Faster, Cheaper, Cleaner

Assessing Urban Mobility

Gonc¸alo Salazar

, Jo˜ao Pedro Silva

and Bernardo Ribeiro

CEiiA - Centro de Excelˆencia e Inovac¸˜ao na Ind´ustria Autom´ovel, Rua Engenheiro Frederico Ulrich, 2650, Maia, Portugal

IN+ Center for Innovation, Technology and Policy Research, Instituto Superior T´ecnico,

Av. Rovisco Pais, 1049-001 Lisboa, Portugal

Keywords:

Urban Mobility, Portugal, Emissions, Cost, Time, Mobility Index.

Abstract:

This work aims at systematically characterizing urban mobility. There are several possible approaches to the

problem but this work focus on reducing commuting times and emissions using the current infrastructures and

without increasing the cost to the end user and doing so in an effective and sustainable way. A set of pairs of

destinations and origins was chosen to represent usual commutes and cover a signiﬁcant area of a city. These

locations were connected by car, motorcycle, public transportation, cycling and walking. The routes were

performed in both directions at three different times of day which allowed for a better understanding of daily

trafﬁc variations. For each route and for each mode of transportation the route and time taken were collected

and the emissions and cost were estimated . The data was treated to adjust the route and gradient difference

between the collected data and actual roads. Measurements and estimates were compared and averaged for

all the routes and for each means of transportation providing an overall view of the commutes. The different

means of transportation were compared and the limit to which one mode has the advantage over another was

evaluated, this advantage is however dependent on the chosen route.

1 INTRODUCTION

The current fast growth of urban population is likely

to lead to an unsustainable situation if not handled

properly. This increase creates multiple problems not

seen before, being mobility one of the most affected

areas.

New emerging urban mobility problems cannot be

tackled effectively using traditional solutions. Inno-

vative approaches and creative solutions are needed

to tackle this ever increasing problem.

This paper aims at analysing one such approach

and present some preliminary results and ﬁndings.

The focus was mainly on the analysis of soft and

shared means of transport and how could they be ben-

eﬁcial to the population in three different areas: time,

cost and emissions.

2 MOTIVATION

Historically, the solution to urban mobility problems

are based in the construction of more roads, bigger

infrastructures and the displacement of industries and

services to unused areas close to the cities. However

with the fast growth of the urban population (World

Health Organisation, 2010) and the increasing cost of

land these solutions are no longer efﬁcient.

However, to tackle this new problem, one must

ﬁrst have the means to analyse and quantify it. While

several methodologies exist (Tomtom, 2013; Gabi-

nete de Estudos e Planeamento and Direcc¸˜ao Munic-

ipal da Via P´ublica, 2007) to evaluate mobility in ur-

ban areas they do not consider nor optimize for an

intermodal route calculation with different weights

of cost, emissions, distance travelled and commuting

time.

The proposal for this project was to try to develop

an index that included considerations for time, emis-

sions and cost for a certain route travelled in urban

areas. The methodology could then be used to evalu-

ate and compare urban mobility in different areas of

the city and different cities.

3 METHODOLOGY

To ensure the coherence of the results and their ap-

Salazar G., Silva J. and Ribeiro B..

Faster, Cheaper, Cleaner - Assessing Urban Mobility.

DOI: 10.5220/0005446900430048

In Proceedings of the 4th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS-2015), pages 43-48

ISBN: 978-989-758-105-2

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

plicability to other cities and to different contexts,

a structured approach was deﬁned. These methods

could then be used to analyse and then compare dif-

ferent cities.

Due to geographical proximity and presence of

multiple modes of transport the city of Porto was cho-

sen as a testbed for the pilot. Several routes were se-

lected and analysed using the different available trans-

port modes.

The gathered data was analysed using available

web tools, some of the data was estimated using val-

ues from available sources.

The overall methodology (cf. Figure ??) consists

of a ﬁrst selection routes and gathering of data fol-

lowed by the deﬁnition of mobility models that would

allow to expand the analysis to other areas and there-

fore validate the models. The process would be con-

tinuous and cyclic to improve the model as well as

evaluate the mobility within the city.

Figure 1: Methodology overview.

3.1 Modes of Transport

The project considered multiple means of transport

within the city of Porto. The modes of transport

should represent the most commonly used and allow

for different scenarios to be evaluated.

The evaluated modes of transport were:

• Car;

• Motorcycle;

• Public Transport (Rail and Road based);

• Bicycle;

• Pedestrian.

Using these different modes of transport allows

for a greater understanding of the mobility in the city

and considers most of the available options to the cit-

izens.

3.2 Route Selection

In order to have a pilot for the methodology different

pairs of origins and destinations within the urban area

of the city of Porto were selected. The different pairs

should be connected by public transport and have

access by the other means of transport.

The criterion for selection of routes was as fol-

lows:

• The routes should cover the biggest area possible;

• The routes should have connections by public

transport;

• The routes should connect the most travelled from

and to areas of the city.

After an analysis with the municipality of Porto,

the selected routes were as follows:

• From the city hall to CEiiA’s building in Matosin-

hos

– Connects the city center to an important adja-

cent city. Multiple interfaces exist between the

two cities. It is an important connection for the

two cities

• From Casa da M´usica to the Faculty of Engineer-

ing

– Connects an important multimodal interface

and entry point of the city with a widely used

area close to the outskirts of the city and to one

of the most important hospital of the city

• From Campanh˜a’s train station to the hotel area

near Avenida da Boavista

– Connects one of the major entry points to the

city with a area with good access near the city

center and good opportunities for visitors

• From Anemona’s Roundabout to Ribeira

– Connects two areas on the outside of the city

bypassing the center. Usual route for tourists

and citizens going to the riverside area

Figure 2: City of Porto with routes connecting origins and

destinations highlighted.

Using the modes of transport deﬁned in chapter

3.1 and the selected pairs of origins and destinations,

SMARTGREENS2015-4thInternationalConferenceonSmartCitiesandGreenICTSystems

a total of 320 km within 20 unique routes were iden-

tiﬁed (cf. Figure 2). Figure 2 shows that the almost

totality of the city area is covered providing a testbed

extensive enough for the purpose of this work.

3.3 Data Collection

To evaluate the parameters considered for this paper

GPS data for all routes was collected.

3.3.1 GPS Data

The GPS data was collected using GPS Logger for

Android, a smartphone application (cf. Figure 3).

This application records points along the route in

CSV (Comma Separated Values) format.

The recorded GPS data is extensive, however,

within the scope of the project only some information

was used. Table 1 describes the data collected with

the GPS application.

Table 1: Data collected from the GPS Logger.

Data type Units

Time Hours,Minutes,Seconds

Latitude Degrees

Longitude Degrees

Altitude m

Speed m/s

Figure 3: Screenshot from the GPSLogger Application.

The data collected will allow to deﬁne the exact

route used to connect the origin to the destination and

the time taken to travel such route.

3.4 Estimates

Some of the parameters could not be directly mea-

sured and were therefore estimated or used from other

sources.

The use of estimated values for part of the results

is needed due to the constraints of the problem and the

high complexity required to measure some values.

This is acceptable because the project focuses not

on a particular type of displacement but tries to aver-

age and create an approximate model for cost, time

and emissions. However to ensure the validity of the

results the number of sources was kept to a minimum.

3.4.1 Costs

Within the scope of this paper the cost estimation con-

sidered those that directly affect the user and were cal-

culated as a function of distance. Public transport was

the exception as the costs for the user are not a func-

tion of distance but depend on the number of travels

necessary to take. This would give an approximate

value for each mean of transport that could be used

for comparison along their route.

For car the cost per kilometre was obtained from

the Portuguese legislation (Portugal, 2012) since it

deﬁnes the value that should be paid by companies to

their employees when using the private car for work

related travels. It was considered that during the mak-

ing of the law an extensive study was done so as to

deﬁne a fair value that included not only gas, but also

vehicle maintenance, wear, insurance and taxes. The

value should reﬂect the average cost per kilometre of

the global Portuguese car ﬂeet (Balsa, 2013).

For motorcycle the value was also obtained from

the Portuguese Legislation (Portugal, 2010) due to the

reasons presented above.

For public transport the cost is not a function of

distance but instead depends on the number of trips.

Both rail and road based transport in the city of Porto

have the same cost and use the same tickets. One

single-use trip in the city of Porto has a cost of 1.20

euros. However, considering a daily commute with

two trips (home-work and back) for 22 days of work

per month and a public transport card costing 30.10

euros per month would result in a cost of 0.64 euros

per trip.

Cycling has a cost of 0.03 euros per kilometre

considering maintenance for a lifetime of 10 years

according to the working paper on the Costs and

Beneﬁts of Cycling (Belter et al., 2013).

The cost of walking is negligible and considered

to be zero. Although one could argue shoes and

clothes are deteriorated while walking, their cost

cannot be attributed to the activity itself as people

Faster,Cheaper,Cleaner-AssessingUrbanMobility

have to buy clothing and shoes anyway. Recent

reports (Fundac¸˜ao Calouste Gulbenkian, 2014) have

shown there are even potential savings in terms of

healthcare costs, as walking regularly does reduce

the risk of cardiovascular diseases and diabetes, for

instance. However, it is difﬁcult to quantify direct

savings and for the purpose of this study the authors

decided not to consider them.

Overall the obtained estimated cost values for

each mode of transport were:

Table 2: Cost estimation per mode of transport.

Transport Cost

Car 0.36AC per km

Motorcycle 0.14AC per km

Public Transport 0.64AC per trip

Bicycle 0.03AC per km

Pedestrian 0AC per km

3.4.2 Emissions

For this project emissions were not measured due to

the granularity of data required and the project times-

pan. To ensure that the model represented the actual

Portuguese car ﬂeet (Balsa, 2013) extensive measure-

ments had to be made and those would produce results

highly dependent on the driving style (Panis et al.,

2006).

More comprehensive models like the COPERT

and MOVES were not used due to some problems,

the MOVES model does not represent the Portuguese

car ﬂeet (Duarte and Costa, 2010) while the COPERT

model does not have a data set representing Portugal

individually, only as part of the EU (Emisia, 2014).

Furthermore emission estimation can introduce a

great overhead if not properly evaluated, therefore,

within the scope of the work only the equivalent CO2

emissions for each mode of transport were consid-

ered.

Car emissions were obtained from the Carbon

Footprint Calculator (Carbon Footprint, 2012) for a

EU average petrol car with an engine up to 1.4L. The

value obtained was of 160.61 grams of CO

per km.

It is considered that this value reﬂects the average car

from the Portuguese ﬂeet and therefore is adequate.

Motorcycle emissions were obtained from the

same source as the car ones (Carbon Footprint, 2012)

and the value obtained is 106.21 grams of CO

per

km.

Public transport emissions need to consider both

road based and rail based public transport. A previ-

ous work done by Quercus (Portuguese Environmen-

tal Agency), with results available online (Ecocasa,

2012). For road based public transport the value is

82 grams of CO

per km while for Portos rail based

transport system is 63 grams of CO

per km.

The European Cyclist Federation (European Cy-

clist Federation, 2012) states that cycling emissions

are around 21 grams of CO

per km.

For this paper it was considered that walking pro-

duced no emissions.

Overall the obtained estimated emission values for

each mode of transport were:

Table 3: Emission estimation per mode of transport.

Transport Emissions

Car 160.61 gCO

per km

Motorcycle 106.21 gCO

per km

Road-based Public Transport 82 gCO

per km

Rail-based Public Transport 63 gCO

per km

Bicycle 21 gCO

per km

Pedestrian 0 gCO

per km

4 DATA ANALYSIS

Using the collected and estimated values for each

mode of transport and pair of source and destination

the data was analysed to perceive patterns that would

allow to generate a model that could be extended to

other routes with different characteristics. The col-

lected GPS data was converted from CVS to GPX

format to be easier to interpret with mapping appli-

cations.

Due to errors induced by the measurements the

data points were ﬁtted to the road to more correctly

represent reality. For each pair of source and destina-

tion the time taken and estimated cost and emissions

were evaluated for each mode of transportation. The

results were then averaged for all routes to give an

overall vision of the city along those 3 vectors.

The obtained results were extended to a model

where the cost, emission and time can by evaluated

for each mode of transport over distance (cf Figures

4,5, and 6)

Using the results obtained and the analysis along

each individual vector, an index that comprehends the

three parameters was developed. This index considers

different weights for time, emissions and cost that can

be tuned for different values (cf. Equation 1).

Index = Average(Time × Weight,

Distance× Emission× Weight,

Distance× Cost× Weight)

(1)

SMARTGREENS2015-4thInternationalConferenceonSmartCitiesandGreenICTSystems

Figure 4: Time over distance.

Figure 5: Cost over distance.

Figure 6: Emissions over distance.

With this index value for each distance the results

can be extrapolated and values for all distances can be

obtained.

A urban commuter such as a businessman will

make the mode of transport choice with most of the

weight on time. A possible weight distribution for

such user would be: 98% for time, 1% for emissions

and 1% for costs. With this distribution, car and mo-

torcycle are the best options up to 10 km. After that,

cycling is the best option, due to the small accumulat-

ing effects of cost and emissions in the index, never-

theless these results are based on trend lines and can

be change once further data gathering is done. After

Figure 7: Triple Parameter Mobility Index versus distance

(lower is better) for 98% time, 1% emissions and 1% costs.

Figure 8: Triple Parameter Mobility Index versus distance

(lower is better) for 85% time, 1% emissions and 14% costs.

20 km motorcycle becomes the best option. Public

transport is one of the worst options for the evaluated

distances.

A typical middle class worker will consider that

time is of importance but cost will also be an impor-

tant factor therefore the weight distribution will be:

85% for time, 1% for emissions and 14% for costs. In

this case, cycling is always the best option while the

car is the worst. Cycling is almost two times better

than a car for the evaluated distances. Public trans-

port is, again, worse than cycling. Walking, although

worst than cycling is still better than driving up to 20

km, without considering other external factors, as fa-

tigue for instance.

An urban commuter with great environmental

conscience would consider emissions to be the top

priority. A possible weight distribution for such com-

muter would be: 25% for time, 50% for emissions and

25% for costs. In this situation, bicycle and walking

are always the best option. Due to the high weight for

emissions all other modes of transport are worst. Pub-

lic transport is again the least desirable mode of trans-

port. However, when considering a zero emission ve-

hicle, such as an electric vehicle, the car is again the

Faster,Cheaper,Cleaner-AssessingUrbanMobility

Figure 9: Triple Parameter Mobility Index versus distance

(lower is better) for 25% time, 5% emissions and 70% costs.

best option, since not only the emissions will be re-

duced but the costs as well.

5 CONCLUSIONS

After evaluating the obtained results it can be seen

that cycling, contrary, to what was expected is faster

than public transport while being cheaper and less

pollutant.

Car is one of the fastest modes of transport, how-

ever is the most pollutant and expensive. However

this only considers one occupant per car, a higher

number of occupants per vehicle could make car a

better option. This opens the possibility for further

studies considering the impact of urban car-pooling

schemes in commutes.

The index shows that without a considerable

weight for time, car is never the best option. Cost

has a great inﬂuence on the index. Public transport is

rarely the best option, it is however a valid mode of

transport if one cannot walk or cycle over long dis-

tances.

Public transport might have a bad performance

due to the lack of dedicated lanes for road based pub-

lic transport. While rail-based public transport is not

affected by trafﬁc and has a better performance than

cycling the overall results of public transport are neg-

atively affected by the poor results of the road based

counterpart. In further studies, rail and road based

public transport modes should be analysed separately.

It might be possible that electric assist bicycles

could further improve the advantages of cycling over

public transport. However further studies need to be

made.

While the selection of the city of Porto as a case

study provides an insight on the methodology to eval-

uate the urban mobility, however further studies in

different cities are recommended as a validation tool

for the methodology.

REFERENCES

Balsa, J. (2013). Avaliac¸˜ao do impacto da introduc¸˜ao de

ve´ıculos el´etricos na procura de combust´ıveis em por-

tugal. Master’s thesis, Universidade de Coimbra.

Belter, T., Harten, M., and Sorof, S. (2013). Working paper

on the costs and beneﬁts of cycling. SustraMM.

Carbon Footprint (2012). Carbon footprint cal-

culator. Available from: http://calculator.

carbonfootprint.com/calculator.aspx. Last accessed:

04/12/2014.

Duarte, I. and Costa, G. (2010). An´alise do modelo MOVES

como ferramenta de c´alculo de emiss˜oes. Master’s

thesis, Universidade de Aveiro.

Ecocasa (2012). Simulador emiss˜oes. Available

from: http://www.ecocasa.pt/simuladores/Simulador

EmissoesCO2/simuladorEmissoesCO2.php. Last ac-

cessed: 04/12/2014.

Emisia (2014). Copert. Available from: http://

www.emisia.com/copert-shop. Last accessed:

06/01/2015.

European Cyclist Federation (2012). Cycle more often 2

cool down the planet. European Cyclist Federation.

Fundac¸˜ao Calouste Gulbenkian (2014). Um futuro para a

sa´ude. Fundac¸˜ao Calouste Gulbenkian.

Gabinete de Estudos e Planeamento and Direcc¸˜ao Munici-

pal da Via P´ublica (2007). Mobilidade na cidade do

porto - an´alise das deslocac¸˜oes em transporte individ-

ual.

Haan, P. and Keller, M. (2000). Emission factors for passen-

ger cars: application of instantaneous emission mod-

eling. Atmospheric Environment, 34.

Panis, L., Broekx, S., and Liu, R. (2006). Modelling in-

stantaneous trafﬁc emission and the inﬂuence of trafﬁc

speed limits. Science of the Total Environment, 371.

Portugal (2010). Decreto-Lei n. 137/2010 de 28 de Dezem-

bro. Lisbon, DGCI.

Portugal (2012). Circular da DGCI n19/93 de 20 de Agosto.

Lisbon, Di´ario da R´epublica.

Tomtom (2013). TomTom European Trafﬁc Index.

World Health Organisation (2010). Bulletin of the world

health organisation.

SMARTGREENS2015-4thInternationalConferenceonSmartCitiesandGreenICTSystems