Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
1
TITLE: MEASURING THE EFFECTS OF TRAFFIC CONGESTION ON FUEL
CONSUMPTION
SUBMISSION DATE: 11 November 2014.
WORD COUNT: 3,998 (without references) + (7 figures + 5 tables) x 250 = 6,998
Alvaro Garcia-Castro, Corresponding Author
Transport Research Centre (TRANSyT), Escuela de Ingenieros de Caminos, Canales y Puertos,
Universidad Politécnica de Madrid, Ciudad Universitaria, 28040 Madrid, Spain
Tel: +34 91-336-5234 Fax: +34 91-336-5362; Email: [email protected]
Andres Monzon
Transport- Civil Eng. Department, Escuela de Ingenieros de Caminos, Canales y Puertos,
Universidad Politécnica de Madrid, Ciudad Universitaria, 28040 Madrid, Spain
Tel: +34 91-336-5373 Fax: +34 91-336-5362; Email: [email protected]
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2
ABSTRACT
Nowadays, road transportation sector is responsible for a considerable amount of the total energy
consumption, also contributing to about 20-30% of the total man-made CO2 emissions
worldwide.
In metropolitan areas, the problem is compounded by the effect of traffic congestion.
Transport models, online route planners, and the latest versions of on-board navigators do take
into account traffic variability in their calculation of travel times. However, the effect of traffic
congestion on fuel consumption is hardly taken into consideration, thereby introducing a
considerable distortion in the result and consequently, the traveler decision process.
The aim of this paper is to investigate in detail, how traffic congestion affects fuel
consumption and to evaluate the importance of including traffic congestion in modeling and
driver information systems. This analysis is based on empirical data collected from some
itineraries of a metropolitan motorway and streets of the metropolitan area of Madrid, Spain.
Specifically, a total of 3,800 trips were recorded with three vehicles under different traffic
conditions and driving styles. The results show considerable increase in fuel consumption under
congested traffic conditions compared with free flow conditions, but also points to the
differences depending on the type of itinerary.
Keywords: Congestion, Fuel consumption, Route planner, Navigator, Energy, CO2
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
3
1. INTRODUCTION
The emission of greenhouse gases (GHG) and excessive consumption of energy resources is a
global problem, due to both, its causes and consequences (1). The transport sector is one of the
largest emitters despite the advances in the field of engines technology.
According to statistics provided by the European Environment Agency (2), road
transportation sector has begun to reduce their emission of GHG; but still contributes to about
93% of the emissions attributable to the transportation sector representing approximately 20.4%
of the total emission. In the U.S., the contribution percentage of road transportation to total GHG
emissions is even higher, reaching almost 22% (3). In terms of energy consumption, the
transportation sector accounted for 26% of the global energy consumption in 2010, and
transportation energy use is expected to increase by 1.1 percent every year from 2010 to 2040,
according to the International Energy Outlook 2013 Reference case (4).
It is worth mentioning that a great amount of fuel consumption and GHG and pollutants
emissions from road transportation sector are concentrated in metropolitan areas. In fact,
according to the European Environmental Agency, metropolitan areas account for 40% of the
total CO2 produced by the road transportation sector (5).
In this context, many efforts to reduce energy consumption focus on the road transport
sector. The European Commission (6) proposes an integrated policy to tackle the problem from
different directions, among them, worth to mention are: the demand management, the shift to
cleaner modes, improving vehicle technologies, and the use of information and communication
technologies (ICT).
2. ARE CONGESTION EFFECTS ON CONSUMPTION PROPERLY CONSIDERED?
Section 2 presents a review of different models and driver information systems with regard to
congestion treatment. It is divided into 3 parts. The first part discusses the most remarkable
features of route planners and navigators. The second part summarizes how congestion is treated
in transportation demand models, finishing with some remarks about emission models and the
disadvantages of using average speed models at micro scales.
2.1 Cost Calculations in Route Planners and Navigators
An online route planner provides information on optimal routes based on different parameters.
Usually, it suggests the shortest or fastest route based on complex algorithms developed by
Dijkstra (7), but many of them take also into account other factors, such as speed limits, safety
and even the current traffic situation.
For instance, in the case of Google Maps (8), it displays the optimal route and estimated
travel time, according to the traffic situation and congestion level. In its first version, this option
was based on historical data, while nowadays it computes utilizing data extracted from the
instantaneous position of smartphone users. Moreover, these applications also provide an
estimate of fuel consumption for the selected route, allowing the user to select from general
categories of vehicles or to define the average consumption. Mappquest (9) calculates the cost of
journey using the average price of gas along the route, length of trip, mileage of the vehicle, and
route characteristics. In other websites, the user can also customize route characteristics: shortest
journey time, shortest distance or the most fuel-efficient route (10). However, none of these online route planners take into consideration the traffic conditions to calculate fuel consumption.
Quite similar is the functioning of the GPS navigation systems, which are able to
recalculate and guide the user through a faster route in case of congestion. However, the fastest
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
4
route is not always the one with the lowest fuel consumption (11). Hence, in recent years, "ecorouting" algorithms have been developed to inform the user about the most economical and
lower emissions route (12), although very few methodologies integrate real time emissions and
traffic information (13).
2.2 Congestion Costs in Transportation Demand Models
Demand models in transportation use a generalized cost function while distributing trips between
different modes and alternative routes. These cost functions include both monetary and nonmonetary costs; basically travel time.
Generally, fuel consumption is added in the formulas of generalized cost as a function of
travel distance, which may be a good enough approximation if static traffic conditions are taken
into consideration. To estimate environmental externalities and test congestion pricing policies,
some studies also consider average speed as a factor to include environmental costs in the model
(14). With these algorithms, traffic variability is taken into account; but the use of emission
models based on the average speed can result in erroneous estimates depending on the type of
route, as discussed in the following section.
2.3 Traffic Congestion in Energy and Emission Models
The energy and emission models are usually developed from a large number of measurements
made, both in laboratories and field trips. These models usually emphasize the importance of the
fleet composition and, depending on the level of detail, these vehicles are aggregated into
categories with similar values of consumption and emission.
The fuel consumption and emission models can be classified according to their level of
disaggregation. They range from the very detailed ones based on the speed profiles of each
vehicle to the aggregate ones used for national emissions inventories.
After examining some emission models, Smit et al. (15) concluded that most of them do
explicitly consider the traffic situation in the modeling process. However, the main problem lies
in the use at a local scale of models designed for emission inventories at regional or national
level, as their improper use can lead to serious over- or underestimations.
Macro-level models are applied for the analysis of large networks. In this family of
models, the average speed models consider the average speed, traffic demand, and fleet
composition of each itinerary to estimate emissions and fuel consumption. Examples of this type
of models are MOBILE (16) and COPERT (17).
Average speed is considered as a key factor in fuel consumption and CO2 emissions.
However, from a traffic engineering point of view, the average speed itself does not sufficiently
explain the variability of traffic conditions and its implications (11). For example, an average
speed of 40 km/h in an arterial road with traffic light coordination could reflect free-flow
conditions, achieving an almost constant speed. The same average speed on a motorway section
can represent high congestion levels, which has clear implications on fuel consumption and
emissions. Figure 1 shows the speed profiles of two trips made with the same vehicle and with a
similar average speed in two different itineraries. However, the differences in traffic conditions
lead to an increase in fuel consumption of about 40% for the congested section.
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
5
FIGURE 1 Urban and ring motorway itineraries with the same average speed but with
different fuel consumption.
3. OBJECTIVES AND METHODOLOGY
The literature review shows that the effects of congestion on fuel consumption are treated very
differently depending on the objectives of the model. In the very detailed micro-level models, the
influence of congestion is treated directly while at macro level, only a few models enables
analysts to incorporate local driving patterns which reflect congestion (18).
Under these premises, the aim of this paper is to investigate in detail how traffic
congestion affects the consumption and to evaluate the importance of considering traffic
congestion in modeling processes and in driver information systems. This analysis was
performed on different roads within the metropolitan area of Madrid with the objective to
analyze the differences between urban motorways and urban street itineraries.
As detailed below, the methodology is based on the analysis of fuel consumption and
speed profiles collected by floating vehicles which performed a number of trips in the city of
Madrid.
3.1 Data Collection Campaign
The data collection campaign was developed under the framework of the European research
project "ICT-Emissions", which aims to develop a methodology to simulate in detail the effects
on GHG emissions of a number of ICT measures applied to road transportation. As an added
value of this project, it is also intended to validate the methodology with case studies conducted
in the partner cities of Rome, Turin, and Madrid. In each of these cases, data collection
campaigns (induction loops, cameras, floating cars, etc.) were conducted with the aim of
evaluating the evolution of traffic and other vehicle parameters before and after the
implementation of the ICT measures.
In March and April 2013, the data collection campaign took place in coordination with
the Department of Traffic Technologies of Madrid. The main objective was to obtain speed
profiles and fuel consumption data at various itineraries of the M30 ring motorway (itineraries 1,
2, 3, and 4) and some adjacent urban streets (itineraries 5, 6, and 7) (Figure 2). Urban itineraries
do not include arterials, but only local streets and signalized collectors.
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
6
FIGURE 2 Monitored itineraries in the data collection campaign in Madrid.
This data collection campaign consisted of conducting a representative number of trips
through the selected itineraries using floating cars (FCD). With this procedure, it is possible to
obtain the fuel consumption and speed profiles. Three FIAT passenger cars were involved in this
study, a Punto 1.2L gasoline and two diesel variants, Punto and Bravo, with 1.3L and 1.6L
engines, respectively. The number of drivers involved was 9 and the total number of recorded
trips was approximately 3,800.
The vehicles were equipped with a mobile device with GPS connection to collect data on
distance travelled, instantaneous position, and speed at a frequency of 1 Hz, thus enabling to
record speed and acceleration profiles. The average fuel consumption (in L/100km) was recorded
directly from the on-board display at the end of each itinerary.
3.2 Review and Selection of Traffic Congestion Indicators
Due to the availability of data and the nature of demand models, the most commonly used
indicator has been the volume-to-capacity ratio, although this indicator has a number of
shortcomings when analyzing traffic congestion in detail (19). The main issue is that drivers, in
general, ignore this kind of information, but rather give importance to travel times.
Based on the literature review (20-23) and trends in software and application developers,
the following congestion indicators are highlighted:
• Delay: Difference between current travel time and free flow travel time
• Delay rate: Difference between current travel time and free flow travel time per kilometer
• Congestion Index (CI): Ratio between travel time and free flow travel time
• Speed reduction congestion index (SRCI): Mean speed reduction from free-flow to
current conditions over free flow mean speed. Index normalized to a 0 to 10 scale.
• Tom Tom congestion index: Difference between current travel time and free flow travel
time over free-flow travel time.
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
7
•
Level of Service (LOS) is a qualitative measure of the quality of driving conditions. On a
scale of A to F, it describes a range of flow conditions where A represents free-flow
conditions; and F severe congestion.
This study will analyze the traffic congestion using the delay rate. On one hand, from the
point of view of drivers, delay is the most understandable and interesting indicator, as it reflects
the time lost in congestion. The use of a length ratio makes possible the comparison between
sections with similar characteristics but different length.
3.3 Data Processing
As mentioned before, the data collection campaign was performed in the framework of the “ICTEmissions” research project, which includes testing some traffic management measures, such as
variable speed limits and section speed control. For the purpose of this study, the trips on which
these measures were in practice were not considered, as they could distort the fuel consumption
and travel time results.
Once filtered and rejected some trips due to GPS reception failures, the remaining were
grouped by vehicle and itinerary. The total number is 2,412, distributed as specified in Table 1
and following the itinerary ID number set in Figure 2.
TABLE 1 Number of Analyzed Trips According to Itinerary ID Number and Vehicle
Itinerary number
and vehicle
M30 ring Motorway itineraries
Urban itineraries
1
2
3
4
5
6
7
Total per
vehicle
Length (km)
5.8
5.3
5.4
6.7
1.2
1.3
1.6
na
Fiat Bravo 1.6D
72
124
46
44
124
86
82
578
Fiat Punto 1.3D
104
215
72
74
214
146
133
958
Fiat Punto 1.2G
Total per
itinerary
103
172
70
63
191
140
137
876
279
511
188
181
529
372
352
2,412
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
The delay rate is calculated using the following formula for each trip:
𝐷𝐷 =
(𝑡−𝑡0 )
60
𝑙
(1)
where,
t = current travel time (s)
t0 = free flow travel time (s)
l = length of itinerary (km)
The fastest trip of each itinerary was set as the free flow travel time of that itinerary. For
these reason, specific peak-off measurements were performed with low traffic intensities,
although overnight hours were not considered. The selected reference trips where investigated in
detail to confirm the absence of any reception failure or anomalous driving behavior.
Measurements of ring motorway trips have been grouped, according to the levels of
service (LOS) defined by the AAHSTO (24). Since most of the registered trips are close to the
area of flow instabilities (when congestion is about to start) and to achieve more accuracy,
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
8
service levels A and B have been divided equally into three sublevels, C and D into two, while E
and F are kept as a unique level of service.
4. RESULTS. FUEL CONSUMPTION AS A FUNCTION OF DELAY RATE
After calculating the delay and consumption values following the methodology proposed in the
previous section, the trips have been filtered by itinerary and vehicle, obtaining the relation
between the delay rate and fuel consumption, both graphically and analyzing best fitting curve.
The following paragraphs discuss some of these relations.
4.1 Fuel Consumption and Delay Rate of Urban Motorway Itineraries
In the case of itineraries of M30 Ring Motorway, the graphs show a growing trend of fuel
consumption due to the increase in the level of congestion. As an example, Figure 3 shows
graphically the relation between fuel consumption and delay rate for the same vehicle on two
different itineraries; values grouped by level of service.
FIGURE 3 Relation between fuel consumption and delay rate - Example of two itineraries
in M30 Ring Motorway.
Thus, for every vehicle and itinerary, the regression can be obtained that best fit the
consumption values as a function of the delay rates. Table 2 shows the slopes of the regression
lines.
TABLE 2 Slope and Goodness-of-Fit of the Regression Line for each Motorway Itinerary
and Vehicle
M30 ring Motorway itineraries
Itinerary number and
vehicle
Fiat Bravo 1.6D
Fiat Punto 1.3D
Slope
R
2
Slope
1
2
3
4
0.8388
0.9405
1.2212
1.3697
0.71
0.91
0.17
0.90
0.5549
0.1888
-0.8125
0.8100
Garcia-Castro and Monzon
Fiat Punto 1.2G
1
2
3
4
5
6
7
8
9
10
11
0.19
0.41
0.90
Slope
0.7761
0.9035
-3.0272
0.6454
0.63
0.70
0.51
0.56
R
2
M30 ring Motorway itineraries
Itinerary number and
vehicle
Slope
Fiat Bravo 1.6D
R2
1
2
3
4
-1.3797
1.4155
1.2212
0.207
0.40
0.22
0.17
0.01
Slope
-2.0475
-3.8528
-0.8125
-0.7977
0.93
0.88
0.41
0.29
-6.076
-3.8255
-3.0272
-4.1423
0.95
0.95
0.51
0.67
R
2
Slope
R
2
Consequently, figure 4 shows the graphs of the same itineraries presented in Figure 3, but
considering only LOS A and LOS B. We can observe high values of R2 when the slope is
negative, but very low in case of positive or nearly horizontal regression lines.
Adjusting to a second degree polynomial, we obtain the following values of goodness-of
2
fit (R ):
TABLE 4: Goodness-of-Fit of the second degree polynomial in Service Levels and
Sublevels A and B, for each Itinerary and Vehicle
Itinerary number
and vehicle
Fiat Bravo 1.6D
Fiat Punto 1.3D
Fiat Punto 1.2G
24
0.74
TABLE 3: Slope and Goodness-of-Fit of the Regression Line in Service Levels and
Sublevels A and B, for each Itinerary and Vehicle
Fiat Punto 1.2G
21
22
23
R2
We would expect all the slopes to be positive but, surprisingly, itinerary 3 presents
declining fuel consumptions as the delay rate increases. This can be explained by the fact that
high levels of congestion were not experienced in this itinerary, so that increase in delay does not
involve braking and acceleration phenomena. The delay in this case, is due to mere slight
reduction in the current average speed compared to free flow conditions. When only the service
levels close to the free flow (LOS A and LOS B) were analyzed, the regression lines present
mostly negative slopes, as reflected in Table 3.
Fiat Punto 1.3D
12
13
14
15
16
17
18
19
20
9
M30 ring Motorway itineraries
1
2
3
4
R
2
0.94
0.92
0.90
0.52
R
2
0.93
0.95
0.92
0.65
R
2
0.99
0.98
0.99
0.96
As expected, second degree polynomial achieves a better goodness-of-fit, obtaining this way a
curve with a minimum for delay rate values between 0.05 and 0.10. Examples are shown in
Figure 4.
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
10
FIGURE 4 Fitting curve of fuel consumption for delay rates close to free flow conditions
(LOS A and LOS B).
Analysis of Fuel Consumption with Reference To the Homogeneity of the Speed Profile
In order to investigate the reason for the occurrence of minimum in consumption, it is necessary
to analyze an indicator that reflects the speed homogeneity along an itinerary. Thus, GarciaCastro and Monzon (25) define Positive Accumulated Acceleration (PAA) as an indicator
obtained from the analysis of the speed profile in an itinerary which graphically represents the
area under the positive accelerations curve. Thus, a trip in which the speed varies considerably
tends to have a comparatively much greater PAA value than those of other trips on the same
itinerary but with uniform speed.
FIGURE 5 Example of Positive Accumulated Acceleration and fuel consumption as a
function of delay, for an itinerary of M30 urban motorway.
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Analyzing where the minimum of fuel consumption and PAA curves are situated in the
plot, it is possible to set 3 different areas as a function of the delay rate (Figure 5).
For delay rates close to zero, fuel consumption decreases because the average speed also
drops while maintaining a homogenous speed along the itinerary.
In the second zone, increase in traffic intensities lead to instabilities producing increasing
braking and acceleration, consequently decreasing the average speed. When the minimum is
reached, the average speed reduction is not enough to counter the effect of acceleration-braking
process on fuel consumption.
Finally, in the third stage, fuel consumption increases since both the negative effects of
increasing accumulated acceleration and the fact that at those high speeds the vehicle no longer
lies in the area of maximum efficiency.
4.2 Fuel Consumption and Delay Rate in Urban itineraries
For urban itineraries, it was not necessary to group the values based on LOS, as the slopes of the
regression lines show consistent and linear tendencies, and the trips recorded are well distributed
among the whole range of traffic conditions. Table 5 shows the slope of regression lines for
every urban itinerary and vehicle. It is remarkable that the vehicle with a higher engine
displacement presents consistently a larger slope for each of the urban itineraries studied.
Graphically, this fact can be also observed in the examples shown in Figure 6.
TABLE 5 Slope and Goodness-of-Fit of the Regression Line for each Urban Itinerary and
Vehicle
Urban itineraries
Itinerary number and
vehicle
Fiat Bravo 1.6D
Fiat Punto 1.3D
Fiat Punto 1.2G
24
11
5
5
7
1,514
1,0353
1,3394
0.34
0.40
0.63
Slope
1,3506
0,6922
1,0964
R2
0.38
0.12
0.37
Slope
1,4401
0,9041
1,1746
0.50
0.44
0.62
Slope
R
R
2
2
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
12
FIGURE 6 Fuel consumption as a function of the delay rate in urban itinerary 5.
In contrast to the motorway itineraries analyzed in the last section, it is not possible to
observe a decrease in consumption with increasing delay rates. This is in line with the fact that
for the range of speeds in urban itineraries, a drop of the average speed on an itinerary always
involves an increase in fuel consumption following the performance curves of vehicles’ engines
(26). In urban areas, both consumption and cumulative positive acceleration (PAA) behave
linearly, as opposed to what was observed in the case of urban motorways (Figure 7).
FIGURE 7 Regression lines of fuel consumption and positive accumulated acceleration as
a function of the delay rate in urban itinerary 5.
5. CONCLUSIONS
Garcia-Castro and Monzon
13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
It has been observed that the congestion effects on fuel consumption are not taken into
consideration in some route planners and transport demand models. However, the results of the
case study presented in this paper show a clear relationship between congestion and fuel
consumption. In the case of urban motorways as the case of M30 in Madrid, the data collected
show that small increases in traffic starting from free flow conditions are positive in terms of fuel
economy, but the general trend is that congestion increases fuel consumption.
Thus, for each itinerary of motorway it is possible to achieve optimum fuel consumption
as a function of the delay rate. Once this optimum is exceeded, increasing traffic intensities cause
speed profiles to become less homogeneous, and hence increases consumption.
On the other hand, the data analysis in urban areas presents a different behavior. For the
speed range analyzed in these areas, fuel consumption increases linearly as a function of
congestion.
The analysis of the speed profile homogeneity reveals that the fuel consumption largely
depends on the point at which increasing traffic intensities cause flow instabilities. Hence, the
optimum is reached at a different point depending on the type of road.
Therefore, the need to include a congestion parameter in fuel consumption models is
obvious, since the increase due to the level of traffic can exceed 100%. Moreover, the use of a
single indicator (usually average speed) involves some risk, since the optimal fuel consumption
depends also on the accumulated acceleration in the section, i.e., speed homogeneity. Similar
average speeds do not mean similar fuel consumptions for different types of roads.
For greater accuracy in the calculation of fuel consumption without resorting to
microscopic models which require very detailed data, it would therefore be advisable to adjust
the average speed models by adding a factor of congestion or traffic homogeneity based on the
type of road. Heavy duty vehicles should also be object of further research, as they are major
contributions to CO2 emissions, especially in interurban itineraries.
27
28
29
30
31
32
33
34
35
ACKNOWLEDGMENT
This work was supported in part by the European Commission under the ICT-Emissions project,
"Development of a methodology and tools for assessing the impact of ICT measures in road
transport emissions”. The authors also acknowledge the collaboration of the City of Madrid and
Madrid Calle-30.
36
37
38
39
40
41
42
43
44
REFERENCES
1. Ramanathan, V., and Y. Feng. Air Pollution, Greenhouse Gases and Climate Change:
Global and Regional Perspectives. Atmospheric Environment, Vol. 43, No. 1, 2009, pp.
37-50.
2. European Environment Agency. Annual European Union Greenhouse Gas Inventory
1990–2011 and Inventory Report 2013. No 8/2013, European Environment Agency,
Copenhagen, 2013.
3. The United States Environmental Protection Agency. Inventory of U.S. Greenhouse Gas
Emissions and Sinks: 1990 - 2011. EPA 430-R-13-001, U.S. Environmental Protection
Agency, Washington D.C., 2013
4. Energy Information Administration. International Energy Outlook 2006. DOE/EIA0484(2013), U.S. Department of Energy, Washington D.C., 2013.
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
14
5. European Environment Agency. The European Environment. State and Outlook 2010.
Urban Environment. Publications Office of the European Union. ISBN 978-92-9213-1517, Copenhagen, 2010.
6. European Commission. Mobilising Information and Communication Technologies to
Facilitate the Transition to an energy-efficient, low-carbon economy. COM (2009) 111,
European Commission, Brussels, 2009.
7. Dijkstra, E. W. A Note on Two Problems in Connexion with Graphs. Numerische
Mathematik, Vol. 1, No. 1, 1959, pp. 269-271.
8. Barth, D. The Bright Side of Sitting in Traffic: Crowdsourcing Road Congestion
Data. http://googleblog.blogspot.com.es/2009/08/bright-side-of-sitting-in-traffic.html,
Accessed July 17, 2014.
9. Mapquest. Mapquest Routeplanner. http://www.mapquest.com/routeplanner, Accessed
July 17, 2014.
10. Viamichelin. http://www.viamichelin.com/, Accessed July 17, 2014.
11. Ahn, K., and H. A. Rakha. Energy and Environmental Impacts of Route Choice
Decisions. In Transportation Research Board 86th Annual Meeting, 2007.
12. Minett, C. F., A. M. Salomons, W. Daamen, B. Van Arem, and S. Kuijpers. Eco-Routing:
Comparing the Fuel Consumption of Different Routes between an Origin and Destination
using Field Test Speed Profiles and Synthetic Speed Profiles. In Integrated and
Sustainable Transportation System (FISTS), 2011 IEEE Forum on, 2011, pp. 32-39.
13. Barth, M., K. Boriboonsomsin, and A. Vu. Environmentally-Friendly Navigation.
In Intelligent Transportation Systems Conference, 2007. ITSC 2007. IEEE, 2007, pp. 684689.
14. Benedek, C., and L. Rilett. Equitable Traffic Assignment with Environmental Cost
Functions. Journal of Transportation Engineering, Vol. 124, No. 1, 1998.
15. Smit, R., A. L. Brown, and Y. C. Chan. Do Air Pollution Emissions and Fuel
Consumption Models for Roadways Include the Effects of Congestion in the Roadway
Traffic Flow? Environmental Modelling & Software, Vol. 23, No. 10–11, 2008, pp. 12621270.
16. Parrish, D. D. Critical Evaluation of US on-Road Vehicle Emission
Inventories. Atmospheric Environment, Vol. 40, No. 13, 2006, pp. 2288-2300.
17. Ntziachristos, L., and Z. Samaras. Speed-Dependent Representative Emission Factors for
Catalyst Passenger Cars and Influencing Parameters. Atmospheric Environment, Vol. 34,
No. 27, 2000, pp. 4611-4619.
18. Liu, H., and M. Barth. Analysis of US EPA’s MOVES Model’s Operating Modes and
Handling of History Effects. In 90th Annual Meeting of the Transportation Research
Board, Washington, DC, 2011.
19. Boarnet, M. G., E. J. Kim, and E. Parkany. Measuring Traffic Congestion. Transportation
Research Record: Journal of the Transportation Research Board, No. 1634,
Garcia-Castro and Monzon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
15
Transportation Research Board of the National Academies, Washington, D.C., 1998, pp.
93-99.
20. Smit, R. An Examination of Congestion in Road Traffic Emission Models and their
Application to Urban Road Networks. Griffith University, Nathan, 2006.
21. Taylor, M. A. Network Modelling of the Traffic, Environmental and Energy Effects of
Lower Urban Speed Limits. Road and Transport Research, Vol. 9, No. 4, 2000, pp. 4956.
22. Cohn, N., E. Kools, and P. Mieth. The TomTom Congestion Index. In 19th ITS World
Congress, 2012.
23. Schrank, D., B. Eisele, and T. Lomax. TTI’s 2012 Urban Mobility Report. Texas A&M
Transportation Institute, 2012.
24. AASHTO. Policy on Geometric Design of Highways and Streets. American Association
of State Highway and Transportation Officials, Washington, D.C., 2001.
25. Garcia-Castro, A., and A. Monzon. Homogenization Effects of Variable Speed
Limits. Transport and Telecommunication Journal, Vol. 15, No. 2, 2014, pp. 130-143.
26. Barth, M., and K. Boriboonsomsin. Traffic Congestion and Greenhouse Gases. Access,
No. 35, 2009, pp. 2-9.