Smart Cities

©SHUTTERSTOCK/JOZEF BAGOTA

Smart mobility solutions

for a smart city
Alok Prakash

S
mart mobility is a key com-
ponent in smart city initia-
tives that are currently
being explored around
the world by authori-
ties, industr y players, and
academics alike. The term
smart mobility encompasses
many facets of mobility,

©SHUTTERSTOCK/METAMORWORKS
including improving public
transport services, provid-
ing guidance to commuters
and motorists, and real-
time traffic monitoring and
management, among others.
This article provides a
brief overview of various tech-
nologies that are being devel-
oped to enable smart mobility
in modern smart cities, with a
special emphasis on techniques for
improving public transport services.
It also touches briefly on other as-
pects of smart mobility, such as le-
veraging smart phones for informa-
tion dissemination and guidance to
travelers as well as low-cost sensors
for rea l-t ime t ra ffic monitor ing
and management. S i n gap or e h a s b e en push i n g What does it mean to be a smart city
for wa rd w it h its sma r t nat ion or nation? In other words, what does
Toward smart mobility initiatives since 2014, while India a city or country need to achieve to
Developing a smart city or even launched a smart cities mission in be deemed smart? Is it better trans-
making an entire nation “smart” 2015 with an aim to develop 100 portation services, enhanced govern-
has caught the imagination of cities around the country. ment services through digitization,
aut hor it ie s a nd go ve r n me nt s As is evident, introducing smart- more accessible health-care op-
around the world. For example, ness into our lives is the goal of big tions—or all of these? It is through
and small countries alike. Typical improvements in all of the areas
Digital Object Identifi er 10.1109/MPOT.2020.3023539
questions that are often asked in mentioned here, and more, that a city
Date of current version: 26 December 2020 this context include the following: or nation can truly be called smart.

24 ■ J a n u a r y/Febr uar y 2021 IEEE POTENTIALS 0278-6648/21©2021IEEE

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on October 12,2021 at 20:56:54 UTC from IEEE Xplore. Restrictions apply.
 This article primarily explores large industrial or residential cam- when needed. This smarter mode
the various smart mobility initia- pus; and, finally, to 4–7-seat taxis of operation comes with numerous
tives that are being developed for for more private, point-to-point benefits for both commuters and taxi
smart cities around the world. The transfers. Traditionally, the larger drivers. As a rider, the waiting time
term smart mobility encompasses vehicles are used for scheduled ser- for a taxi is not only more determin-
many facets, such as improving pub- vices with a fixed route and longer istic, but the fare is also known in
lic transport systems (PTSs), provid- travel time but cost significantly advance, thereby mitigating the risk
ing timely and relevant guidance to less than the smaller vehicles, such of getting overcharged or “taken for a
commuters and motorists and en- as taxis, which, in turn, provide ride.” Additionally, the mobile applica-
suring real-time traffic and road faster and dedicated transit, albeit tions (apps) typically used for booking
management. Each of these tasks at a higher cost. such services also allow customers
is a challenging problem in its own In recent years, all forms of road- to track the booked taxi, both before
right and needs to be tackled indi- based PTSs discussed have seen and during the ride. This provides
vidually. Several research, academic, significant changes in their mode of immense relief to passengers in new
industrial, and government estab- operation, largely enabled by the use places as well as to more vulnerable
lishments have made great contribu- of technology in one form or another. riders traveling by themselves.
tions in these areas to continuously The most prominent of these chang- At the same time, taxi drivers also
improve the system, as discussed in es has been in the way we use taxi benefit from this approach by driving
the subsequent sections. services. In the past, it was common to areas with more demand, thereby
for empty taxis to drive around a reducing the amount of time they
PTSs city in the hopes of picking up pas- drive without a passenger. The new
PTSs are designed to serve a large sengers by the roadside. To ensure system also provides taxi drivers an
number of commuters and are typical- sufficient availability of taxis, city opportunity to select the trips they
ly run using a two-pronged approach authorities mandated a minimum wish to make based on their prefer-
within a city: road and rail (Fig. 1). distance that a taxi must cover per ences. The drastic change in the taxi
day. For example, in Singapore, the operations has been made possible
Road-based PTSs Land Transport Authority (LTA) is- through the intelligent use of tech-
Road-based PTSs, as the name sug- sued an advisory in 2013 that a taxi nology, such as mobile apps, better
gests, use the road network to must cover a minimum distance of GPS-based tracking, and digital pay-
extend the reach of public transport 250 km every day to make it easier ment systems. Ride-hailing services
to commuters’ doorsteps. The types for commuters to hail one. have become so pervasive that a city
of vehicles used for road-based pub- However, with the advent of ride- cannot be called smart unless it of-
lic transport range from single or hailing services, such as Grab, Uber, fers this service to its residents.
double-decker buses t hat ca n and Ola, as a smarter alternative to The popularity and success of
accommodate 80–125 passengers; traditional taxi services, this require- ride hailing with cars has motivated
to 12–15-seat minivans for shuttling ment was scrapped in 2017, since companies like Grab and Uber to ex-
smaller numbers of people around a commuters can simply “call” a taxi plore and offer other shared modes

MAGLEV COURTESY OF WIKIMEDIA COMMONS/MARYLAND GOVPICS. HYPERLOOP

COURTESY OF WIKIMEDIA COMMONS/NEUHAUSENGROUP.

FIG1 The vehicles used for public transport.

IEEE POTENTIALS Januar y/Februar y 20 2 1 ■ 25
Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on October 12,2021 at 20:56:54 UTC from IEEE Xplore. Restrictions apply.
 of transport, such as bicycles or per- Additionally, many passengers fleet may achieve the intended goals
sonal mobility devices (PMDs), and were already familiar with book- of improving user experience and be-
others. With such services, users are ing taxis through a phone call, even ing economically viable.
allowed to borrow a mode of trans- before the advent of mobile phone Other innovations to improve
portation, such as a parked bicycle, app-based ride hailing. In contrast, road-based PTSs include the use of
PMD, from one parking lot, use it to public buses are meant to serve ev- dedicated lanes or roads for public
reach their destination, and leave ery citizen, their ability to call or use buses. Cities like Singapore have
the ride in another parking lot near a mobile phone notwithstanding. In hundreds of kilometers of dedicated
the destination. Users pay only for other words, it cannot be assumed bus lanes on major roads to expe-
the one-time use of the vehicle, while that every commuter can be shifted dite PTSs. These lanes are demar-
the operators bear the costs of own- toward a paradigm where public cated with yellow or red lines along
ership and maintenance. This mode buses arrive only when called. This with appropriate signage painted
of operation helps with the utility constraint makes it nearly impos- on the road surface. Since there is
maximization of the assets and is, sible to know the origin and destina- no physical barrier to separate the
as a result, economical. tion of every passenger waiting for a bus lane from other lanes, a large
On-demand services have also bus, without which it is not possible stretch of the dedicated lanes in
been tried with larger vehicles but to schedule an on-demand public Singapore can also be used by pri-
with limited success. For example, bus with flexible routing to serve the vate vehicles during off-peak hours
the Kutsuplus service in Helsinki, entire community. [Fig. 2(a)], while a small portion is
Finland, used an automated system Even with these limitations, com- meant to be used only by public bus-
to perform on-demand scheduling panies like Shuttl in India have had es at all times [Fig. 2(b)].
and flexible routing of 15 minibuses some level of success in offering an Studies have shown that the ef-
according to real-time bookings made on-demand bus service by leverag- fectiveness of such dedicated lanes
by passengers through their smart- ing a large enough pool of commut- in improving the travel time of pub-
phones. The system, which operated ers using their service to sustain lic buses hinges critically on the en-
for three years, enjoyed considerable their businesses. At the same time, forcement of such lanes to allow only
popularity among passengers. How- a similar service, Grabshuttle, by authorized vehicles. In other cities,
ever, the Kutsuplus project was sus- Grab in Singapore, has announced such as the upcoming Iskandar
pended at the end of 2015 due to a lack the closure of its offering, highlight- Malaysia Bus Rapid Transit system
of funds to support the expansion of ing the difficulties faced by on-de- in Johor Bahru, Malaysia, physical
the fleet to a large number of vehicles. mand services using larger vehicles. barriers are preferred to ensure the
A similar on-demand bus system was It has also been suggested that, usage by authorized vehicles only.
proposed by Tokyo University and although it may not be economically While dedicated bus lanes, when
tested through field trials in three sustainable to use such on-demand sufficiently enforced, have been
cities in Japan. A study of the afore- services to provide anywhere-to- shown to improve public bus servic-
mentioned projects suggests that if anywhere transport, it may be pos- es, critics of this system also argue
a demand-responsive transport sys- sible to offer such services for first-/ that, in a modern urban city with
tem is intended to serve passengers last-mile transit using smaller vehi- limited land resources, dedicating
spread over an entire city with highly cles within a small defined operating an entire lane to only public services
diverse demands in terms of origins zone. Examples of such zones may may not be feasible and often leads to
and destinations, a large fleet size is be a housing estate or an academic overall traffic congestion due to lim-
necessary to make the system both campus, with constraints placed on ited road availability for other road
reliable to commuters and economi- the number of possible origins and users. This problem is particularly
cally viable for the operator. destinations, where a much smaller severe during peak hours on major
roads, where many public and private
vehicles need to compete for limited
road resources, thereby resulting in
a degraded user experience for many.
Recently, researchers from TUM-
CREATE proposed a virtual right-
of-way concept, where bus lanes are
“dedicated” only when a bus is ap-
proaching. At all other times, even
private vehicles are allowed to use
(a) (b) the lane. The management and en-
fo r c e m e nt h a s b e e n p r op o s e d
FIG2 (a) A peak-hour bus lane, Singapore. (Source: Google Street View.) (b) A full-day through the large-scale deployment
bus lane, Singapore. (Source: Google Street View.) of low-cost camera-based sensors.

26 ■ J a n u a r y/Febr uar y 2021 IEEE POTENTIALS

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on October 12,2021 at 20:56:54 UTC from IEEE Xplore. Restrictions apply.
 The proposed system is currently ferred mode of transport within mod- el in the carriages of arriving trains.
being developed and will be trialed ern cities as a rapid transit service for The information was meant to nudge
in the near future on public roads to a large number of urban commuters. the passengers toward carriages that
demonstrate its effectiveness as the The London Tube, New York Subway, are less crowded, making it easier
best of both worlds. Mass Rapid Transit Singapore, and and faster to board. Such systems
The use of autonomous vehicles Mass Transit Railway Hong Kong are have also been employed on the Ya-
(AVs) in road-based public transport just a few examples of such urban, manote Line in Tokyo, Japan, where
is yet another exciting idea that is rail-based, public rapid transit sys- commuters can use mobile apps to
currently being explored by many tems from around the globe. These view the locations of trains as well as
academic and commercial enter- systems have been immensely suc- their loading statuses in real time.
prises around the world. The typical cessful in satisfying large numbers of
fixed routes of public transport ve- passengers, albeit with a high initial Guidance for commuters
hicles greatly benefit AVs by limiting installation cost. and motorists
the variations in their paths to make New and improved technolo- A discussion on smart mobility
navigation easier. The vast majority gies have made trains run faster efforts cannot be complete without
of research for AVs is focused primar- than ever before. High-speed trains considering the impact of smart
ily on making such vehicles more in- around the world, such as the bullet phones in our daily commuting or
telligent to accurately perceive the trains in Japan, regularly travel at driving behavior. While operators
environment around them by using speeds of 300 km/h and more. Even and authorities are continuously
an array of sophisticated sensors. so, designers have been making con- working toward improving the reli-
However, it has been shown that stant inroads to further increase the ability and speed of road- and rail-
the success of AVs hinges critically speed of trains with technologies based PTSs, it is also important to
on the accompanying infrastruc- such as magnetic levitation and pro- accurately, and in real time, “sense”
ture to provide timely and relevant pulsion. The proposed next-gener- current road/traffic conditions, and
information for the safe and smooth ation bullet train in Japan will rely the current location and crowded-
operations of the AVs. For example, on magnetic levitation technology ness of PTSs, among others. Such
camera-based sensors can be de- and has already set several speed information, when disseminated in a
ployed along the entire road network records during trial runs, reaching timely manner, allows for efficient
to provide lanewise, real-time infor- more than 600 km/h in April 2015. journey planning and even dynamic
mation on traffic density, accidents/ Recently, the concept of using dedi- repla nning for commuters a nd
incidents on the road, road work, cated paths, without relying on rail motorists. To this end, in current
and so on to enable AVs to plan a tracks, for high-speed transport ser- and future smart cities, suitable
safe and efficient route. vices is also being explored in the form infrastructure needs to be installed
Similarly, the sensors deployed of a hyperloop, a concept described in to provide such sensing capabilities.
at road intersections and bends can detail by Elon Musk in his 2013 white Scores of developers have also
assist AVs in negotiating curves and paper. The primary idea of the pro- worked with authorities to provide
maneuvering around corners by pro- posed system is to use air-tight tubes relevant information, wherever avail-
viding information on other vehicles through which pods can travel with- able, in a seamless manner through
and pedestrians that may otherwise out encountering air resistance, there- smartphone apps to users. Prime ex-
be hidden from them. Several re- by making transportation extremely amples of such apps include Google
searchers are focusing on developing fast as well as energy efficient. Since Maps, Waze, and dozens of city-spe-
a cost-effective sensor infrastructure the introduction of the hyperloop idea, cific apps that estimate the times of
for citywide deployment by propos- several start-ups and even large com- bus or train arrivals (Fig. 3). Most of
ing low-complexity algorithms for re- panies have been exploring the com- these apps can be categorized into
al-time sensing that can be mapped mercial viability of such systems. two types: those for commuters and
on low-cost embedded platforms In an effort to leverage modern those for motorists or drivers.
while still being resilient to adverse technology to further improve user
weather and lighting conditions. experience, operators and research- Apps for commuters
ers are exploring the demand-respon- Several state and private developers
Rail-based PTSs sive scheduling of trains to reduce the have created mobile apps, for both
Unlike the road-based PTSs dis- waiting time of commuters. Another iOS and Android, to help commut-
cussed in the previous section, rail- area of interest is to devise techniques ers in their daily travels. For exam-
ways have the distinct advantage of to spread passengers along the plat- ple, Google Maps allows passengers
having a dedicated track for their forms to improve the boarding pro- to plan their route using multimod-
operations. Traditionally, railways cess at subway stations. For example, al transportation options, including
have been used to move very large train operators in Singapore trialed taxis, buses, light rail, and sub-
numbers of travelers from one town a passenger information system in ways. The app provides alternative
to another. Railways are also the pre- 2018 that displayed the crowding lev- routes while taking into account

IEEE POTENTIALS Januar y/Februar y 20 2 1 ■ 27
Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on October 12,2021 at 20:56:54 UTC from IEEE Xplore. Restrictions apply.
 user preferences, such as less walk- transport operators, such as the LTA avoid the traffic jams and congestion
ing and fewer transfers. Other apps, of Singapore, that provide the neces- that typically evolve continuously on
such as MyTransport.SG, are widely sary information through app pro- urban roads. The Google Maps app,
used in Singapore to check the gramming interfaces for developers once again, dominates the market
arrival times of buses at a particu- to integrate into their apps. Similar with its free-to-use business model,
lar bus stop so that commuters can efforts are also being made by trans- reliable and dynamic route recommen-
leave their homes just in time to port operators in other cities to make dations, as well as continuous on-the-
catch the bus and avoid waiting at relevant information accessible to go traffic updates. These apps can be
the bus stop for too long. commuters in a timely manner. used to not only save travel time but
Similarly, some apps also no- also to better plan one’s journey based
tify users of nearby traffic jams and Apps for motorists on current and even estimated near-
offer alternative choices to save Similar to apps developed for commut- future traffic conditions.
travel time. These apps and oth- ers, apps have also been created to aid It should be noted that, while
ers like them have been made pos- motorists and drivers in journey plan- some apps, such as Waze, rely on the
sible through conscious efforts from ning and even dynamic replanning to crowdsourcing of information regard-
ing traffic conditions, accidents, and
work zones, most others largely de-
pend on existing infrastructure and
government entities to obtain most of
the relevant information. This moti-
vates the designers of smart cities to
incorporate relevant sensors and in-
frastructural support to provide the
necessary information in real time.

Traffic monitoring and

management
Leveraging technology for traffic
monitoring and management is not a
new concept. Roadside cameras have
traditionally been used to monitor
traffic conditions on roads, such as
heavy or light traffic, road incidents,
or accidents. However, new and
improved technologies make it
increasingly seamless to achieve
these goals in real time so that appro-
priate steps can be taken as soon as
(a) (b) possible. For example, in the case of
accidents, vital seconds or minutes
FIG3 (a) A Google Maps image of public transport options. (Source: Google Maps An- can be saved by the latest cameras
droid app.) (b) The Waze app for current traffic status. (Source: Waze.) with edge computing capabilities that
can automatically detect such occur-
rences and send alerts to the appro-
priate authorities together with the
precise location of the accident.
In a smart city with smart mobility
solutions in place, it is also desirable
to keep the roads clear of any obstruc-
tions, such as illegally parked vehicles,
particularly near the curb. Since pub-
lic buses typically use the left lane to
embark and disembark passengers, it
is imperative to keep these lanes free
of obstructions to achieve faster travel
time for public transport users.
FIG4 An embedded inductive loop detector (marked with red arrows) for vehicle detec- Typically, cameras are installed at
tion. (Source: Google Street View.) hot spots along the side of the road

28 ■ J a n u a r y/Febr uar y 2021 IEEE POTENTIALS

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on October 12,2021 at 20:56:54 UTC from IEEE Xplore. Restrictions apply.
 to monitor these lanes. A large ma- that are being leveraged under the aggregator.” Techcrunch.https://tech
jority of these cameras record the larger umbrella of smart city initia- crunch-com.cdn.ampproject.org/
video footage on network video re- tives. The major focus of govern- c/s/techcrunch.com/2019/11/25/
corders that are also installed near ments and authorities across the shuttl-india-bus-aggregator/amp/
the camera. An operator collects the world is to entice more people to use (accessed Oct. 1, 2020).
video data periodically and analyzes PTSs instead of private vehicles. To • Z . A bdu l la h. “On-dema nd
the video for violations in a command this end, the article focused mainly bus services Beeline, GrabShuttle
and control center, either manually on strategies to improve PTSs. It shutting dow n end-2019.” Chan
or automatically, using video-ana- a lso ex plored ot her mea ns of nel ne w sa si a ht t ps://w w w.chan
lytics software. In future smart cit- improving passenger experience n e l n e w s a s i a . c o m / n e w s/sin
ies, camera-based sensors with edge through smartphone apps that have gapore/on-demand-bus-ser vice s
computing capabilities can perform been made possible through effec- - b e e l i n e - g r a b s hut t l e - s hut t i n g
such analytics in real time on the tive collaborations between govern- -down-end-12118442 (accessed Oct.
edge, thereby obviating the need for ments, transport authorities, and 1, 2020).
manual intervention to a large extent. operators as well as private app • “RRT-2—Technologies for holis-
Traffic management, on the oth- developers. Finally, a brief discus- tic traffic management and control.”
er hand, is primarily done through sion was presented regarding how https://w w w.tum-create.edu.sg/
traffic lights by controlling the time modern technology can be leveraged research/project/rrt-2-%E2%80%94
of the green and red signals along for effective traffic monitoring and -tech nolog ies-hol ist ic-t ra f f ic
the various directions of a traffic management. With the great strides -management-and-control%C2%A0
junction. Smart traffic lights rely being made currently for smart (accessed Oct. 1, 2020).
on inductive loop detector technol- mobility, smart and even smarter • “Factsheet: Passenger load infor-
ogy, or loop detectors (Fig. 4), to de- cities will be a reality very soon. mation system piloted on downtown
tect passing vehicles at junctions line for smoother boarding,” Land
and control the traffic light tim- Read more about it Transport Authority, Singapore, May
ing accordingly. • ”Smart Cities Mission.” Minis- 14, 2018. [Online]. Available: https://
However, loop detectors are ex- try of Housing and Urban Affairs, w w w.lt a.gov.sg/content/lt agov/
pensive to install and maintain Government of India. http://smart en/newsroom/2018/5/2/factsheet
because they need to be embedded cities.gov.in/content/ (accessed Oct. -passenger-load-information-system
in the road surface. They are also 1, 2020). -pi lot e d- on- do w nt o w n-l i ne -for
unable to accurately distinguish • “Taxi availability improves for sec- -smoother-boarding.html (accessed Oct.
between different types of vehicles ond consecutive year,” Land Transport 1, 2020).
and, as a result, cannot be relied Authority, Singapore, Nov. 26, 2014. • DataMall. https://www.mytra
upon to prioritize specific types of Accessed: Oct. 1, 2020. [Online]. Avail- nsport.sg/content/mytransport/
vehicles, such as public buses. able: https://www.lta.gov.sg/content/ home/dataMall.html (accessed Oct.
With the ever-growing popularity ltagov/en/newsroom/2014/11/2/ 1, 2020).
of using machine learning techniques taxi-availability-improves-for-second • Spacex. https://www.spacex
for weather-invariant classification -consecutive-year.html .com/sites/spacex/files/hyperloop_
of objects, researchers from Nanyang • A. Lim. “LTA removes require- alpha-20130812.pdf (accessed Oct.
Technological University, Singapore, ment for taxi drivers to clock a 1, 2020).
have been experimenting with using daily minimum mileage.” Strait-
camera-based sensors for estimating s t i m e s . h t t p s : // w w w . s t r a i t About the author
the queue length at traffic lights and to st i me s.com/singapore/transport/ Alok Prakash ([email protected])
distinguish between public and private lta-removes-requirement-for-ta xi earned his Ph.D. degree from the
vehicles. Such sensors have numerous -drivers-to-clock-a-daily-minimum School of Computer Science and
benefits when compared to traditional -mileage (accessed Oct. 1, 2020). Engineering, Nanyang Technologi-
loop detectors in terms of installation • “Kutsuplus: Final report,” Hel- cal University, Singapore, in 2014.
and maintenance cost as well as sup- sinki Regional Transport Authority, He is a senior research fellow at
ported features. It is envisioned that, in Helsinki, Finland, 2016. Accessed: the School of Computer Science
the future, the large-scale deployment Oct. 1, 2020. [Online]. Available: and Engineering, Nanyang Tech-
of low-cost camera-based sensors for https://w w w.hsl.fi/sites/default/ nological University, Singapore.
both traffic monitoring and manage- files/uploads/8_ 2016_kutsuplus He is also the principal investiga-
ment can be relied upon to enable _finalreport_english.pdf tor of the Sensing and Manage-
smart mobility in a smart city. • Iskandar Malaysia Bus Rapid ment for Agile Transport project
Transit. https://imbrt.com.my/ (ac- under the TUMCREATE Phase 2
Conclusion cessed Oct. 1, 2020). program.
This article provided an overview of • M. Singh. “India’s Shuttl raises
various solutions for smart mobility $36M to expand its app-based bus 

IEEE POTENTIALS Januar y/Februar y 20 2 1 ■ 29
Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on October 12,2021 at 20:56:54 UTC from IEEE Xplore. Restrictions apply.