Data Sets and Friction Estimation for Safety of Heavy Vehicles

This article discusses the importance of data sets and friction estimation in improving safety of heavy vehicles in challenging conditions. It covers various data sets such as speed, geometric design elements, loading, vehicle characteristics, road conditions, road signs and control systems, and safety management interventions. It also explores different ideas of enhancing friction estimation safety, including the use of measured values from sensors, system models, environment and road sensing, cooperative methods, vehicle dynamics observation techniques, and machine learning.

Software Engineering Fundamentals 1
SOFTWARE ENGINEERING FUNDAMENTALS
Name
Course
Professor
University
City/state
Date

Software Engineering Fundamentals 2
Software Engineering Fundamentals
1. Data sets
Safety of heavy vehicles in challenging conditions is very crucial. One of the best ways
of improving safety of these vehicles is by creating appropriate simulations so as to identify the
right combination of parameters, variables or strategies that can be applied in enhancing safety of
heavy vehicle drivers and other road users. Many studies have found various data sets that can be
used to create simulation models for estimating and improving safety of heavy vehicles. Some of
these data sets include the following:
i) Speed
Speed has a significant impact on safety of heavy vehicles. When heavy vehicles travel at
very high speed, they become vulnerable to high damage in the event of an accident because the
driver cannot control them in case of an emergency. Driving risks increases with increasing
speed standard deviation of the vehicle. In most cases, heavy vehicles tend to reduce their speed
when travelling on comparatively steep longitudinal grades so as to enable the driver have more
control of the vehicle and reduce accident risks (Bassan, 2016). The heavy vehicles should travel
at recommended speed especially when on steep slopes.
ii) Geometric design elements
Safety of heavy vehicles is largely affected by various geometric design elements of the road.
Some of the data sets that should be used include: horizontal curvature (transition curve, super-
elevation, curve radius, sight distance, etc.), vertical alignment/grade (gradients, crest curves, sag
curves and vertical curves), cross slope, etc. In general, various data about the complexity of
road alignment should b considered.

Software Engineering Fundamentals 3
iii) Shoulder width
When the rod width is small, the driver cannot control the vehicle properly. Small road width
also increases the likelihood of heavy vehicles colliding with other vehicles such as passenger
cars. Therefore lanes should be of adequate width so as to improve safety of heavy vehicles.
iv) Loading
Many studies have shown that the risk of infrastructure damage and traffic accidents increase
when heavy vehicle exceed the legal and acceptable weight limits. When a heavy vehicle is
overloaded, its kinetic energy increases significantly thus increasing impact forces together with
damage to the infrastructure and other vehicles – in the event of an accident (Jacob & Beaumelle,
2010). Different countries have varied maximum permitted weight limits for heavy vehicles.
Therefore various loadings should be tested so as to determine the suitable mass limits that the
heavy vehicle should be allowed to carry without exceeding legal limits.
v) Vehicle characteristics
There are several vehicle characteristics that can be used to improve safety of heavy vehicles
in challenging conditions (Tagesson, 2014). Some of these include: braking system, steering
system, driver assistance systems, anti-roll bars, and other sensing systems, which are essential
in controlling directional stability of heavy vehicles (Tagesson, et al., 2016). These systems are
very crucial in situations such as when cornering or when the vehicle lose control (such as lateral
stability and rollover) (Trigell, et al., 2017). Drivers should understand and use them effectively
to improve safety on the road.
vi) Tunnels

Software Engineering Fundamentals 4
Movement of heavy vehicles through tunnels is largely restricted as drivers cannot perform
some actions such as a U-turn maneuver. Therefore when a road has tunnels, any physical
obstacles that may restrict maneuverability of heavy vehicles should be prevented or minimized
(Caliendo, et al., 2013). The data to be used when creating road tunnels’ crash-prediction models
include: tunnel length, number of lanes, annual average daily traffic per lane, presence of
sidewalk, and percentage of heavy vehicles. This data can be analyzed used methods such as
Bivariate Negative Binomial regression model, Random Effects Binomial regression model,
Negative Multinomial regression model, Inverse Gaussian regression model, Maximum
Likelihood Method, and Cumulative Residual method, among others (Meng & Qu, 2012).
vii) Road conditions or pavement surface characteristics
The condition of the road also affects the safety of heavy vehicles. When the road is in poor
condition, the risks of accident are very high. For instance, a road that is full of potholes or one
that is worn out makes it difficult for the driver to control the vehicle. Poor roads also reduces
grip between the vehicle tires and the road, making it easy for the vehicle to slip. For instance,
when friction level is different between the right and left side of the vehicle, even braking
becomes a challenge for drivers (Tagesson, et al., 2014). It is therefore important to create
simulations showing high possibility of heavy vehicle accidents on poor roads as this may drive
governments and other relevant private stakeholders to maintain roads.
viii) Road lighting, light condition of the vehicle and weather conditions
When weather conditions, light condition of the vehicle and road lighting are poor, the heavy
vehicle and driver cannot perform as expected. Poor weather and lighting conditions make it
difficult for the driver to see clearly and may accidently hit oncoming vehicles, pass over bumps

Software Engineering Fundamentals 5
at high speed or find it difficult to negotiate a corner. Therefore road lighting, light condition of
the vehicle and weather conditions should be used to create simulations that will determine when
and how the driver should use lighting of the vehicle, depending on daylighting level (which may
be affected by fog, excess smoke, heavy rain, etc.). Weather conditions such as strong winds and
thunderstorms also affect safety of heavy vehicles and should be used to create safety models of
heavy vehicles. These conditions largely affect drivers’ behavior, ability and flexibility to handle
unstructured elements (Tagesson, 2017).
ix) Road signs and control systems
Properly painted and signalized roads are less risky to accidents because drivers know in
advance what is ahead and what to do so as to avoid accidents. These systems are very essential
along steep or meandering roads, at intersections and mid-blocks, etc. Therefore data about
control systems and road signs should be used to create models that improve safety of heavy
vehicles.
x) Safety management interventions
This is another very important, but sometimes overlooked, data set that has a huge impact on
safety of heavy vehicles. When designing a comprehensive plan to improve safety of heavy
vehicles in challenging conditions, it is important to consider the following factors: safety
training, management commitment, journey planning and scheduling, environmental and vehicle
conditions, involvement of employees, support/communication, incentives, type of freight
(freights such as flammable products increases severity of damage in case of an accident), and
safety technologies of the vehicle, among others (Mooren, et al., 2014).
2. Ideas of enhancing friction estimation safety

Software Engineering Fundamentals 6
Friction is a major contributing factor to safety of heavy vehicles. Therefore it is important to
calculate tire-road friction coefficient when developing safety models for heavy vehicles in
challenging conditions. Some of the ways of enhancing friction estimation include:
Using measured values from yaw rate sensors and wheel angular velocity – this involves
identifying longitudinal and lateral velocities using Kalman filter observer, estimating
longitudinal and lateral tire forces using recursive least square algorithm, and estimating friction
coefficient using estimated values from the previous stages and multilayer perception neural
network.
System models – this is where friction estimation is done using models such as single
wheel model, anisotropic brush model, vehicle lateral dynamics, enhanced adaptive observer,
recursive least square method, etc.
Environment and road sensing – this involves use of sensors to determine various
parameters affecting tire-road friction.
Cooperative methods – this involves use of information shared between vehicles and road
infrastructure. The method combines environmental and road sensing methods together with
vehicle dynamics observation methods.
Vehicle dynamics observation techniques – this method is used to determine tire
behavior, depending on tire loading. Some of these methods include: longitudinal dynamics
based, lateral dynamics based, and combination of lateral and longitudinal dynamics (Prokes,
2015).
Machine learning – this ideas involves used of data collected from weather stations and
historical friction data collected from connected vehicles. The road friction coefficients are then

Software Engineering Fundamentals 7
calculated using models such as support vector machine, logistic regression and neural networks
(Ghazaleh, Nasser & Zenuity, 2017).
In general, ideas of estimating friction for heavy vehicles entail use of in-vehicle
measurements to observe tire behavior.
References
Bassan, S., 2016. Overview of Traffic Safety Aspects and Design in Road Tunnels. IATSS Research, 40(1),
pp. 35-46.
Caliendo, C., De Guglielmo, M. & Guida, M., 2013. A Crash-Prediction Model for Road Tunnels. Accident
Analysis & Prevention, Volume 55, pp. 107-115.
Jacob, B. & Beaumelle, V., 2010. Improving Truck Safety: Potential of Weigh-in-Motion Technology.
IATSS Research, 34(1), pp. 9-15.
Meng, Q. & Qu, X., 2012. Estimation of Rear-End Vehicle Crash Frequencies in Urban Road Tunnels.
Accident Analysis & Prevention, Volume 48, pp. 254-263.
Mooren, L., Grzebieta, R., Williamson, A. & Friswell, R., 2014. Safety Management for Heavy Vehicle
Transport: A Review of the Literature. Safety Science, Volume 62, pp. 79-89.
Ghazaleh, P.G., Nasser, M., Zenuity, A.B., 2017. Road Friction Estimation for Connected Vehicles Using
Supervised Machine Learning. Goteborg, Sweden: Chalmers University of Technology.
Prokes, J., 2015. Realtime Estimation of Tyre-road Friction for Vehicle State Estimator. Goteborg,
Sweden: Chalmers University of Technology.
Tagesson, K., 2014. Truck Steering System and Driver Interaction. Goteborg, Sweden: Chalmers
University of Technology.
Tagesson, K., 2017. Driver-centered Motion Control of Heavy Vehicles. Goteborg, Sweden: Chalmers
University of Technology.
T Tagesson, K; Eriksson, B; Hulten, J; Pohl, J; Laine, L; Jacobson, B., 2016. Improving Directional Stability
Control in a Heavy Truck by Combining Braking and Steering Action. Goteborg, Sweden, Chalmers
University of Technology.
Tagesson, K., Jacobson, B. & Laine, L., 2014. Driver Response to Automatic Braking under Split Friction
Conditions. Tokyo, JSAE.

Software Engineering Fundamentals 8
Trigell, A., Rothhamel, M., Pauwelussen, J. & Kural, K., 2017. Advanced Vehicle Dynamics of Heavy
Trucks with the Perspective of Road Safety. International Journal of Vehicle Mechanics and Mobility,
55(10), pp. 1572-1617.