Study of the greenhouse gas emissions and fuel

consumption effect, caused by passive and

aggressive behavior in driving public service
buses in Pasto city⋆

Fredy Dulce1,3[0000−0001−8937−2334] , Jackeline Murillo2[0000−0003−0644−0923] , and

Eduardo Caicedo1[0000−0003−0727−2917]
1
Programa de Posgrado en Ingenierı́a Eléctrica y Electrónica - PPIEE, Universidad
del Valle, Cali, Colombia
2
Escuela de Ingenierı́a Civil y Geomática - EICG, Universidad del Valle, Cali,
Colombia
3
Programa de Ingenierı́a Electrónica, Universidad CESMAG, Pasto, Colombia

Abstract. Public transport systems have great importance in planning

any city; currently, the increase in population is not the only factor to
consider, but also the pollution levels and the social impact. In general,
the public transport system in Pasto has aggressive behavior from the
bus drivers due to the goals that they should achieve. These goals are
related to the route time and directly affect the salary of the bus drivers,
which is why they tend to have aggressive driving behavior. This article
focuses on the simulation of driving behaviors, aggressive and passive,
considering the rush hour and peak-off-hour traffic to determine what
effect it has on fuel consumption and greenhouse gas emissions. Four
public transport routes were used in Pasto city as the study case. The
simulations were carried out in Eclipse SUMO, a widely used tool for per-
forming complex analyses that include many variables to make a realistic
enough simulation.

Keywords: Driving behavior · Rush hour · Peak-off-hour · Public trans-

port · Greenhouse Gas Emissions · SUMO

1 Introduction

It is very known that the transport sector is one of the most significant con-
tributors to greenhouse gas emissions, making the research related to this sector
relevant. Considering this, Colombia aims to reduce greenhouse emissions by 36
million tCO2eq, and one of the challenges is to migrate forward to sustainable
transport due to 78% of these emissions coming from this sector [1]. Vehicle fuel
⋆
This work has been supported by the project “Desarrollo de un modelo alternativo
de energı́a y movilidad con fuentes no convencionales en la Universidad de Nariño”
BPIN 2020000100041
2 F. Dulce et al.

consumption and carbon emissions are influenced by driving behaviour, as well

as traffic conditions and average driving distance for safety [2].
Traffic-related emissions, such as CO2, CO, NOx and HC, have negatively
affected the environment and led to poor air quality [3]. Highlighting that CO2
has the most significant impact [4]. According to [5], improving the driving con-
ditions in major Latin American cities could bring potential economic benefits
on the order of billions of US dollars per year.
For example, in the case of Bogotá, implementing driving patterns similar to
those represented by test driving cycles used in Europe or Japan could reduce
CO, CO2, and HC emissions by at least 11% and by as much as 20% for NOx.
Modifying driving conditions usually requires hefty, expensive, and difficult-to-
build infrastructure. In this context, eco-driving becomes a cost-effective alter-
native to reduce vehicle emissions, especially if the practice is prioritized for
large-engine vehicles and those lacking emission control technologies [5].
Considering the previous fact, we aim to assess the driving behavior (passive
and aggressive) on public bus transport systems, adding the rush hour and peak-
off-hour traffic to determine its effect on fuel consumption and greenhouse gas
emissions. For this, we take the Strategic Public Transport System (SETP for
its acronym in Spanish) of Pasto city as a study case. Moreover, using all input
variables such as speed, routes, stops, acceleration, driving behavior, rush hour
and peak-off-hour, Euro norm, and vehicle type allows us to simulate the reality
of the city, which is a significant contribution.

2 ”Simulation of Urban MObility” (SUMO)

SUMO is an open-source, highly portable, microscopic, and continuous traffic

simulation package designed to handle large networks. It allows for intermodal
simulation, including pedestrians, and comes with a large set of tools for scenario
creation. It was mainly developed by the Institute of Transportation Systems
employees at the German Aerospace Center. [6].
Several studies have used SUMO to simulate road systems with a high level
of detail, thus the simulations and their results can be as realistic as possible.
SUMO has two emission and fuel consumption models, the Passenger Car and
Heavy-Duty Emission Model (PHEM) and a model based on the Handbook of
Emission Factors for Road Transport (HBEFA). Both models are based on Euro-
pean Union Standards (EUS). The HBEFA model can be empirically calibrated
for each vehicle class (passenger, bus, motorcycle, etc.). The HBEFA model has
also been shown to provide more stable results than PHEM on different SUMO
simulation setups. [7].
Articles like [8] create a traffic prediction application that can help solve
congestion problems that have become an annual routine. On the other hand,
research like [9] uses this tool to simulate a scheduling strategy for charging and
dispatching electric heterogeneous Bus Rapid Transit (BRT) fleets equipped with
batteries.
 Title Suppressed Due to Excessive Length 3

Research has even been carried out that contributes to the prevention of
traffic accidents. A clear example is a work presented in [10] that surveys various
technologies and applications simulated with SUMO for multiple purposes and
concludes that SUMO is an effective tool for minimizing accidents in Industry
4.0 era, where cars are talking with other vehicles and devices on the run.
In short, it is possible to find a large number of similar researchs, determining
that SUMO is a powerful tool for the simulation of realistic vehicular traffic
scenarios in any city regardless of its characteristics, such as population size or
vehicles in transit.
For this article, SUMO is used to analyze driving behavior, adding the rush
hour and peak-off-hour traffic to determine what effect it has on fuel consumption
and greenhouse gas emissions with the SETP of Pasto city as the study case.

3 Methodology

The simulation process was defined as follows. First, the Pasto city map with
the road network was downloaded from Open Street Map (OSM). Secondly, four
routes of the SETP were defined to make the simulations, and we created the
road corridors in NetEdit. For this, the streets were extracted from the SETP
website [11]. The next step involves defining the vehicle characteristics such as
speed, acceleration, deceleration, emission standards, and weight, among others.
Finally, the bus stop points for passenger boarding and alighting are defined.
On the other hand, two simulation scenarios were defined. These are periods
of high demand (morning and afternoon rush hour) and periods that consider
free-flow volume, i.e., without congestion. Because of that, the bus in the simula-
tion will have more stops during rush hour and fewer stops during peak-off-hour.
Likewise, the duration of stops is larger during rush hour than peak-off-hour.
Subsequently, the acceleration of the buses is changed. The work carried
out in [13] was taken as a reference to calibrate the acceleration. It has been
estimated that aggressive driving occurs when the acceleration is carried out
with 4.0 m/s2, and passive driving is assumed when the driver accelerates at a
rate of 2.0 m/s2.

3.1 Road Corridors

Pasto city had about 825 buses until 2017, with an average age of 12 years for
the entire fleet; the fuel used for all vehicles is Diesel or ACPM. The SETP
operation is under the organization ”Unión Temporal Ciudad Sorpresa,” which
is made up of four companies. In addition, the fare for a bus ticket is 2000 COP
for the year 2022 and has an approximate annual increase of 100 COP.
Routes E1 and E3 are part of the so-called strategic routes. They have been
assigned this name because they make relatively short journeys to reduce trans-
port times, connecting strategic points such as the city center and some outlying
sectors. On the other hand, routes C1 and C16 are part of the so-called comple-
mentary routes. Unlike the strategic ones, these routes were not designed to have
4 F. Dulce et al.

shorter routes but to connect the most significant number of sectors possible so
that most of the population has access to public transport services and can be
mobilized from anywhere in the city.
The Table 1 and Figure 1 present the set of selected collective public transport
routes.

Table 1. Selected Routes - Collective Public Transportation. Source:

http://181.49.177.91/

Road Approximate
Origin - Destination Distance [Km]
Corridor Time [min]
(North - South)
E1 Barrio Dolores – 10.4 56
Universidad de Nariño
(North - South)
E2 9.4 51
Altos de chapalito – Torobajo
(South - North - West)
C1 Obonuco – Altos de 12.1 59
Chapalito
(East - West Outside
C16 of the municipal capital) 14.6 66
Cabrera/La Laguna – Anganoy

Fig. 1. SETP selected routes, a) E1 route, b) E2 route, c) C1 route and d) C16 route.
Source: http://181.49.177.91/
 Title Suppressed Due to Excessive Length 5

3.2 Critical transport hours

Pasto city has three daily rush hours between 6 am and 7 am, 12 pm and 1 pm,
and 6 pm and 7 pm. Likewise, the peak-off hours are between 9 am and 10 am
and 3 pm and 4 pm. Table 2 shows the average daily travel people make using
the SETP [12].

Table 2. Average of travel per day in the SETP. Source [12]

Hour 6am - 7am 9am - 10am 12pm - 1pm 3pm - 4pm 6pm - 7pm
Number of travels 13000 4000 14000 4000 10000

The people in Pasto start their activities around 7 am, take lunch at 1 pm,
and return to work or study at 2 pm; finally, they finish the activities at 6 pm.
Due to this schedule, there are three rush hours daily. Instead, while most people
work or study, few people need to get away from home, and the number of travel
is minor. Because of that, there are two peak-off hours daily. This behavior is
taken into account too. Thus the simulations could be more realistic and the
results more accurate.

3.3 Simulation Parameters

The simulation parameters are described in Table 3. The parameters acceleration
and acceleration duration have two values because these are the parameters used
to simulate driving behavior, a regular acceleration is obtained with 2.0 m/s2 ,
and an aggressive acceleration is obtained with 4.0 m/s2 .

Table 3. Simulation Parameters

Parameter Value
MaxSpeed [m/s] 16.67
Acceleration [m/s2 ] [2.0, 4.0]
Deceleration [m/s2 ] 3.0
Vehicle Mass [kg] 11000
Emission Class HBEFA3/HDV D EU3

The EURO III standard is used since most of the SETP buses in Pasto
comply with this standard as Figure 2 shows [12].

4 Simulations and results

SUMO is used in conjunction with Python to perform the simulations, which
allows for speeding up the process. It is possible to perform a better analysis of
6 F. Dulce et al.

Fig. 2. Percentage of buses per European standard in Pasto

the information through Python. Figure 3 shows an explanatory diagram of the

communication between SUMO and Python.

Fig. 3. Simulation SUMO-PYTHON

 Title Suppressed Due to Excessive Length 7

4.1 Validation
First, we compared the CO emissions at different speeds to validate that the Euro
III standard is met, for which a trial test was carried out under track conditions.
The result shows that the vehicle complies with the European standard. Once
this was verified, the experiments were carried out with the four selected routes.
The figure 4 shows that all the values are below the maximum value of the
Euro III norm; in this case, the maximum value allowed by the norm is 2.1
[g/kwh] [14]. With this outcome, we can validate the results given by SUMO.

4.2 Results
Figure 5 contains the simulated cases’ outcomes. Each subplot shows the result
for a single variable, for example, fuel consumption in Gl or CO2 emissions in
Kg. Moreover, each subplot has a bar plot with results for aggressive behavior
appointed with ”A” and passive behavior with ”P” on the axis below the figure.
Likewise, there are results related to the traffic, either rush hour or peak-off-
hour traffic. Finally, there are results for each selected route, C1, C16, E1, or
E2. Notably, the CO2 emissions are in kilograms scale while the other emissions
are in grams scale, highlighting that CO2 is the most gas emission generated by
fuel combustion in a diesel bus.

Fig. 4. For the Euro III norm, CO emissions in g/kWh vs. vehicle speed [km/h].
8 F. Dulce et al.

Visibly the emissions and fuel consumption are most considerable when the
bus transit in rush hour than peak-off-hour. Because of that, we calculated the
increasing percentage of rush hour fuel consumption and emissions concerning
peak-off-hour. The results were plotted in the figure 6.

Fig. 5. Total emissions and fuel consumption for each simulated case
 Title Suppressed Due to Excessive Length 9

It is interesting to observe that the general emissions and fuel consumption

increase around 50% or more, only in the C16 case, the increase is around 20%.
Moreover, the increase is most considerable for passive behavior than aggressive
behavior. The above indicates that rush hour traffic affects passive drivers more;

Fig. 6. Percentage increase of emissions and fuel consumption of rush hour traffic
respect peak-off-hour traffic
10 F. Dulce et al.

one possible explanation is that aggressive drivers have high values in both hours
of traffic cases; thus, the increase between each case is not that significant.
Finally, we expected the most significant increase in fuel consumption and
emissions for aggressive behavior than passive behavior. To check it, we calcu-
lated the percentage increase of all aggressive outcomes concerning passive, and
the results were plotted in the figure 7. However, the expected idea was only
fulfilled for fuel consumption, CO2, and NOx emissions, but this idea was wrong
for CO, HC, and PMx emissions.
As figure 7 shows, there are some negative bars. This fact indicates a de-
crease in emissions for aggressive behavior concerning passive behavior. Doing a
literature review, we found two references that support this finding. The articles
show that CO and HC decrease with higher engine loads. The load on the motor
can be assumed to be directly proportional to the acceleration since if the load
is increased, the acceleration must increase so that the force will be greater and
thus maintain the speed approximately constant. In this order of ideas, if the
acceleration increases, the CO and HC tend to decrease. A plausible reason for
the decrease in CO, HC, and PMx emissions could be the increase in combustion
temperature caused by a higher engine load, thereby enhancing the oxidation of
unburnt hydrocarbons in the exhaust. Moreover, aggressive driving character-
ized by rapid acceleration might have required the engine to operate at a higher
peak speed (rpm). A previous study showed that an internal combustion engine
could emit lower concentrations of hydrocarbons with increasing engine speed
[15], [16].
On the other hand, the case of route C1 is different from the others because
this route has more direction changes, which requires a more frequent decelera-
tion and acceleration, preventing the vehicle from reaching a high speed. If the
speed is necessarily low, emissions will be higher in any case. On the other hand,
the other routes do not have as many direction changes, which can allow them
to reach higher speeds through higher accelerations, which is consistent with the
results of HC, CO, and PMx.

5 Conclusions

The results obtained in the four simulation cases show several generalized be-
haviors. First, fuel consumption and gas emissions are always higher during rush
hours. Regarding fuel consumption costs, moving during rush hours is more ex-
pensive than during peak-off-hours. However, the behavior also influences since
passive behavior has lower fuel consumption and, therefore, a lower fuel expense.
Regarding emissions, moving around a traffic jam harms the environment
since emissions at peak hours are much higher than at off-peak hours, around
50% higher. Regarding behavior, it is impossible to state that passive driving
will be better for avoiding significant emissions of polluting gases. This state-
ment is only correct for CO2 and NOx, but in the case of intermediate driving
behavior between aggressive and passive could have fewer emissions in cases of
mobilization within the cities. This fact could be different if mobilizations on
 Title Suppressed Due to Excessive Length 11

highways or between cities, where it is possible to reach higher speeds and less
aggressive behavior regarding acceleration.

Fig. 7. Percentage increase/decrease of emissions and fuel consumption of aggressive

behavior respect passive behavior
12 F. Dulce et al.

References
1. Departamento DE Nacional Planeación. (2019). Plan Nacional de Desarrollo 2018-
2022. pacto por colombia.
2. Keyvanfar, Ali & Shafaghat, Arezou & Muhammad, Nasiru & Ferwati, M Salim.
(2018). Driving Behaviour and Sustainable Mobility—Policies and Approaches Re-
visited. Sustainability. 10. 1152. 10.3390/su10041152.
3. Shafaghat, Arezou, et al. ”Environmental-conscious factors affecting street microcli-
mate and individuals’ respiratory health in tropical coastal cities.” Sustainable Cities
and Society 21 (2016): 35-50
4. Sun, Longsheng, Mark H. Karwan, and Changhyun Kwon. ”Incorporating driver
behaviors in network design problems: Challenges and opportunities.” Transport Re-
views 36.4 (2016): 454-478.
5. Keyvanfar, Ali, et al. ”Driving behaviour and sustainable mobility—policies and
approaches revisited.” Sustainability 10.4 (2018): 1152.
6. sumo.dlr.de. (2022). Documentation - SUMO Documentation. SUMO.
https://sumo.dlr.de/docs/index.html
7. Almatrudi, Sulaiman & Parvate, Kanaad & Rothchild, Daniel & Vijay, Upadhi.
(2022. Using Automated Vehicle (AV) Technology to Smooth Traffic Flow and Reduce
Greenhouse Gas Emissions. 10.7922/G2JS9NRB.
8. Aditya, F., Nasution, S. M., & Virgono, A. (2020). Traffic Flow Pre-
diction using SUMO Application with K-Nearest Neighbor (KNN)
Method. International Journal of Integrated Engineering, 12(7), 98–103.
https://doi.org/10.30880/ijie.2020.12.07.011
9. Ruiz, S., Arroyo, N., Acosta, A., Portilla, C., & Espinosa, J. (2018). An optimal
battery charging and schedule control strategy for electric bus rapid transit. IET
Seminar Digest, 2018(1), 23–30. https://doi.org/10.1049/ic.2018.0004
10. Srivastava, R., & Kumar, V. (2020). Accident avoidance simulation us-
ing SUMO. Proceedings of the 2020 9th International Conference on Sys-
tem Modeling and Advancement in Research Trends, SMART 2020, 179–183.
https://doi.org/10.1109/SMART50582.2020.9337079
11. RUTA C1 - SETPasto - Sistema Estratégico de Transporte Público de Pasto.
(n.d.). Retrieved July 3, 2022, from http://181.49.177.91/index.php/rutas/rutas-
complementarias/c1
12. Movilidad Sostenible LTDA. (2018). Elaboración del Plan Maestro de Movilidad
y Espacio Público para el Municipio de Pasto, Nariño. 170.
13. Perrotta, Deborah & Macedo, José & Rossetti, Rosaldo & Afonso, J.L. & Kokkino-
genis, Zafeiris & Ribeiro, Bernardo. (2014). Driver Attitude and Its Influence on the
Energy Waste of Electric Buses. 99-108. 10.1007/978-3-662-45079-6 8.
14. Emission Standards: Europe: Cars and Light Trucks, url:
https://dieselnet.com/standards/eu/ld.php
15. Dhital, Narayan Babu, et al. ”Effects of driving behavior on real-world emissions
of particulate matter, gaseous pollutants and particle-bound PAHs for diesel trucks.”
Environmental Pollution 286 (2021): 117292.
16. Shirneshan, Alireza. ”HC, CO, CO2 and NOx emission evaluation of a diesel engine
fueled with waste frying oil methyl ester.” Procedia-Social and Behavioral Sciences
75 (2013): 292-297.