accidents in Chinese freeway tunnels. Tunnelling and underground space technology, 24(3), pp.350-355.
  9. Massoudian, N. and Eshghi, M., 2015, December. Design an intelligent curve to reduce the accident rate. In Signal Processing and Intelligent Systems Conference (SPIS), 2015(pp. 185-188). IEEE.
  10. Mohamedshah, Y.M. and Kohls, A.R., 1994. Accident rates using HSIS. Public Roads, 58(1).
  11. Obaidat, M.T. and Ramadan, T.M., 2012. Traffic accidents at hazardous locations of urban roads. Jordan Journal of Civil Engineering, 6(4), pp.436-447.
  12. Thomas, P., Morris, A., Talbot, R. and Fagerlind, H., 2013. Identifying the causes of road crashes in Europe. Annals of advances in automotive medicine, 57, p.13.
  13. Timothy, R.N., Glennon, J.C. and Saag, J.B., 1983. Accident Analysis for Highway Curves. Transportation Research.
  14. Wedajo, T., Quezon, E.T. and Mohammed, M., Analysis of Road Traffic Accident Related of Geometric Design Parameters in Alamata- Mehoni-Hewane Section.
  15. World Health Organization, 2015. Global status report on road safety 2015. World Health Organization
  16. Ziegler, Charles M. “How roadside features affect traffic accident experience: 2nd progress report on the Michigan study”, Michigan State Highway Department, 1951.
  17. National Vital Statistics Reports Volume 66, Number 6, November 27, 2017. 75. Caliendo Ciro, and Guida Maurizio. Microsimulation Approach for Predicting Crashes at Unsignalized Intersections Using Traffic Conflicts. J. Transp. Eng. 138:1453–1467, 2012.
  18. Erskine, P. INTERSECTION DESIGN. 81, 2006.

RESEARCH TEAM

SEMIH BEZCI

Researcher

ANNA MATSOKINA

Researcher

JONATHAN ONG

Researcher

CLARA LIM

Researcher

ERIC ZHEVEL

Researcher

LUKA BRADESKO

SCIENCE LEAD