www.ietdl.

org
Published in IET Intelligent Transport Systems
Received on 3rd April 2013
Revised on 26th September 2013
Accepted on 26th October 2013
doi: 10.1049/iet-its.2013.0076

ISSN 1751-956X

Fuel consumption and gas emissions of an automatic

transmission vehicle following simple eco-driving
instructions on urban roads
Grégoire S. Larue, Husnain Malik, Andry Rakotonirainy, Sébastien Demmel
Centre for Accident Research and Road Safety – Queensland (CARRS-Q), Queensland University of Technology (QUT),
Kelvin Grove, Queensland, Australia
E-mail: [email protected]

Abstract: Following eco-driving instructions can reduce fuel consumption between 5 and 20% on urban roads with manual cars.
The majority of Australian cars have an automatic transmission gear box. It is therefore of interest to verify whether current eco-
driving instructions are efficient for such vehicles. In this pilot study, participants (N = 3) drove an instrumented vehicle (Toyota
Camry 2007) with an automatic transmission. Fuel consumption of the participants was compared before and after they received
simple eco-driving instructions. Participants drove the same vehicle on the same urban route under similar traffic conditions. The
authors found that participants drove at similar speeds during their baseline and eco-friendly drives, and reduced the level of their
accelerations and decelerations during eco-driving. Fuel consumption decreased for the complete drive by 7%, but not on the
motorway and inclined sections of the study. Gas emissions were estimated with the (Virginia Tech) VT-micro model, and
emissions of the studied pollutants (carbon dioxide (CO2), carbon monoxide, oxides of nitrogen and hydrocarbons) were
reduced, but no difference was observed for CO2 on the motorway and inclined sections. The difference for the complete lap
is 3% for CO2. The authors have found evidence showing that simple eco-driving instructions are efficient in the case of
automatic transmission in an urban environment, but towards the lowest values of the spectrum of fuel consumption reduction
from the different eco-driving studies.

1 Introduction there is no international standard or benchmark on the

optimal method to instruct and perform eco-driving.
Transport accounts for nearly 27% of total CO2 emissions Eco-driving generally attempts to change drivers’ behaviour
from fossil fuel combustion. Transport is the second largest through instructions or tips such as driving more smoothly
CO2 emitting sector after electricity production [1]. Fuel by anticipating changes in the traffic, shifting gear sooner,
combustion emits water and CO2 exhaust gases (perfect operating the vehicle within an optimum revolution range,
combustion), carbon monoxide (CO), oxides of nitrogen avoiding jerky braking/acceleration and avoiding traffic
(NOX) and unburned hydrocarbons (HCs). These emissions jams [9, 10]. European Union regulations already require
are pollutants whose emissions are controlled by eco-driving to be taught during training of novice drivers.
government standards [2]. Vehicle emissions and energy Japan achieved its 2010 goal of reducing CO2 emissions by
consumption are largely influenced by driving style, 31 million tons below 2001 levels by encouraging drivers to
particularly in terms of acceleration, choice of speed and use their vehicles more efficiently through eco-driving [1].
idling time [3, 4]. Many countries have therefore promoted Traffic has increased in urban transport systems and driving
eco-driving as a key element of national strategies to reduce pattern over a specified route varies to a great extent (e.g. road
fuel consumption and CO2 emissions. grade) [11]. To effectively reduce gas emissions, it is necessary
Eco-driving consists of a variety of driving techniques to change driver behaviour or style in a way such that
(style) that save fuel and lower emissions. Eco-driving eco-driving becomes the norm rather than the exception [12].
instructions in Europe, US and Japan have been shown to Research has shown that vehicle’s emissions can be
reduce fuel consumption between 5 and 20% on urban accurately estimated with models taking into account the
roads for combinations of manual and automatic speed and acceleration of the vehicle, such as the
transmission vehicles [5–7]. Eco-driving instructions are a comprehensive modal emissions model and the VT-micro
cost-effective solution which can be implemented rapidly model [13]. The VT-micro model has been shown to be
on a large scale, improving energy consumption and above 90% accuracy [14]. Such models allow to assess the
emissions without significant increase of travel time [8]. effects of driving style and hence can be used to evaluate
Eco-driving instructions vary considerably per country. the effects of eco-driving on emissions.
Apart from Sweden, which has introduced a new law in The claimed advantages of the eco-driving approach are
2007 to test eco-driving skill during a practical vehicle test, that they can apply to vehicles of any age or size, they can

590 IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597
& The Institution of Engineering and Technology 2014 doi: 10.1049/iet-its.2013.0076
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org

Fig. 2 Instrumented vehicle

that almost 50% of fuel consumption in urban environments

is spent on accelerating and decelerating, as illustrated in
Fig. 1 [16].
In France, St-Pierre and Andrieu [17] conducted a study
which comprehensively characterised the immediate effects
of eco-driving instructions on driver behaviour in a vehicle
Fig. 1 Smart Drive study has established that almost 50% of fuel with manual transmission on semi-rural roads. Our study
consumption in urban environments is spent on accelerating and builds on their findings and focuses on the effects of
decelerating eco-driving instructions on driver’s steady speed,
acceleration and deceleration, in a vehicle with an automatic
transmission and on urban roads. We use the simple advice
take effect across the entire fleet of vehicles immediately at for attaining eco-driving behaviours, as Andrieu and
low cost (as opposed to being phased in), and that they can St-Pierre [9] showed that even though a complete
result in immediate savings to individuals from greater fuel eco-driving course provides higher benefits, differences are
efficiency, better safety and perhaps lower insurance rates [12]. small. Our experiment will test the following two hypotheses:
One of the most popular eco-driving instructions for drivers
operating vehicles with a manual gear box is to shift to the † simple eco-driving instructions improve fuel consumption
highest gear as soon as possible without lugging the engine. of automatic passenger vehicle in an urban environment.
Such instructions reduce engine revolutions per minute † simple eco-driving has positive effects on gas emissions
(RPM) and consequently the fuel consumption. However, when driving an automatic car in an urban environment.
such instructions are very general and do not take into
account driving conditions that could require different This study comprises a within subject comparison of
strategies. For instance, are such instructions relevant for driving performance with and without eco-driving
urban driving, motorway and inclined road sections? As instructions using a medium size instrumented passenger
most passenger vehicles in Australia are equipped with an vehicle featuring an automatic gear box travelling in an
automatic transmission (also called automatic gear box), are urban environment.
such instructions adapted to automatic transmission
vehicles? Drivers of automatic transmission vehicles are not
aware of the gear they are driving in nor are they able to 2 Method
control gear shifts. Japan has conducted many trials to
evaluate the effect of eco-driving instructions on fuel 2.1 Experimental design
consumption in urban road. Shinpo’s [15] study has shown
positive eco-driving outcomes with a fleet of automatic and 2.1.1 Instrumented vehicle (Fig. 2): A medium size
manual transmission vehicles, but without focusing on the Toyota Camry Altise (2007) with automatic gear box was
transmission type. The Smart Drive study has established fitted with:

Fig. 3 Architecture of the in-vehicle system

IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597 591
doi: 10.1049/iet-its.2013.0076 & The Institution of Engineering and Technology 2014
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org

Fig. 4 Test track specification

† IBEO laser scanner for leading vehicle detection. All the above-mentioned sensory data were recorded in a
† Vigil system [global positioning system (GPS) receiver, system using RTMaps software. The architecture of the
accelerometer and cameras] for vehicle position, vehicle in-vehicle system is illustrated in Fig. 3.
dynamics and to record the images of the road ahead and
images inside the vehicle. 2.1.2 Test track specification: The driving experiment
† NeoOBD on-board diagnostic device (OBD-II) for was conducted in an urban environment at Kelvin Grove,
retrieving vehicle’s RPM, speed, instantaneous fuel Brisbane, Australia. The test track features a small portion
consumption and the percentage of throttle open. of two lane motorway (3 km) and suburban single lane

592 IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597
& The Institution of Engineering and Technology 2014 doi: 10.1049/iet-its.2013.0076
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org
roads (2 km). Driving such a small distance allows us to stopped at the positions identified as ‘A’ and ‘C’ in Fig. 4.
control for driver’s fatigue and habituation. Fig. 4 illustrates Each participant started from point ‘A’ as shown in Fig. 4
the test track used for the experiment. Fig. 4b is a GPS and completed the lap (anticlockwise) in which the
point tracking of one of the participants. This image is also motorway section (speed limit 80 km/h) started 800 m from
used to highlight the track’s length. The track’s length will the starting point A. The motorway ended at 2200 m from
be used to explain the beginning/end of the different road the starting point. The section of road between 2500 and
types (i.e. motorway, incline etc.) in the procedure section. 3500 m, from A, was a moderately inclined road (speed
The map (GoogleMap) of the track is presented in Fig. 4a. limit 60 km/h). The ascending angle of inclination was
Specifications for the test track/circuit are given below: measured to be 4°. The rest of the roads were classified as
suburban roads.
† a fixed start point and end point from where the test vehicle
starts and stops, respectively. 2.4 Data collection
† clearly visible marked lanes (two lanes). The length of the
track was approximately 5000 m for encompassing different The data related to the vehicle dynamics and obstacle
classifications of the road (motorway, suburbs, incline road etc.). positioning were recorded with the in-vehicle technology
described in Fig. 3. Data were collected from the sensors of
the instrumented vehicle and the surrounding environment at
2.1.3 Traffic conditions: The experiment was designed varying frequencies. Data from the GPS were retrieved at 1
for an urban environment without heavy traffic in which Hz, vehicle dynamics (i.e. RPM, speed and throttle position)
reaching the optimal speed of 80 km/h for fuel consumption at 50 Hz, cameras at 30 Hz and laser scanner at 50 Hz. The
was unlikely. The time and location of the experiment were parameters that were successfully analysed from controller
chosen to maximise free flow opportunities in most of the area network (CAN)-bus included RPM, speed, the
route. percentage of throttle open and the mass air flow rate. We
calculated the instantaneous fuel consumption using speed
2.2 Participant recruitment and air flow (throttle position) from CAN-bus readout. Lee
et al. [18] demonstrated positive relationships between
Participants were recruited via word of mouth within the engine RPM, throttle position and actual fuel consumption.
Queensland University of Technology. To be eligible, The distance travelled per unit of fuel miles per gallon
participants were required to have a driver’s licence issued (MPG) can be obtained with the following formula (from
within Australia, be younger than 60, have no medical circuit cellar), which includes conversion form MPG to km/l
conditions that affect their driving, be familiar with an
automatic transmission, have driven a medium passenger MPG = (14.7 × 6.17 × 4.54 × VSS × 0.621371)
car similar to the test vehicle and be an experienced driver

/ 3600 × MAF/100
(more than 5 years driving licence).
Participants for this study were 13 licensed car drivers,
where 14.7 = grams of air to 1 g of gasoline (ideal air/fuel
9 males and 4 females, aged between 25 and 60 years.
ratio); 6.17 = pounds/gallon density of gasoline; 4.54 =
g/pound (conversion); VSS = vehicle speed in km/h;
2.3 Procedure 0.621371 = miles/h/km/h (conversion); 3600 = s/h
(conversion); MAF = mass air flow rate in 100 g/s; and 100 = c
On their arrival, each driver was briefed about the itinerary,
MAF in g/s (conversion).
track geometry and the driving manoeuvres they would
have to perform. Participants drove on a 5 km circuit
around Kelvin Grove, which encompassed different road 2.5 Data analysis
types (motorway and suburban roads) and significant
The characteristics of driving performance such as speed,
variations in elevation. The suburban roads feature traffic
RPM, throttle, fuel consumption with and without
lights and stop signs. Each participant drove three laps of
eco-driving instructions were compared within participants
the circuit, with a total duration of approximately 30 min/
by Wilcoxon tests using the R statistical software version
session. A research officer sat on the back seat to operate
2.11.1. Power analysis shows that 13 participants are
the equipments and to assist the participants if questions
required with such (one-tailed) tests to detect large size
were asked. The first lap was a familiarisation drive. The
effects with a power β of 0.95 at α = 0.05. We extracted and
second and third laps were randomly driven with or without
analysed the speed when there is no vehicle 30 m in front
any eco-driving instructions to eliminate a learning effect of
of the test vehicle (free flow). The detection of vehicles in
the journey (counterbalanced). The eco-driving instructions
front was made with the IBEO laser scanner and confirmed
were given just before the corresponding lap. The
manually by visualisation of the front camera footages.
instructions were the following:

† accelerate and brake smoothly. 2.5.1 Gears: We were not able to retrieve the gear level
† slow down and watch speed. from the OBD-II. However, the gear selected is related to
† anticipate the road ahead and avoid unnecessary abrupt the speed and RPM. In vehicles with internal combustion
braking. engines, the output conversion between the engine and the
† choose the appropriate speed. drive wheels is achieved by the combined action of the
† monitor RPM and avoid excessive RPM. assemblies of the power-train [19]. The power-train is
characterised by a total power-train ratio iA between
Each participant’s driving task was divided into three road engine-speed RPM and the road wheel speed u as follows
sections corresponding to the road characteristic (i.e.
motorway, incline and suburban). Every lap started and ·
u = iA RPM

IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597 593
doi: 10.1049/iet-its.2013.0076 & The Institution of Engineering and Technology 2014
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org
The resulting velocity-/engine-speed diagram is therefore
linear for each gear, with a different slope for each gear
level. We were able to obtain the gear in which the vehicle
was by recording data with the Toyota Camry at steady
pace for various speeds. The ratio RPM/u is characteristic
of the gear, and the clustering was done by two different
researchers, who analysed the RPM graphs and RPM/u
ratios to identify the gears of the vehicle.

2.5.2 Gas emissions: The VT-micro emissions model

was developed as a statistical model consisting of linear,
quadratic and cubic combinations of speed and acceleration
levels using chassis dynamo-meter data collected at the Oak
Ridge National Laboratory and the Environmental
Protection Agency [13]. The VT-micro models expresses
the emissions of a gas Emissiongas as a function ‘of the
speed u and the acceleration a of the vehicle as follows
⎧ 3 3
⎨ i=0 j=0 Lgas i j
i, j u a
e , if a ≥ 0
Emissiongas (u, a) = 3 3 gas i j
⎩ i=0 j=0 Mi, j u a
e , if a , 0

where Lgas gas

i, j and Mi, j are regression coefficients calibrated for
the given gas and a given type of vehicle for positive and
negative accelerations, respectively. Such coefficients have
been evaluated for the following gases: CO2, CO, HC and
NOX [20]. We used the matrix L gas and M gas that were
calibrated in that paper for the (light duty vehicles) LDV3
Fig. 5 Mean speed before and after the eco-instructions for the
category of vehicles, which is the category our Toyota
complete lap, on the incline (4°) and on the motorway
Camry belongs to. This category was obtained by data
mining techniques classification and regression tree and is
defined as light duty vehicles built after 1995, with an observed (2.23 against 2.33 kph/s). Decelerations decreased
engine size lower than 3.2 l and a mileage <83 653 km. for the complete drive after eco-driving instructions, from
This model is used in conjunction with boundary conditions 2.42 to 2.13 kph/s (P < 0.001). For deceleration, no
in terms of speed and acceleration, in order to ensure differences for deceleration were observed on the motorway
realistic gas emissions estimates [14]. and inclined sections. The average RPM (presented in
Fig. 6) was significantly higher after eco-driving instruction
3 Results for the motorway (P = 0.017) with an increment from 1730
to 1913 RPM. No statically significant difference was
The synchronisation software RTMaps was able to observed for the incline section (P = 0.30). Analysis of the
successfully record, synchronise and replay data from the positive kinetic energy shows a reduction from 0.519 to
sensors illustrated in Fig. 3. Participants completed their 0.468 m/s2 after eco-driving instructions (P < 0.001),
drive on average in 435 s during the baseline. No statistical showing a better anticipation of the traffic ahead. This
difference was observed for the eco-driving lap, with an analysis shows that participants have followed the posted
average of 438 s. The average following distance clearance, speed limit and reduced both their accelerations and
as measured by the IBEO laser device during the decelerations, anticipating traffic ahead, following partly the
experiment was approximately 28 m. Therefore participants eco-driving instructions they received. They tried to
were generally able to drive in free flow traffic. improve the smoothness of their driving, while not reducing
First, we ensured that participants did follow the their travel time, and failed to reduce their RPM.
eco-driving instructions. We analysed first speed on both The instantaneous fuel consumption variation (pattern)
sections and for the overall lap. All participants complied to with or without eco-driving instruction was similar
the posted speed limit. The mean speed before and after the within participants. For example, Fig. 7 illustrates the
eco-instructions for the complete lap, on the incline (4°) instantaneous consumption (in ml/s) for a couple of
and on the motorway were very similar, as illustrated in participants as obtained from CAN-bus data. Red line
Fig. 5. Wilcoxon tests show that no statistical difference corresponds to the consumption during the baseline drive,
was observed for speed before and after instructions whereas the green line represents consumption with
(P-value of 0.58, 0.78 and 0.60 for the complete lap, the eco-driving instructions. Although some participants
motorway and incline sections, respectively). Then, we showed improvements after eco-driving instructions as for
assessed the level of accelerations and decelerations. the left image of Fig. 7, some did not show any
Accelerations for the complete lap decrease from 2.95 to improvements of even consumed more fuel (right of Fig. 7).
2.64 kph/s after eco-driving instructions (P = 0.001). The fuel consumption for the complete lap, on the
Accelerations on the highway section were on average of motorway and inclined sections is illustrated in Fig. 8.
3.48 and 2.93 kph/s for the baseline and eco-driving, Although fuel consumption was slightly smaller on the
respectively. This difference was statistically significant motorway after eco-driving instructions (7.26 l/100 km
(P < 0.001). On the inclined section, no difference was instead of 8.37 l/100 km), Wilcoxon tests did not show any

594 IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597
& The Institution of Engineering and Technology 2014 doi: 10.1049/iet-its.2013.0076
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org

Fig. 6 Average RPM

Fig. 8 Fuel consumption for the complete lap, on the motorway

statistically significant decrement in fuel consumption
and inclined sections
(P = 0.38). On the incline section, average fuel consumption
varied from 13.83 l/100 km before instructions to 11.35 l/
100 km. This decrease of 17.9% of fuel consumption nevertheless failed to reach statistical significance (P = 0.11).
For the complete drive, fuel consumption decreased from
11.47 to 10.64 l/100 km (P = 0.048), which corresponds to
a 7.2% decrement during eco-driving.
Gears were identified by the following conditions on the
RPM/u ratio:

† first gear is defined by a ratio between 45 and 53.

† second gear: 33–42.
† third gear: 26–30.
† fourth gear: 19–23.
† other values of the ratio are considered as transitions
between gears.

The proportion of time spent in the four different gears of

the vehicle was obtained and are presented in Fig. 9 and
Table 1 for the different driving sections and conditions.
The baseline showed that on the motorway section,
participants spent most of the time in the fourth gear (60%).
Following eco-driving instructions resulted in a 21.3%
reduction in time spent in the fourth gear (P = 0.027). For
the incline section, the baseline showed that the second gear
was the most used and that the third gear was used 17% of
the time. After eco-driving instructions, the time spent in
third gear decreased to 4% (P = 0.076) and the time spent
in second gear increased by 15% (P = 0.024). For the
complete lap, the time spent in the fourth gear decreased
from 12.9 to 8.4% (P = 0.026). No other differences
observed were statistically significant.
Fig. 7 Instantaneous consumption (in ml/s) for a couple of Gas emissions were estimated with the VT-micro model
participants as obtained from CAN-bus data from the speed and acceleration of the vehicle at each time.

IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597 595
doi: 10.1049/iet-its.2013.0076 & The Institution of Engineering and Technology 2014
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org

Fig. 9 Proportion of time spent in the four different gears

Table 1 Time spent in the different gears (percentage)

Complete lap Motorway section 4° incline section

Baseline Eco-driving Δ* Baseline Eco-driving Δ* Baseline Eco-driving Δ*

first 14 16 — 6 7 — 35 30 —
second 41 44 — 8 12 — 49 63 +14% (0.024)
third 32 32 — 25 42 — 16 6 —
fourth 13 8 −38% (0.027) 60 39 −21% (0.027) 0 1 —

*Statistically significant differences (P-value)

Table 2 Gas emissions (in g)

Complete lap Motorway section 4° incline section

Baseline Eco-driving Δ* Baseline Eco-driving Δ* Baseline Eco-driving Δ*

CO2 969.80 941.34 −2.9% (0.033) 246.73 227.77 — 34.94 33.37 —

CO 2.264 2.040 −9.9% (0.001) 0.822 0.638 −22.4% (<0.001) 0.070 0.067 —
HC 0.096 0.089 −7.3% (0.001) 0.033 0.027 −18.2% (0.002) 0.0031 0.0029 —
NOX 0.381 0.367 −3.7% (0.033) 0.145 0.136 — 0.011 0.009 —

*Statistically significant differences (P-value)

Table 2 presents the total emissions of CO2, CO, HC and NOX 0.033) from 969.8 to 941.3 g, CO emissions decreased by
gases for the complete drive and the motorway and inclined 9.9% (P = 0.001) from 2.3 to 2.0 g; HC gas emissions
sections. No differences were observed for the inclined decreased from 96 to 89 mg (P = 0.001), which corresponds
section. For the motorway, CO2 and NOX emissions to a 7.3% decrement. Finally, NOX emissions were reduced
remained similar, whereas emissions of other gases were by 3.7% (P = 0.033), from 381 to 367 mg.
reduced. Emissions of CO gas were reduced by 22.5% from
0.822 to 0.638 g (P < 0.001); HC gas emissions were 4 Limitations
reduced by 18.2% from 33 to 27 mg (P = 0.002). Although
the fuel consumption did not decrease after eco-driving As a pilot study, this experiment has several limitations.
instructions on the motorway section, the decrement of CO Firstly, we had a small sample size (N = 13) and sample
and HC gas emission tends to show that the combustion type which might have impacts on its statistical validity.
was more efficient during eco-driving (i.e. the combustion The gas emissions were obtained using a model that was
was more complete). Unfortunately, no direct benefit was not calibrated on our vehicle. Therefore the values obtained
observed for CO2 and NOX emissions. For the complete can only be used as a basic approximation of the effects on
lap, gas emissions were reduced for the four gases under gas emissions of eco-driving instructions for automatic cars.
investigation. CO2 emissions were reduced by 2.9% (P = In addition, the psychological profile of the drivers was not

596 IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597
& The Institution of Engineering and Technology 2014 doi: 10.1049/iet-its.2013.0076
 17519578, 2014, 7, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-its.2013.0076 by Universidade Federal Da Paraiba, Wiley Online Library on [02/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.ietdl.org
controlled for (e.g. sensation seekers). Furthermore, a direct 6 Acknowledgments
comparison between manual and automatic vehicles would
have provided a more powerful evaluation of the effects of We would like to thank Queensland Cyber Infrastructure
eco-driving with automatic cars. Despite these limitations, Foundation (QCIF) and QFleet for their support in this
this pilot study showed statistically significant change in the project. We also would like to thank Professor Rakha and
RPM, gears and gas emission variables when participants Dr. Ahn from the Virginia Tech Transportation Institute for
received eco-driving instructions. providing us with the VT-micro model parameters.

7 References
5 Conclusions
1 International Transport Forum: ‘The cost and efficiency of reducing
Both hypotheses described in the introduction were transport ghg emissions – preliminary findings’, 2010
confirmed. Eco-driving instructions for manual cars is 2 Ribbens, W.B.: ‘The basics of electronic engine control’
expected to improve fuel consumption by 5–20% [5], we (Butterworth-Heinemann, Oxford, 2013), pp. 177–231, Ch. 5
3 Nie, Y., Li, Q.: ‘An eco-routing model considering microscopic vehicle
have shown that eco-driving instructions using a vehicle operating conditions’. 91st Annual Meeting of Transportation Research
with an automatic gear box in an urban environment Board, 2013
results in reduced fuel consumption of a magnitude close 4 Kamal, M.A.S., Mukai, M., Murata, J., Kawabe, T.: ‘Ecological vehicle
to the lower limit of what is observed with manual cars. control on roads with up-down slopes’, IEEE Trans. Intell. Transp. Syst.,
2011, 12, (3), pp. 783–794
The significant increase of RPM during eco-driving might 5 EcoMove: ‘Cooperative mobility systems and services for energy
have offset the theoretical fuel consumption gain from efficiency’, 2010
smoother accelerations. Simple eco-driving instructions 6 Ando, R., Nishihori, Y.: ‘How does driving behavior change when
failed to prove benefits on the inclined section of this following an eco-driving car?’, Procedia – Soc. Behav. Sci., 2011, 20,
study, suggesting that more comprehensive instructions pp. 577–587
7 Vagg, C., Brace, C.J., Hari, D., Akehurst, S., Poxon, J., Ash, L.:
should be provided to drivers for this particular task, ‘Development and field trial of a driver assistance system to
particularly focusing on releasing the accelerator before encourage eco-driving in light commercial vehicle fleets’, IEEE Trans.
passing crests and using momentum for the following hill Intell. Transp. Syst. , 2013, PP, (99), pp. 1–10
by accelerating before reaching the bottom of the previous 8 Nouveliere, L., Mammar, S., Luu, H.T.: ‘Energy saving and safe driving
assistance system for light vehicles: experimentation and analysis’.
hill. Further analysis showed that the increase in engine Networking, Sensing and Control (ICNSC), 2012 9th IEEE Int. Conf.,
RPM in eco-driving mode was a consequence of driving 2012, pp. 346–351
in lower gears for a much longer time, and a consequence 9 Ando, R., Nishihori, Y.: ‘A study on factors affecting the effective
of the smoother accelerations. These results suggest eco-driving’, Procedia – Soc. Behav. Sci., 2012, 54, pp. 27–36
refining the gear changing points in automatic vehicles 10 Ben Dhaou, I.: ‘Fuel estimation model for eco-driving and eco-routing’.
2011 IEEE Intelligent Vehicles Symp. (IV), 2011, pp. 37–42
might be required for improving eco-driving efficiency. 11 Saboohi, Y., Farzaneh, H.: ‘Model for developing an eco-driving
Further research is needed to assess whether such issue is strategy of a passenger vehicle based on the least fuel consumption’,
widely found for a wide range of automatic vehicles, and Appl. Energy, 2009, 86, (10), pp. 1925–1932
to assess the potential improvements of eco-driving 12 Barkenbus, J.N.: ‘Eco-driving: an overlooked climate change initiative’,
benefits with fine-tuning of gear shifting points. Benefits Energy Policy, 2010, 38, (2), pp. 762–769
13 Rakha, H., Kamalanathsharma, R.K.: ‘Eco-driving at signalized
of eco-driving instructions in terms of gas emissions were intersections using v2i communication’. 2011 14th Int. IEEE Conf.
also proven to be lower than expected. For the complete Intelligent Transportation Systems (ITSC), 2011, pp. 341–346
lap, all pollutants studied were reduced, but only by 3% 14 Ahn, K., Rakha, H., Trani, A., Van Aerde, M.: ‘Estimating vehicle fuel
for CO2. On the motorway section, we observed reduced consumption and emissions based on instantaneous speed and
acceleration levels’, J. Transp. Eng., 2002, 128, (2), pp. 182–190
emissions of approximately 20% for CO and HC, but not 15 Shinpo, K.: ‘Japanese eco-driving initiative’. Workshop on Ecodriving:
for CO2 and NOX. This is probably because of the Successful Approach to Ecodriving, IEQ-HQ Paris, 22 November 2007
reduced average acceleration during eco-driving. Future 16 Smart Drive: Energy conservation Center, Japan, 2006
work should include detailed analyses of the 17 St-Pierre, G., Andrieu, C.: ‘Characterisation de l’eco-conduite et
environmental effects, and should also examine the effects construction d’un indicateur dynamique pour vehicules thermiques’,
2010
of the eco-driving instructions (technology or education) 18 Lee, M.L., Park, Y.K., Jung, K.K., Yoo, J.J.: ‘Estimation of fuel
on safety indices (e.g. following distances). Our analysis consumption using in-vehicle parameters’, Int. J. u- and e- Serv. Sci.
on the impacts of different urban road types on Technol., 2011, 4, (4), pp. 37–46
eco-driving behaviour will assist in designing 19 Harald, N., Aaron, K.: ‘Automotive transmissions: fundamentals,
selection, design and application’ (Springer Verlag, DE, 2010)
comprehensive eco-driving strategies for urban contexts. 20 Rakha, H., Ahn, K., Trani, A.: ‘Development of vt-micro model for
These eco-driving strategies will have the ability to adapt estimating hot stabilized light duty vehicle and truck emissions’,
to the changing environmental scenarios. Transp. Res. D, Transp. Environ., 2004, 9, (1), pp. 49–74

IET Intell. Transp. Syst., 2014, Vol. 8, Iss. 7, pp. 590–597 597
doi: 10.1049/iet-its.2013.0076 & The Institution of Engineering and Technology 2014