Emerging Techniques for Effective Road Safety

Management

Abstract. The rise in the number of deaths and accidents on the roads has drawn the
attention of stakeholders to the need for more effective "road safety management"
procedures. The term "road safety management" refers to the coordinated efforts put
into planning, analyzing, and putting into action various road safety operations and
tactics (RSM). The majority of research efforts have been focused on analyzing the
structure of RSM plans and outlining the RSM best practices that are currently in use.
Because of the subjective nature of the project circumstances and the structural
intricacy of the RSM design, selecting an emerging technique might be challenging.
This endeavor is made more difficult by the absence of straightforward,
understandable literature that is relevant to RSM techniques. As a result, the purpose
of this study is to expand upon the RSM techniques, as well as their respective
applications, in more depth. This study analyses the efficiency of the most often used
techniques and advance tools during a certain phase of a project, taking into account
any potential duplication that may occur. The findings of this study will be
summarized into within conclusion that will be useful for highway agencies in their
efforts to analyze current road safety conditions, identify potential safety issues, and
locate positive strategies to enhance road safety.

Keywords. Road Safety Management, Techniques, Tools, Project phase, Application

1 Introduction
The integrated task of planning, evaluating, and implanting road safety activities and
strategies is defined as Road Safety Management (RSM). The RSM system includes
policies related to road safety improvements with details and activities and
implementation of safety programmes with continuous examinations and development
[1]. Some literature suggests that the proper implementation of the RSM plan concludes
with good advancements within existing road safety practices [1][2][3]. Nicole
Muhlrad et al. [4] developed a model of investigation as part of DaCoTa (Road Safety
Data Collection, Transfer, and Analysis) project. The model supports the performance
investigation of any RSM planned system by introducing qualitative variables.
However, the assumed ideal RSM system within DaCoTa model did not provide clear
results for achieving the highest level of road safety [4]. Some literature elaborated on
the relationship between RSM components and safety performance with the help of
road accident data [5][6]. But the comparison of different RSM plans for different
countries is difficult as complex structures. Countries having similar RSM plans may
get different results due to different administrative structures and local governance.
Wegman et al. [7] investigated the road accident scenario of Nine European countries and
recommended an RSM system with intervention for every single county. The detailed
analysis concluded that some important components of an ideal RSM are: 1) Defining rod
safety responsibilities to a particular ministry; 2) establishment of a centralized road safety
governing agency to channel safety funds with countrywide safety activities; 3) setting the
safety-centric target; 4) intensification of road safety improvement related resources; 5)
improvement in stakeholders engagements within the RSM structure; 6) improvement in
accident data collection systems. Supporting the same findings, Bliss and Breen [3]
concluded some recommendations within their report to improve countrywide road safety
performance with an integrated RSM system. The report suggested the foundation of the
RSM system with the safe system approach and with the principles of institutional
management. Johnston summarized the key to the best RSM system as the concept of
“4 Cs”, by integrating Constituency, Commitment, Cooperation, and Coordination [8].
But complex structures and systems are often lacking with the procedural integration
of RSM tools. RSM procedures and tools are aimed at improving road safety at different
road infrastructure project stages [9]. Many of them can be functional with the available
road sections, allowing a supplementary responsive attitude, while the rest are applied in
the planning and designing stages (i.e., planning and design) with a more active attitude.
Therefore, the need for a clear and easy elaboration of RSM tools is evident to ensure
the effective implementation of RSM plans. The present paper aims to provide a
detailed overview of RSM tools and highlights the application of each tool.

2 Road safety management (RSM) Techniques and tools

2.1 Safety Audit (SA)
A safety audit (SA) is a methodical evaluation of drawings/plans/ documents of new road
schemes/projects. The audit procedure targets to minimize possible road crashes in the
future by eliminating hazardous features. Thus, it is a practical tool of RSM. Great Britain,
Denmark, and Australia are the countries where RSAs were executed for the first time
around 1990. Since then, the RSA has become a benchmark process for most road
infrastructure projects around the world. Every country has its standards, checklists, and a
distinct RSA procedure. Generally, road safety audits are:
 Performed by a certified person who has been qualified and accredited as a ‘road
safety auditor’ by a country-specific government body
 Performed in a pre-standardized format of checklists and within which the
compiling and appraisal of other audits are also possible
 Organised in a way that safety auditors are independent of the design & planning
team of a consulting project
 Documented in the form of an ‘audit report’ with details of specific improvements
that must be done to ensure the safety of road users
  Responded by designing, planning, and operating agencies with a compliance
report to the safety auditor
2.2 Safety Inspections (SI)
A Safety Inspection (SI) is an organized safety evaluation tool applied to an operating road
stretch. SIs have also referred to as RSAs for already constructed and fully operating roads.
The RSI process aims to classify hazardous components of a roadway that are unnoticed
within the accident records or to identify the novel issues of roadway engineering design
and modification that can lead to an accident. Therefore, the process of SI remains the same
as SA, with a change in the safety checklist. Ultimately, RSI is classified as a thematic
examination such as an assessment of road signs. Therefore, SI covers a detailed analysis
of every possible aspect of an existing road than a common check. Recent advances
concludes SI and RSA with the use of virtual reality (VR) tools and drone usage [10].
2.3 Network Screening (NS)
Network screening (NS) technique and tools are applied where the deviation in the road
accident numbers of different stretches of a roadway network is statistically investigated.
The ultimate object of NS is to detect road stretches with the following type of safety issues:
 Strangely high accident numbers
 A higher proportion of fatal/ severe accidents
 A higher proportion of a certain accident type
Many NS tools are available, such as simple rankings of road stretches with available
accident datasets and advanced statistical techniques with accident prediction modeling.
There are many applications and software available to perform NA with existing road
accident data stretches such as Safety Analyst [11] and RISMET [12].
2.4 Accident Modelling (AM)
Accident models (AMs) are based on the advanced regression analysis methods which
statistically assessing the variation of accident/ fatalities/ injury numbers with a change in
a set of premeasured variables. This RSM tool aims to evaluate the influence of different
factors on accident and injury numbers and conclude their extent and effects. With advanced
data structure modeling and mathematical advances, research advances in AM-based RSM
tools are demanding. The following are some crucial components of the accident model
(AM):
 Accident models must have a large sample with a higher mean of accident numbers
[13]
 Variable data must be recorded in such a way that the variation of between-section
variation is maximized and within-section is minimized [14]
 Endogeneity bias control attribute must be controlled within the accident modeling
analysis with safety improvement/ treatment variables [15]
2.5 Road protection scoring
Road protection scoring provides an idea about the road’s forgivingness. As of now, many
scoring systems have been already established in different parts of the world. International
Road Assessment Programme (iRAP) has a unique and widely applicable road protection
scoring system, which provides star ratings of a road stretch out of five [16]. This scoring
system has been established in New Zealand (KiwiRAP), Australia (AusRAP), Europe
(EuroRAP), and, India (IndiaRAP). The protection scoring system allows to record the
safety features of road stretches and concludes the road safety risk in the form of star ratings.
These results are published in the form of graphs to inform about the road safety levels of
a particular road stretch. A higher star rating suggests the highest level of safety with the
highest level of road forgiveness.
2.6 Identification and analysis tool of accident blackspots/ hazardous locations
Most countries are having their methods for concluding and evaluating hazardous locations/
accident blackspots. But these methods are insufficient to be a great RSM tool [17]. The
state-of-the-art blackspot identification and evaluation methods must have the following
characteristics:
 The desired number of sites must be evaluated to identify accurate numerical
criteria for selecting the hazardous location
 Sliding window approach must be avoided to finalize the hazardous location
 The Empirical Bayes (EB) estimation of road safety must be applied to precisely
identify the hazardous location by considering the expected road accidents
numbers
 Hazardous locations/ blackspot must lie within the upper-most percentiles of road
site distribution concerning the expected road accidents numbers
 Past 3 to 5 year accident data records are suitable for the identification of the
hazardous location
 Accident injury severity levels must also be considered while analyzing hazardous
location

2.7 Impact assessment

Impact assessment (IA) is a planning technique that estimates the desired impact of
recommended safety measures and investments on accident severity. Many software
performs a cost-benefit analysis of road improvements and performs an impact assessment
of safety measures. Nonetheless, most applications of this software are limited to large-
scale road projects and major road upgradation programs. In general, the safety investments
are small and of lower implementation costs. Some literature suggests the impact of these
regular minor road improvements from which tool development of minor IA becomes
possible [9].
Fig. 1. Impact Assessment in PTV VISSIM [18]

2.8 Road user behavior monitoring tools

Road users’ behavior influences a safer road environment. Therefore, it becomes crucial to
analyze the behavior of every possible road user over time. The most supervised behavioral
aspects are:
 Speeding
 Use of seat belt
 Use of helmets
 Driving Fatigue
 Intoxicated driving
 Pedestrian behavior
The lack of tools with high accuracy and high data reliability for user behavior stuck is
evident. Further, there is no reliable device present in the market which directly detects the
use of mobile phones and other devices while driving for a longer period. The modern
technology with artificial intelligence (AI) and machine learning (ML) architecture may
detect driver distraction with relatively high accuracy [19] [20].
2.9 Conflict study tool
Conventional traffic conflict study investigates the available images and videos through
proper software. Manual coding and observations are applied as subjective conflict study
techniques. The observers and coders may need proper training for effective conflict
analysis via this kind of software. Advanced technologies eliminate human interference in
traffic conflict studies via image processing and geo-referenced stereo sequences [21].
Some techniques estimate the trajectory and speed of road users to detect collision time
[22].

Fig. 2. Reference coordinate system for conflict study using Data Analysis Module [21]

There are also some priority-based tools and methods which also help to strengthen the
application of RSM [23] [24].

3 Possible integration of techniques and tools with standard RSM

procedure
Conventionally, there are ten major procedures involved in RSM [25]. First, Road safety
impact assessment (RSIA) is carried out before project approval of a new road or before
applying any modification to the existing road. Efficiency assessment (EA) of prioritizing
safety budgets is a second procedural step of RSM. Impact assessment (IA) tools generally
conduct RSIA and EA with ease. Then, the procedure of Road safety audit (RSA) and Road
safety inspection (RSI) is conducted by Safety Audit (SA) and Safety Inspection (SI).
Further, SI is also a helpful tool to analyze, identify and prioritize accident-prone stretches
as High Risk Site (HRS) investigation process. Road network serviceability and safety are
ensured by Network Operation (NO) process with the help of the network screening (NS)
tool, too. Figure IV represents all possible integrations between RSM procedure and tools.
Fig. 3. Possible integration of techniques and tools with standard RSM procedure

Here, it is interesting to note that some RSM procedures and tools have a similar application.
For example, RSA, RSI, and HRS have a similar purpose to the application of SA, SI, and
RPS. The data required, time of application, and methodology to apply these procedures
and tools are quite different.

4 STAGE-WISE APPLICATION
The road infrastructure projects are divided into the following six stages [9]:
 1. Planning and Construction
2. Opening to traffic and initial adjustment stage
3. Normal operation stage
4. Periodic inspection, maintenance, and renewal of equipment
5. Correlation of error and treatment of hazardous locations
6. Major upgrading and renewal

Each of these stages has its unique operation and importance in the RSM procedure.
Therefore, every stage has its applicability to RSM tools. Table 1 neatly represents the
phase-wise applicability of RSM tools. Accept, network screening, and accident modeling,
every tool is applicable within more than one infrastructure project stage. A properly
audited road section in the planning and construction stage may directly eliminate the
application of identification tools for hazardous locations [9]. But, the continuously
changing traffic patterns with other road parameters arise with the application of most RSM
tools.

Table 1. Phase-wise applicability of RSM techniques and tools

Tool Project phase

Safety audit (SA) Planning & construction stage,
Opening to traffic and initial adjustment
phase
Safety inspection (SI) Operation stage,
Periodic inspection, maintenance, and
renewal of equipment,
Major upgrading and renewal stage
Network screening (NS) tool Operation stage
Accident modeling (AM) Operation stage
Road protection scoring Correction of errors and treatment of
hazardous locations
Identification and analysis tool of Correction of errors and treatment of
accident hazardous locations/ hazardous locations
blackspots
Road user behavior monitoring tool Normal operation, Correction of errors,
and treatment of hazardous locations
Impact assessment (IS) Correction of errors and treatment of
hazardous locations
Conflict studies Correction of errors and treatment of
hazardous locations
5 CONCLUSIONS
Substantial studies have been directed around safety management (RSM) to identify the
structure, evaluate the process, and prepare proper plans. However, the changing traffic
conditions and human behavior make it difficult to apply RSM on a larger scale without the
required tools. These tools satisfy the predefined procedures of RSM and satisfy the
procedural aims within every project stage. Safety audits (SAs) are applied several times
through the planning and construction stage. Similarly, Safety inspections (SIs) are
practiced for the operational stage of the road to detect defects and errors. Network
screening (NS) tools and accident modeling (AM) are applicable to prioritize road sections
based on the accident data and to analyze the possible variation in the number of accidents
within a particular road stretch or network. Road user behavior monitoring tools are the
most advanced tools of RSM with multipurpose applications. It provides crucial
information on normal and risky user behavior. This information becomes the foundation
for the special RSM plans to improve defective road behavior. Furthermore, road protection
scoring and identification and analysis tools for blackspots highlight the factors which may
lead to an accident. Finally, the conflict study tools are most important to establish the
interaction between road infrastructure and human factors. All applications of RSM tools
complete every aspect of road accident analysis and integrate most procedures of RSM.
Proper integration of these tools within the indicated project phases in research may
improve the overall effectiveness of RSM.

REFERENCES
[1] N. Muhlrad, Road safety management systems, a comprehensive diagnosis method adaptable to low and
middle income countries. 2009.
[2] J. Shah, K. Bhatt, P. Trivedi, and S. Easa, “Road Safety Conditions and Management in India:
Challenges and Opportunities,” pp. 3–23, 2022, doi: 10.1007/978-981-16-8837-9_1.
[3] T. Bliss and J. Breen, “World Report on Road Traffi c Injury Prevention Country Guidelines for the
Conduct of Road Safety Management Capacity Reviews and the Specifi cation of Lead Agency Reforms
, Investment Strategies and Safe System Projects,” 2004.
[4] R. Alfonsi, L. Persia, T. Antonino, and D. S. Usami, “Advancements in Road Safety Management
Analysis,” Transp. Res. Procedia, vol. 14, pp. 2064–2073, 2016, doi: 10.1016/j.trpro.2016.05.105.
[5] S. C. Wong and N. N. Sze, “Is the effect of quantified road safety targets sustainable ?,” Saf. Sci., vol.
48, no. 9, pp. 1182–1188, 2010, doi: 10.1016/j.ssci.2009.12.020.
[6] R. Elvik, “Road safety management by objectives : A critical analysis of the Norwegian approach,” vol.
40, no. 0001, pp. 1115–1122, 2008, doi: 10.1016/j.aap.2007.12.002.
[7] F. Wegman, F. C. Engineering, J. Commandeur, V. U. Amsterdam, E. Doveh, and V. Gitelman,
“SUNflowerNext: Towards a composite road safety performance index,” 2008.
[8] I. Johnston, “Beyond ‘ best practice’ road safety thinking and systems management - A case for culture
change research,” Saf. Sci., vol. 48, no. 9, pp. 1175–1181, 2010, doi: 10.1016/j.ssci.2009.12.003.
[9] R. Elvik, “Assessment and applicability of road safety management evaluation tools: Current practice
and state-of-the-art in Europe,” 2010.
[10] J. Seymour, A. Gross, C. Chestnutt, and M. C. Drive, “Preparing for a Virtual Road Safety Audit ( RSA
),” 2020.
[11] D. W. Harwood, D. J. Torbic, K. R. Richard, and M. M. Meyer, “SafetyAnalyst-Software Tools for
 Safety Management of Specific Highway Sites,” 2010.
[12] G. Schermers et al., “Guidelines for the development and application of evaluation tools for road safety
infrastructure management in the EU,” Road Infra¬ Struct. Saf. Manag. Eval. Tools, 2011.
[13] D. Lord and L. F. Miranda-Moreno, “Effects of Low Sample Mean Values and Small Sample Size on
the Estimation of the Fixed Dispersion Parameter of Poisson-Gamma Models for Modeling Motor
Vehicle Crashes: A Bayesian Perspective,” Saf. Sci., vol. 46, no. 5, pp. 751–770, 2008, doi:
10.1016/j.ssci.2007.03.005.
[14] S. Cafiso, A. Di Graziano, G. Di Silvestro, G. La Cava, and B. Persaud, “Development of comprehensive
accident models for two-lane rural highways using exposure, geometry, consistency and context
variables,” Accid. Anal. Prev., vol. 42, no. 4, p. 1072—1079, Jul. 2010, doi: 10.1016/j.aap.2009.12.015.
[15] D.-G. Kim and S. Washington, “The significance of endogeneity problems in crash models: an
examination of left-turn lanes in intersection crash models.,” Accid. Anal. Prev., vol. 38, no. 6, pp.
1094–1100, Nov. 2006, doi: 10.1016/j.aap.2006.04.017.
[16] International Road Assessment Programme IRAP, “iRAP Star Rating and Investment Plan Coding
Manual,” 2014.
[17] J. Stipancic, L. Miranda-Moreno, N. Saunier, and A. Labbe, “Surrogate safety and network screening:
Modelling crash frequency using GPS travel data and latent Gaussian Spatial Models,” Accid. Anal.
Prev., vol. 120, no. January, pp. 174–187, 2018, doi: 10.1016/j.aap.2018.07.013.
[18] W. Kustra, K. Jamroz, and M. Budzynski, “Safety PL- A Support Tool for Road Safety Impact
Assessment,” Transp. Res. Procedia, vol. 14, no. November, pp. 3456–3465, 2016, doi:
10.1016/j.trpro.2016.05.308.
[19] J. Mafeni Mase, P. Chapman, G. P. Figueredo, and M. Torres Torres, “Benchmarking Deep Learning
Models for Driver Distraction Detection,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif.
Intell. Lect. Notes Bioinformatics), vol. 12566 LNCS, no. May, pp. 103–117, 2020, doi: 10.1007/978-3-
030-64580-9_9.
[20] F. Omerustaoglu, C. O. Sakar, and G. Kar, “Distracted driver detection by combining in-vehicle and
image data using deep learning,” Appl. Soft Comput. J., vol. 96, p. 106657, 2020, doi:
10.1016/j.asoc.2020.106657.
[21] S. Battiato, S. Cafiso, A. Di Graziano, and G. M. Farinella, “Road Traffic Conflict Analysis from Geo-
Referenced Stereo Sequences,” in Image Analysis and Processing - ICAP 2013, 2013, vol. 8156, no.
September, doi: 10.1007/978-3-642-41181-6.
[22] A. Laureshyn, “Application of automated video analysis to road user behaviour,” Lund Institute of
Technology, 2010.
[23] P. Trivedi and J. Shah, “Identification of Road Crash Severity Ranking by Integrating the Multi-Criteria
Decision-Making Approach,” J. Road Saf., vol. 33, no. 2, pp. 33–44, 2022.
[24] P. Trivedi and J. Shah, “Road Crash Severity Ranking by Applying a Multi-criteria Decision-Making
Tool: Analytical Hierarchy Process,” in Intelligent Infrastructure in Transportation and Management,
J. Shah, S. S. Arkatkar, and P. Jadhav, Eds. Springer Singapore, 2022, pp. 123–129.
[25] International Transport Forum, “Road Infrastructure Safety Management Road Infrastructure Safety
Management,” 2007.