Disruptive technology and innovation in transport - Policy paper on sustainable infrastructure - EBRD
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
August 2019 Disruptive technology and innovation in transport Policy paper on sustainable infrastructure
Executive summary
A key objective of the European Bank for key challenge in the development of the identified
Reconstruction and Development (EBRD), disruptive technologies and their applications will
especially in the transport sector, is to support be their successful integration into new business
the promotion of innovative new technology in the and governance models, maximising their combined
economies where the Bank operates to improve benefits to support the end goal.
competitiveness and provide demonstration
effects. The purpose of this paper is to provide an The four applications of the disruptive technologies
overview of the current state of the market and that Section 3 reviews in detail are as follows:
opportunities for the implementation of a range
of (disruptive) digital technologies capable of • Traffic management using intelligent transport
revolutionising the transport sector in the EBRD systems (ITS) – using new technologies to predict
regions. These technologies include: future traffic demand more accurately and optimise
road networks accordingly, providing a wide range
• the internet of things (IoT) – a system of objects, of social and economic benefits, including reduced
processes, data and people connected with each congestion and pollution, improved safety and
other via sensors, and controlled remotely using travel experiences for all road users.
the internet
• Personal travel planning and public transport –
• big data – complex data characterised by high analysing available information on travel demand
volume and requiring the use of advanced and travel patterns of the population, to facilitate
analytics for processing the optimisation of planning, programming and
operation of public transport systems, as well as
• artificial intelligence (AI) – computer science improving personal journey planning for the public.
which enables machines to function like
a human brain • Autonomous and connected vehicles for
mobility – developing applications for AVs which
• drones – unmanned aerial vehicles (UAVs) or can contribute to increased safety, a better user
flying robots. experience, economic savings and reductions
in congestion, by facilitating car sharing and
The paper outlines a range of digital technologies “mobility as a service” (MaaS).
and concepts (Section 2), introduces various
technology application areas with supporting case • Unmanned aerial vehicles/drones for
studies and cost-benefit analysis (Section 3) and monitoring - using technology to revolutionise
discusses a policy roadmap for their successful the way we undertake asset management,
implementation (Section 4). maintenance and inspections (bridges, tunnels
and construction sites) and providing an efficient
The summary of the technologies presented in means to deliver packages (logistics).
Section 2 demonstrates that IoT, big data and AI
do not operate in isolation but instead represent These technology application areas were reviewed
highly complementary technologies. Big data is in the context of their contribution to the following
collected most effectively using IoT systems and policy objectives: (1) transport efficiency, (2) safety and
drones and then processed most efficiently using security, (3) environment and climate change and
AI algorithms and optimisation techniques. The (4) socio-economics. From the analysis of these policy
main applications of these particular technologies objectives we concluded that the technology application
in transportation focus around demand forecasting areas which have the most profound (disruption)
and traffic optimisation resulting in better traffic potential impact were new smart mobility (AVs/MaaS
management, asset management, travel planning and drones) and intelligent transport systems (ITS), each
and operation of autonomous vehicles (AVs). The requiring and leveraging different digital technologies.
Policy paper on sustainable infrastructure August 2019 1The key challenge in the development of the • Identifying requirements for facilitating necessary
identified digital technologies and their applications enabling “public” infrastructure and forms of
will be to integrate the business and governance economic regulation to enable widespread
models for new mobility technologies, services and adoption.
systems successfully. The following challenges are
critical to this process: • Developing cost-benefit analysis methodologies
and the supporting evidence base to promote
• Harmonising existing and new policies related to adoption.
the legal framework for use and operationalisation
of such technologies. • Launching analytical work and developing
innovative operating models.
• Facilitating interoperability and data sharing.
• Developing integrated mobility systems.
• Promoting vehicle-to-infrastructure (V2I) and
vehicle-to-vehicle (V2V) communication. • Sharing data and digital infrastructure.
• Ensuring data security and addressing risk- • Supporting capacity-building, education and
sharing/liability concerns. awareness-raising.
2 August 2019 Disruptive technology and innovation in transportContents
1. Introduction 5
1.1. Objectives 5
1.2. Structure 5
1.3. Background
Economics of new technology 6
2. Disruptive technologies 10
2.1. Internet of things – a data collection and management tool 10
2.2. Big data – the data 11
2.3. Artificial intelligence – a data tool for processing
complex datasets 12
2.4. Drones – an alternative data collection and exploitation tool
for monitoring 14
3. Applications in transport 16
3.1. Traffic management using intelligent transport systems 16
Overview 16
Enablers, barriers and opportunities 17
Cost-benefit analysis 18
Case study – “Talking Traffic” in the Netherlands 20
Case study – City parking 21
3.2. Personal and public transport travel planning 23
Overview 23
Enablers, barriers and opportunities 23
Cost-benefit analysis 24
Case study – Personal travel planning in Perth, Australia 25
Case study – Public transport in Singapore 26
3.3. Autonomous and connected vehicles for mobility 27
Overview 27
Enablers, barriers and opportunities 27
Cost-benefit analysis 29
Case study – CAVs on a freeway in Antwerp modelling study 33
Case study – Port of Rotterdam 33
Case study – Autonomous vehicles (providing mobility as a service) 34
Policy paper on sustainable infrastructure August 2019 3
3. Applications in transport (continued) 00
3.4. Unmanned aerial vehicles/drones for monitoring 35
Overview 35
Enablers, barriers and opportunities 35
Cost-benefit analysis 36
Case study – Amazon Prime Air 37
Case study – Elios indoor drone for bridge inspection in Minnesota 39
4. Policy roadmap 40
4.1. Policy objectives 40
Transport efficiency policy objective 40
Safety and security policy objective 40
Environment and climate change policy objective 40
Socio-economic policy objective 41
4.2. Policy recommendations to assist with barrier removal 41
Barrier 1: Legal and regulatory framework 42
Enabling policy 1: Harmonising existing and new policies 42
Barrier 2: Data fragmentation and multiple platform development 42
Enabling policy 2: Facilitating interoperability and data sharing 42
Barrier 3: Security, insurance and privacy concerns 43
Enabling policy 3: Ensuring data security addressing risk
sharing/liability concerns 43
Barrier 4: Prohibitive cost or lack of economic and financial
evidence of return on investment 43
Enabling policy 4: Developing cost-benefit analysis methodologies
and the supporting evidence base 43
4.3. Cross-cutting issues and opportunities 43
Bibliography 47
Annex A. Policy objectives 51
Glossary of terms 52
4 August 2019 Disruptive technology and innovation in transport1. Introduction
1.1. Objectives These technologies were selected because:
One of the EBRD’s key medium-term priorities is • their technological applications have specific
“digitisation and digital development”. An important relevance to the transport sector
aspect of this objective is the promotion of new • they have potential to provide benefits from
technology and innovation that can improve the structural change, through addressing congestion
competitiveness of key sectors and businesses and pollution to improved safety and a wide range
in the EBRD regions. The Bank is now considering of social and economic benefits
a range of new technologies and innovations that • they have the most potential to revolutionise how
are being developed and adopted in many parts we live, work and travel in the next 10-20 years
of the transport sector. It seeks to understand • they primarily rely on the use of digital technology
the opportunities for EBRD clients to adopt and • they have active programmes of implementation
implement such technologies, including overcoming with test programmes in Europe and in Central Asia
potential barriers to entry and adoption, with the
Bank’s support. Conversely, the following technologies, while they
are very relevant to transport and have the potential
This paper provides an overview of the current to significantly disrupt the sector, are not discussed
state of the market and opportunities for the in detail. This is because they are deemed to rely
implementation of specific digital technologies in the less on the digital transformation of the fourth
transport sector. It discusses a range of potential industrial revolution, and would require engineering
application areas including an assessment of the transformation to be fully captured by the EBRD policy
potential costs and benefits of digital technologies and business model. Further information is available
capable of “disrupting” the transport sector. As such, elsewhere on:
the paper will be of direct interest to the EBRD’s in-
country transport teams, but also to municipalities • electric vehicles (EVs)
and regional/national transport authorities, • advanced materials
illustrating the potential (and successful) application • energy storage technologies (for example,
of selected disruptive technology in different lithium batteries)
contexts within the transport sector, and a roadmap • advanced robotics and manufacturing.
towards leveraging such technologies better. The
following technologies are introduced and discussed 1.2. Structure
in the context of the transport sector:
The structure of this paper is as follows:
• Internet of things (IoT) – system of objects,
processes, data and people connected with each Section 2 provides a summary of different digital
other via sensors, and controlled remotely using technologies and concepts, covering (1) internet
the internet. of things, (2) big data, (3) artificial intelligence and
• Big data – complex data characterised by (4) drones. These technologies are summarised in
high volume and requiring the use of advanced terms of potential application areas, components,
analytics for processing. barriers to implementation, technology and policy
• Artificial intelligence (AI) – computer science enablers and opportunities for further development
which enables machines to function like a human and implementation.
brain.
• Unmanned aerial vehicles (UAVs/drones) This section provides a qualitative description of
or flying robots. the various technologies considered in this paper
and a discussion of areas of commonality and
complementarity. Indeed, many of the technologies
considered here are typically applied in combination
Policy paper on sustainable infrastructure August 2019 5as technology stacks. This is because cost-benefit car. However, the situation is changing rapidly: since
information is not readily available for discrete 2002, the number of kilometres driven per person has
technologies, but rather for areas where these fallen by 8.5 per cent (Deloitte, 2015). Meanwhile, use
technologies, when applied in combination, can yield of public transport has increased. This trend suggests
significant economic and environmental benefits. that urban residents are becoming more likely to
consider new ways of travelling and to move away
In Section 3 we introduce different application from the traditional car ownership model in favour of
areas, related to the transport sector, to help frame new forms of transport such as car sharing, electric
each of these technologies, focusing on their vehicles, autonomous cars and mobility-as-a-service
potential impacts and implications. We explore (MaaS) solutions (Deloitte, 2015).
several case studies for each of the application
areas from Europe, Central Asia and elsewhere, The past 100 years have also been characterised
together with information on the costs and benefits by significant growth in car ownership, which has
of those applications and specific schemes been linked with global drivers of suburbanisation
implemented around the world. as well as increasing incomes and consumer
purchasing power. The proportion of the global
These case studies provide practical examples population living in urban areas continues to
of some of the challenges and opportunities rise faster than the capacity of roads and public
associated with the implementation of these transport. The pressure on transport infrastructure
technologies, and an outline of their disruption is significant and cannot be resolved by simply
potential. After reviewing the most prominent building more infrastructure. New, innovative
disruptive technologies and potential application solutions and approaches are required to address
areas in transport, we identified several areas of these problems. The use of new digital technologies
public policy that might warrant further examination. is a key part of addressing this challenge and can
This is discussed in more detail in Section 4. help to ensure more efficient and sustainable use
A roadmap for implementation is presented, of existing infrastructure. At the same time, it can
comprising a range of applications that can build encourage the public to abandon their cars in favour
on these disruptive technologies, as well as of walking, cycling and shared mobility solutions
potential barriers, bottlenecks and opportunities. (Webb, 2019).
1.3. Background These new forms of transport rely increasingly on
exploiting the use of digital technologies, which are
Traditional methods of overcoming critical transport revolutionising the way we travel and communicate.
and infrastructure challenges are increasingly The ability to collect vast amounts of data (“big
subject to technology-based disruptions, creating data”) and process it in real time using advanced
new opportunities. We are now in the fourth analytics and AI will allow us to predict transport
industrial revolution – but this one is happening demand better and, as a result, improve our ability
much faster than any of its predecessors. to manage existing infrastructure. Ensuring that
The accelerating pace of technology diffusion assets are connected and communicate with each
and its updates, the convergence of multiple other through internet-of-things protocols and
technologies towards human-centric goals, or platforms will provide new ways to organise traffic,
common applications, and the emergence of global travel and logistics, while permitting the remote
platforms are disrupting traditional transport and control and management of assets and networks.
infrastructure development models. The use of robots, such as drones, is already
revolutionising how we manage our assets and
In transport, demand continues to grow each year, undertake infrastructure inspections and surveys,
with Europeans, on average, travelling around as well as supporting logistics and deliveries of
35,000 passenger kilometres per year; with a clear consumer goods and services in many established
majority (64 per cent) of these trips being made by and emerging global markets.
6 August 2019 Disruptive technology and innovation in transportThe digital age has the potential to bring with it Economics of new technology
a range of disruptive technologies. Indeed, the pace
and the scale of the changes is expected to increase The transformative nature of disruptive technologies
due to the rapid development in digital technologies. makes their economic and financial analysis
In the past, disruptive technologies would have been challenging. By definition, disruptive technologies
viewed as unknown and unproven, often considered make more fundamental changes and affect deeper
impractical for real-world application. In many structures – changing the way existing markets
cases these disruptive technologies would displace operate, creating new players and displacing old ones.
established firms in existing markets. For example, The analysis of their impacts requires a corresponding
mainframe computer manufacturers in the 1970s economic and financial methodological approach
and 1980s underestimated the potential demand that considers broader outcomes than those typically
for personal computers. As a result, companies like applied to transport projects and investment.
Apple and Microsoft disrupted the market with their
new products, while major manufacturers dismissed The types of disruptive technologies proposed here
personal computers and overlooked a market that have a wide range of potential applications and
did not yet exist (Baker et al., 2016). impacts across society generally. Furthermore,
changes due to the deployment of, for example, big
Today the term “disruptive” is often used to describe data solutions can have significant implications
technological advancements which are new, evolve across a range of sectors at once. As such, these
rapidly and have a significant impact on how we technologies change the context for transport as well
live and work, as well as on our economy. To ensure as transport itself and result in a changing economic
that society is ready for these new technologies, and financial landscape. In the World Economic
governments, policymakers and lawmakers will Forum’s study Deep Shift: Technology Tipping Points
need to gain a good understanding of how the future and Societal Impact (20), three of the technologies
is going to unfold and make the right investment considered are identified as the subject of “tipping
decisions in infrastructure and education so that point” considerations – big data for decisions
societies continue to prosper. In today’s society, (expected to be common by 2023); driverless cars
digitisation and disruptive technologies such as (by 2026); artificial intelligence and decision-making
big data, IoT, AI and drones have the potential to (also by 2026). For these, cost-benefit analysis is only
change the way the transport sector is organised a partial guide to their feasibility as the structural
and managed, paving the way for new services and changes they both cause and require depend on a
business models. wider range of factors.
For the transport sector, the economic impacts
of new technology will occur through several
mechanisms affecting the demand and supply sides
of the economy: (a) reducing the need for travel
through substitution; (b) improving the efficiency
and convenience of travel by creating new modes,
improved route planning, more efficient vehicles,
and in vehicle services and so on; (c) improving the
efficiency of infrastructure construction, operation
and management; (d) improving the efficiency of
transport operators and other businesses (through
more competition, new services and new market
structures); and (e) externalities such as reduced
emissions, productivity gains, better information for
public planning and so on.
Policy paper on sustainable infrastructure August 2019 7The increasing capability of virtual technologies Issues of data ownership within and across
will reduce the need to travel by allowing remote organisations can complicate aggregation. Owners
observation and communication, but will also of data from one system might not find it in their own
contribute to changes in the relative economic commercial interest to have their data combined
values of both new and old goods and services, with data from other systems (see website link 23
changing incentives for travel and transport. Closely at the back of this report). Another example lies
related to changes that affect overall demand are in the provision of infrastructure that often is or
changes that are fundamental to a certain sub- has aspects of natural monopoly. It is complex to
sector of the transport network. A clear example develop solutions using disruptive technologies
is the retailer Amazon’s proposed use of drones that address this, are timely and coordinated, and
for “final mile” delivery to customers which would permit the benefits of competition. With common
completely substitute an existing part of the limited and widespread infrastructures in place, such as
capacity of the current terrestrial distribution system. roadside or in-vehicle sensors, a range of value-
added services becomes feasible. However, the
Cost-benefit analyses of these technologies show need for, and revenue generated from, any one
that their economic viability is often clear, but their service may be insufficient to cover the costs,
development is inhibited in practice by many barriers thus making the implementation of risk-sharing
to market developments (detailed for each technology arrangements and associated financing structures
and application in the next section of this report). substantially more complex.
These barriers mainly fall into three categories – lack
of transparency over the potential benefits of the In the construction industry, developers in
technology; the distribution of costs and benefits, worksite industries are working on two potential
which may mean that the benefits are not captured applications that are too nascent to reach their
by those bearing the costs; and regulatory barriers market potential today, and present too many
that prevent the adoption of new technology due, barriers for development at this stage: fully robotic
for example, to perceived safety risks. worksites and 3D printing of replacement parts
on-site. Given the labour intensity, unpredictability
For instance, at the technical level, the inability and danger of some worksite environments, being
to capture and use relevant data from multiple able to remove employees from the site entirely
streams generated by different systems (ITS or IoT) would offer substantial productivity and safety
is the result of several organisational, technical benefits, particularly for assets that are difficult
and commercial barriers. In some cases, a lack to reach. Many barriers to full automation remain,
of understanding of the potential to use data has including the need for more sophisticated robotics
led to a failure to invest in deploying tech-enabled and safety concerns about unmanned operations
solutions. But there are also technical challenges, (especially for bridges and tunnels). The ability to 3D
including finding efficient ways to transmit and store print replacement parts on demand could greatly
data. The most fundamental challenges are in data reduce downtime caused by equipment failure and
transmission and storage. Many IoT applications could raise asset utilisation and output. However,
are deployed on remote or mobile equipment. this would require equipment that produces parts
Real-time transfer of all the data being generated by that meet performance standards. If this challenge
the sensors on aircraft engines would require more could be resolved, worksites would be able to reduce
bandwidth than is currently deployed. If data can be substantially the cost of carrying a spare parts
collected and stored, the next obstacle is aggregating inventory and could avoid delays caused by out-of-
it in a format that can be used for analysis. Limited stock parts.
standardisation of data means that substantial
systems integration work is needed to combine data The analysis reported in the literature to date has
from multiple sources. This challenge is accentuated taken a variety of approaches to the definition of
by connectivity and storage challenges. scope for cost-benefit analysis. Skeete (2018)
8 August 2019 Disruptive technology and innovation in transportnotes that “there is no universal, valid definition The scope of cost-benefit analyses in the literature
to acceptance nor a single approach, but a broad has typically not sought to represent the costs for
range of theoretical constructs”. In practice, the overcoming these factors. However, expenditure on
authors choose a fixed set of mainly transport- lobbying, for example to change regulation, would
related assumptions. For example, a study on typically be part of corporate behaviour.
autonomous vehicles notes that “most studies
conduct micro-technical examinations of specific In the assessment of costs and benefits, the types
components within the autonomous vehicle” of benefit commonly considered are the following:
(Skeete, 2018).
• Time savings to individuals (for example, from
While this reduces the complexity of the analysis reduced congestion).
in each study, assumptions are often particular to • Savings from fewer automobile accidents (health,
the individual themes of the studies and this can less disruption).
reduce their comparability. Furthermore, a focus on • Energy savings (from reduced trips and from more
applications (how alternative technologies might efficient use).
solve the same problem) as opposed to individual • Environmental benefits (mainly related to
technologies (the many ways in which each greenhouse gases and improved air quality).
technology can contribute) widens the number
of situations addressed by each technology, Less commonly considered benefits are as follows:
correspondingly increasing the number of cost-
benefit ratios that are relevant to each. This • Maintenance savings (on roads and vehicles) from
makes comparison difficult. lower use or fewer trips.
• Environmental benefits not related to emissions
The scope of economic and financial analysis (such as noise).
tends to be tied to fixed aspects of the existing • Savings in the supply chains (for example, reduced
transport system, notably the volume of trips. demand for road materials) (20).
Using a benchmark of a fixed volume of trips, it
is possible to compare disruptive technologies The types of cost considered are relatively clear
to more traditional ways of achieving the same in the specific studies but subject to variability
impacts. For example, technological solutions can when considered across the studies as a group.
reduce congestion, avoiding the need to increase In general, studies focus more on financial than
road capacity to maintain or improve trip times. economic savings, with elements such as road
There can be associated benefits of reduced fuel costs being excluded (reflecting current charging
consumption and emissions. The methods for structures, where road networks are free at the point
valuing these benefits are already established of use), rather than full economic costs. Similarly,
using traditional methods. The estimation of the elements such as mobile phones may be assumed
costs, arguably the more uncertain element, can to be available at zero additional cost because they
nevertheless be based on detailed knowledge of the are assumed to have already been purchased.
new technology. While the costs and benefits can While this has some impact on the structure of the
be defined, overcoming the issues of transparency, analysis, it also reflects particular perspectives
distribution (allocation) of costs and benefits and on the availability and ownership of pre-existing
regulation may be the greater challenge. Overall, infrastructures which are often key to the incentives
these factors, and their ultimate influence on the for future participation and collaboration.
level of uptake, are likely to be those that determine
the overall viability of a new technology. Fagnant
(2015) identifies that, among a range of missing
research, “one of the most pressing needs is a
comprehensive market penetration evaluation”.
Policy paper on sustainable infrastructure August 2019 92. Disruptive technologies
2.1. Internet of things – a data collection IoT technologies have a number of applications in
and management tool the transport sector, including intelligent transport
systems (ITSs) which use data collected from sensors,
The internet of things (IoT), often referred to as actuators, cameras and micro-controllers to optimise
the “internet of everything” is a system of objects, public transport, reduce congestion, monitor the
processes, data, people and even atmospheric environment and run security applications (Hill et al.,
phenomena, connected with each other via various 2017). From the transport sector’s perspective, the
types of embedded sensors, and controlled remotely IoT could significantly change the way government
using the internet (Witkowski et al., 2017). The entities provide transport services by allowing
applications of the IoT play an increasingly important transport infrastructure assets to be monitored and
role in smart transport and more recently in the operated in real time from remote locations. For
“smart cities” agenda, helping to control traffic, example, International Business Machines (IBM)
monitoring weather and safety risks, providing has developed systems that aggregate data from
information about the state of the roads and infrastructure-based sensors and similar devices
monitoring accidents. IoT platforms help manage, to identify and measure traffic speed and volume
analyse and compile data from a wide variety of on city roads. This provides road agencies, and in
sensors, including proximity, infrared, image and some cases the motoring public, with real-time traffic
motion detection sensors. conditions, which can assist in incident response
and routing activities (Baker et al., 2016).
Table 1. The internet of things – applications, barriers and opportunities
Applications Barriers Opportunities
• Traffic management (ITS) • High cost • Reducing congestion (savings
• Demand modelling and forecasting • Security and privacy concerns in time, fuel, improved air quality)
• Asset management • IoT platforms • Making transport safer and
• Freight tracking • Interoperability and standards more efficient (vehicle tracking,
• Logistics (tracking of deliveries) (integration inflexibility) travel planning)
• Parking management • Legal issues around the internet • More accurate forecasting
• Personal travel planning • Requirement for technical skills (incident detection)
• Public transport planning • Requirement for infrastructure • Logistics (deliveries tracking)
• Autonomous vehicles readiness • Social and economic benefits
• Vehicle-to-vehicle (V2V) and
Components Enablers vehicle-to-infrastructure (V2I)
communication
• IP (internet protocol) software • APIs
platform • Ubiquitous (low-cost or high-speed)
• Computers connectivity
• Device-to-device communication • IP-based networking
(Bluetooth, Z-Wave, ZigBee) • Computing economics
• Device-to-cloud communication • Miniaturisation
(Ethernet, wi-fi) • Artificial intelligence
• Device-to-gateway (application • Advances in data analytics
layer gateway service) • Enhanced computing capabilities
• Cloud computing
• IPv6 IP protocol development
• Blockchain
10 August 2019 Disruptive technology and innovation in transport2.2. Big data – the data The transport sector has always collected and
analysed large quantities of data, including traffic
“Big data” refers to complex data that is surveys, data from timetables, and, more recently,
characterised by high volume (ranging from 1,000 data from traffic cameras, mobile phones and
gigabytes to 1 petabyte, equivalent to 1 million sensors. Historically, quantitative urban research
gigabytes in size), high velocity (in order to be useful, has relied on data from surveys and censuses. All
it needs to be analysed rapidly in “real time”) and of these data sources are expected to continue
high variety (normally comprising several different to play a vital role in urban analysis. However,
sources of data). The rapid increase in the availability recent developments in the quantity, complexity
and complexity of data has led to the term “big and availability of big data, together with advances
data”, although it does not have a universally agreed in computing technology, are presenting new
definition (Houses of Parliament, 2014). However, it opportunities to create more efficient and smarter
is generally accepted that big data tends to be too transport systems. Figure 1 shows the main big data
complex to be analysed using traditional methods sources and the three component layers required to
and requires the use of advanced analytics and support smart infrastructure (Hill et al., 2017).
computational algorithms.
Figure 1. Big data basic layers for smart infrastructure connected by the IoT
g
k in
Rule-based
ma
automation Improved
io n
decisions
c is
Decision Machine
De
support learning
Decreasing data tools
volume, increasing Organisation
data value algorithms
ing
ak
Learning
Big data Data
em
Modelling Analytics Improved
analysis mining
intelligence
ns
Se
nt
me
Data Data Data
ge
cleaning structure storage
na
ma
Assets Customers Costs Activities
ta
Da
SCADA Customer GPS Ticketing/ Social Sensors Drone Laser Satellite GIS Manufacturer's CCTV Scanned Control
systems billing counting media surveys surveys imagery and BIM data images system
Source: Hill et al. (2017).
Policy paper on sustainable infrastructure August 2019 11In the context of the transport sector, big data is 2.3. Artificial intelligence – a data tool for
mostly associated with map data, vehicle location processing complex datasets
data, traffic control information, personal location
data, payment or transaction data and public Artificial intelligence (AI) refers to computer science
transport information. Big data can be collected and algorithms that enables machines to function
from a number of sources and using a variety of like a human brain, analysing complex datasets
methods, such as GPS or satnav, mobile devices for trends and patterns. Examples of AI methods
(Bluetooth or wi-fi), cameras and sensors (for that are being increasingly applied in the transport
example, RFiD2). The IoT often acts as an enabler sector include artificial neural networks, genetic
for big data collection, providing an ecosystem of algorithms, simulated annealing, and fuzzy logic
sensors and data platforms, capable of collecting models (Abduljabbar et al., 2019). By modelling
and processing a vast amount of information quickly a relationship between the cause and effect
and efficiently. of different real-life scenarios, AI helps bridge
uncertainties and gaps within the data that
cannot be resolved using traditional methods.
Table 2. Big data – summary of applications, barriers and opportunities
Applications Barriers Opportunities
• Traffic management (ITS) • Data availability and openness • Reducing congestion (savings
• Strategic planning of data in time, fuel, improved air quality)
• Demand modelling • Data usability or accuracy • Making transport safer and
• Asset management • Data processing more efficient (vehicle tracking,
• Travel planning • Lack of technical skills for advanced travel planning)
• Route guidance data analytics • More accurate forecasting
• Disruption alerts • Privacy issues (incident detection)
• Infrastructure management • Data storage • Integrated cashless payments
• Operational insight • Mobile data on public transport
• Autonomous vehicles • Willingness to share
• Lack of information on private sector
data available
Components Enablers
• Map data • Internet of things
• Weather data • Artificial intelligence
• Personal location data • Machine learning
• Public transport schedules • Advanced analytics (predictive
• Vehicle location data and real-time)
• Fare and pricing data • Blockchain
• Payment or transaction data • Enhanced computing capabilities
• Smartphone sensors (GPS, • Cloud computing
accelerometer, camera) • Social media
12 August 2019 Disruptive technology and innovation in transportThe application of artificial intelligence in the situations and events. An area where AI applications transport sector centres around road and public have also seen rapid development is intelligent transport planning, traffic incident detection transport systems (ITSs), where AI and machine and predicting traffic conditions. The intelligent learning (ML) techniques are used to find patterns computational analytics of these systems are able and features in the captured data to allow real-time to represent uncertainty, imprecision and vague optimisation of traffic control policies and to achieve concepts, hence can be used for route optimisation more connected transport systems. problems in transport, including dynamic traffic Table 3. Artificial intelligence – applied solutions, barriers and opportunities for the transport sector Applications Barriers Opportunities • Big data analytics • Lack of infrastructure • Better detection and prediction • Corporate decision-making, • Dependent on the quality or of travel patterns planning and managing reliability of data • Better traffic forecasts • Accurate prediction and • “Black box” effect – limited • Improvements to public transport detection models understanding of the relationship (enhanced reliability) • Traffic flow/volume forecast between input and output • Integration with shared mobility • Traffic conditions forecast • Not capable of forecasting under (Uber) • Improvements in public transport unexpected events and adverse • Enabling MaaS • Traffic incident prediction weather conditions • Enabling smart city initiatives • Traffic management (ITS) • Computation complexity of • On-demand public transport services • Smart highways AI algorithms • Improved productivity • Smart rail • Lack of advanced analytics skills • Creation of new jobs • Traffic signal control • Lack of technological infrastructure • Asset management to support AI • Travel planning • Fragmentation and incompatibility • Logistics of data • Robotics • Data privacy issues • Autonomous vehicles • Impact of automation (jobs • Drones displacement or loss) • Customer analytics • Predictive maintenance Components Enablers • Knowledge-based system • Big data • Artificial neural network systems • Internet of things • Machine learning • Blockchain • Deep learning techniques • Computing power and speed • Genetic algorithm • Algorithmic improvements • Simulated annealing algorithm • Talent and skills • Ant colony optimiser algorithm • Investment and funding • Artificial immune system algorithm • Bee colony optimisation algorithm • Swarm intelligence systems • Fuzzy logic model • Logistic regression model • Agent-based software engineering Policy paper on sustainable infrastructure August 2019 13
2.4. Drones – an alternative data Maintaining roads, bridges and tunnels at the
collection and exploitation tool optimum level can be very costly. Inspecting the
for monitoring deck of a bridge, for example, could take four
workers an entire eight-hour shift to complete.
A drone, in technological terms, is an unmanned This would also involve heavy-duty equipment
aircraft. Drones are more formally known as and could cost nearly US$ 5,000. In addition,
unmanned aerial vehicles (UAVs) or unmanned a traditional bridge inspection would need to take
aircraft systems (UASs). Essentially, a drone is a place during the daytime and would require the
flying robot that can be remotely controlled or fly re-routing of traffic, which would have additional
autonomously using software-controlled flight plans cost implications. Using a drone to inspect the same
in their embedded systems, working in conjunction bridge would require only two people, no heavy-duty
with on-board sensors and GPS. There are two main equipment and limited traffic control and monitoring,
types of drones: rotor (tricopters, quadcopters, with the entire process taking about two hours.
hexacopters and octocopters) or fixed-wing, which This would provide significant savings on staff and
include the hybrid VTOL (vertical take-off and equipment and improve efficiency and safety (14).
landing) drones.
The summary of the disruptive technologies
Drones are being increasingly used in the transport presented above shows that IoT, big data and AI
sector to improve operational efficiency, save money do not operate in isolation. Instead, they are highly
and time and increase safety. The technology complementary technologies. Big data is collected
has been used to inspect bridges and tunnels, as most effectively using IoT systems and drones and
well as monitoring traffic and in logistics delivery. then processed for optimisation and forecasting
Infrastructure can be inspected and made using AI. The main applications of the three
more resilient through remote inspections and technologies in transport centre around demand
multi-spectral imagery, with drones providing an forecasting and traffic optimisation resulting in
interoperable platform capable of more frequent better traffic management, asset management,
and precise measurements. Furthermore, drones travel planning and operation of autonomous
have several potential applications in logistics. vehicles (AVs). A more detailed explanation of the
Transporting vital goods through the air has been four applications of big data, IoT and AI, supported by
a staple of international commerce for decades, case studies, is presented in Section 4.
but a revolution is taking place at low altitudes, on
demand, for last-mile connectivity (World Economic Drone operations are also inextricably linked with IoT
Forum, 2018). and big data. They can act as data collection devices
and perform tasks that are remotely controlled by
humans, using IoT. The one application of drones
that is transforming the transport sector is in
logistics and deliveries, which Section 4 discusses
in more detail, with supporting case studies.
14 August 2019 Disruptive technology and innovation in transportTable 4. Drones – applications, barriers and opportunities Applications Barriers Opportunities • Asset inspections and maintenance • Regulatory concerns • Improved traffic management (tunnels, bridges) • Safety • Cost savings and/or increased • Infrastructure maintenance • Security efficiency • Design process (provision of • Privacy • Creation of new jobs geospatial data) – integration with • Anonymity and traceability • Increased safety (engineering building information modelling (BIM) • Misuse (for example, terrorism, inspections) • Construction site monitoring drug smuggling) • Improved resilience of infrastructure • Enhancing construction site safety • Insurance implications • Enhancing data processing and • Traffic monitoring • Impact of automation (job losses) accessibility • Logistics (deliveries) • Aviation risk (potential for collisions • Supports BIM • Warehousing and inventory with other aircrafts) maintenance • Remote delivery • Disaster response Components Enablers • Predator drone (military) • Automatous drones • VTOL drone (vertical take-off and • Battery technologies landing) • Logistic configurations • Global navigational satellite systems • Internet of things (GPS and GLONASS) • 3D modelling • Flight controller (central brain • Augmented and virtual reality (AV/VR) of the drone) • Video editing software • Inertial measurement unit (IMU) • Electronic speed controllers (ESC) • Ground station controller (GSC) – smartphone app • Internal compass • First-person view (FPV) video transmission technology • High-performance cameras • Multispectral, LIDAR, photogrammetry, low-light night vision and thermal sensors Policy paper on sustainable infrastructure August 2019 15
3. Applications in transport
As Section 2 shows, the four disruptive technologies 3.1. Traffic management using intelligent
(IoT, big data, AI and drones) have several transport systems
applications in transport. This chapter discusses
the technologies and applications that have the Overview
greatest potential to fundamentally change the
way traffic flow is organised and managed, and ITSs are an emerging field driven by digital
as a result to bring the most significant economic technologies, aimed at improving the efficiency,
and social benefits to the EBRD regions. These safety and environmental performance of road
applications have been identified as areas that transport. An ITS enables vehicles to interact directly
bring the four technologies together, to showcase with each other and with the surrounding road
how they can enable and complement each other to infrastructure. It typically involves communication
provide optimum solutions for the transport sector. between vehicles (vehicle-to-vehicle, V2V), between
The four applications of the disruptive technologies vehicles and infrastructure (vehicle-to-infrastructure,
reviewed in detail in Section 3 are as follows: V2I) and/or infrastructure-to-infrastructure (I2I)
and between vehicles and pedestrians or cyclists
• Traffic management using intelligent transport (vehicle-to-everything, V2X).
systems (ITSs) – using new technologies to
predict future traffic demand more accurately Big data can be collected using a variety of
and optimise road networks accordingly, techniques and is already used all over the world
providing a wide range of social and economic to combat congestion, including through the use of
benefits, including reduced congestion and inductive loop detection (insulated cables embedded
pollution, improved safety and travel experience in the streets), video analysis and infrared sensors
for all road users. (detecting the heat emitted by objects), GPS and
social media. ITSs have been developed since
• Personal travel planning and public transport the beginning of the 1970s, however the recent
– using the available information on travel widespread emergence of big data has allowed the
demand and travel patterns of the population development of new applications for the transport
to facilitate optimisation of planning, programming sector. Over the past decade, there have been
and operations of public transport systems, remarkable new developments in technologies that
as well as improving the public’s personal facilitate ITSs; however, these are far from being
journey planning. used to their full potential, as this section will detail.
• Autonomous and connected vehicles for Big data is a disruptive technological change,
enhanced mobility – developing applications following cloud computing and the internet of things,
for AVs that will contribute to increased safety, which enable large data volume and large data type,
better user experience, economic savings with high commercial value to be processed at
and reductions in congestion by facilitating car a lower cost and higher speed. In the transport
sharing and mobility as a service (MaaS). sector and ITSs, all traffic monitoring, data treatment
and applications can be done at a much lower cost
• Unmanned aerial vehicles and drones and higher frequency. Big data analytics can improve
for monitoring – using the technology to the ITSs’ operational efficiency. Many subsystems
revolutionise asset management and inspections in ITSs that need to handle large amounts of data
(bridges, tunnels and construction sites) and to give information or provide traffic management
delivery of packages (logistics). decisions will be less expensive to operate. Through
fast data collection and analysis of massive amounts
of current and historical traffic data, traffic
management departments will be able predict traffic
flow in real time. Public transport big-data analytics
16 August 2019 Disruptive technology and innovation in transportcan help management departments to learn limited dissemination of traffic information, and
journey patterns in the transport network, a lack of experience in using advanced intelligent
which can be used for better public transport data analysis methods to provide real-time and
service planning. accurate traffic information to travellers and to traffic
management departments to deal with unexpected
Enablers, barriers and opportunities events and illegal traffic behaviour. AI and other
technologies of big data analysis, together with
Technical analysis increased computing power and storage capacity,
bring new opportunities for the development of ITSs.
The use of big data, IoT and AI in traffic management
systems allows for more detailed and accurate The ITS big data analysis cloud platform consists of
predictions in relation to future traffic demand, a basic service layer (data collection), data analysis
traffic flow and any unexpected events or incidents layer (data integration) and terminal publishing layer
on the road network. Having this wealth of real-time (data release). The basic service layer is the basis
information allows for more efficient optimisation for data analysis and its main purpose is to use
of traffic signals and reducing congestion, improving cloud computing technology to integrate data from
traffic safety and preventing or reducing damage different systems (IoT). The data analysis layer’s
to the infrastructure, as well as reducing traffic function is to process the data in real time, using
emissions, enhancing mobility, increasing service advanced analytics and potentially AI to then help
reliability and supporting economic development. the decision-making process by providing trend
predictions and forecasting. The main function of
Despite significant advances in the use of big the terminal distribution layer is to communicate the
data in traffic management, there are still some available information by releasing it in real time via
problems. These difficulties include a lack of cloud services and smart devices (mobile phones,
integration of traffic data, low utilisation rate, PCs) (Hu et al., 2017), as Figure 2 shows.
Figure 2. An ITS in real time
TRAFFIC INFORMATION TRAFFIC INFORMATION
COLLECTION PUBLICATION
www
Video detection Coil detection INTELLIGENT TRANSPORTATION Internet Smart mobile phone
DATA CLOUD PLATFORM
Radar detection Floating car
On-board Variable
equipment message board
TRAFFIC
SCENE MANAGEMENT
TRAFFIC INFORMATION
MANAGEMENT
Signal control Electronic police
HD bayonet Overspeed snap Platform (collection, integration, release)
Video surveillance Mobile policing
Source: Market analysis (Support study for Impact Assessment of Cooperative Intelligent Transport Systems, European Commission (2016) and Hu et al. (2017)).
Policy paper on sustainable infrastructure August 2019 17The uptake of ITSs has been uneven across Europe. Cost-benefit analysis
In 2016, the C-Roads Platform was formed to
provide a single point of contact for cooperation The claims about the benefits of traffic management
between the automotive industry manufacturers and are varied but, in general, high. A typical set of claims
the European Union (EU) member states. Initially, is provided in a summary by Transforming Transport,
eight member states were included, which increased an EU-funded project of a consortium of 48 leading
to 16 as of October 2017. The C-Roads Platform transport, logistics and information technology
aims to facilitate harmonised and interoperable ITS stakeholders in Europe (22). They highlight the
deployment across the EU, encouraging cooperation following points:
and harmonisation between the projects:
• A 10 per cent efficiency improvement can lead
• C-Roads InterCor (2016-19), Belgium, France, the to cost savings of €100 billion from big data,
Netherlands, the United Kingdom. The project as fast data collection and analysis of current
links European ITS initiatives with the aim of and historical traffic data has helped traffic
creating a continuous ITS network that can serve management departments with operational
as a testbed for ITS services deployment and efficiency.
development.
• Improvements in operational efficiency
• NordicWay (2015-17), Finland, Denmark, Norway, empowered by big data are expected to lead to
Sweden. The pre-deployment pilot project aiming US$ 500 billion savings worldwide in terms of time
to test interoperable cellular communication for and fuel, as well as savings of 280 megatonnes of
ITS services, enabled through roaming between CO2 emissions.
different mobile networks and cross-border
services. • The McKinsey Global Institute concluded that in
2013, US$ 400 billion a year globally could be
• C-The Difference (2016-18), France, the saved by “making more of existing infrastructure”
Netherlands. The partners involved in this pilot through improved demand management and
project have been working on bringing ITS services maintenance.
to the market for the past 10 years, investing
significantly in the development and deployment • “The Internet of Things has changed the business
of ITS. performance of many organisations and is
predicted to cut the emissions from trucks in the
• C-Roads (2016-20) Austria, Belgium, the Czech US by 25 per cent” (21).
Republic, France, Germany, Slovenia. The project
outlines the rationale and objectives for ITS In a specific study in the Netherlands, the social
development, including coordinated deployment costs of congestion on the main road network
across borders. in 2015 (2010 prices) are estimated to be
between €2.3 and €3 billion annually. Traffic
management systems (ITSs) contribute to a 9 per
cent improvement in overall travel time, which is
worth €210-€272 million (on a pro rata basis). The
associated system costs are €164 million, giving
a benefit-cost ratio of between 1.3 and 1.7.
18 August 2019 Disruptive technology and innovation in transportIn a more recent study by the European Commission The study estimates the benefits of 25 individual
(20), annual benefits of approximately €15 billion types of improvement that are possible with this
in 2030 are compared with costs of €2.5 billion, equipment, with some providing more than one type
giving a ratio of benefits to cost of 6:1. This further of benefit. Of the total number, 15 provide safety-
corresponds with other studies, which have estimated related benefits, 10 provide efficiency benefits such
benefit-cost ratios in the range of 1.5 to 6.8 shown as savings in travel time, 8 relate to managing traffic
in the EasyWay and DRIVE C2X studies (20). operation such as smoothing the patterns of traffic
flow and 8 provide environmental benefits. Each
This study bases the estimates of costs on the independent improvement is relatively simple, with
installation of four types of hardware and software: examples including “warning of slow or stationary
vehicle”, in-vehicle signage and speed limits,
• In-vehicle information sub-system fitted either information on weather conditions, signal violation
by the vehicle manufacturer or retrofitted at intersections, off-street parking and park-and-ride
to the vehicle and attached to the vehicle information, zone access control for urban areas and
communication buses. protection for vulnerable road users.
• Personal sub-systems such as mobile phones,
tablets, personal navigation satnav-type devices, Over 86 per cent of the costs relate to the hardware
and other handheld devices not attached to the required within vehicles and a further 10 per cent
vehicle’s information bus. to “aftermarket” devices. Three elements make up
• Roadside ITS sub-systems such as beacons and approximately 99 per cent of benefits estimated,
smart traffic lights. with reduced travel times and increased efficiency
• Central ITS sub-systems, which may be part of a accounting for 66 per cent of total benefits, reduced
centralised traffic management system. accident rates for 22 per cent and fuel consumption
savings 11 per cent. As with all the disruptive effects
These provide a range of communication and assessed here, the assumptions on uptake are of
control services (V2V or V2I) but do not provide great importance, as essentially the same hardware
the more advanced capability for autonomous and enables all 25 services. Where services are limited
connected vehicles discussed below. While the by regulatory and other constraints, benefits fall
costs for autonomous driving are estimated as commensurately.
eventually falling to between US$ 1,000 and US$
1,500 per heavy goods vehicle (HGV) per year (see This study primarily addresses what can be seen
below), the upfront costs to consumers of these four as marginal changes from enhanced driver aids
simpler systems for new vehicles are estimated at to adapting existing systems. The advantage of
approximately €275 per car and €315 per HGV (with the approach is that the changes assessed are
ongoing annual costs of €20 and €30 respectively). specific and realistic and occur within a relatively
By 2030, these are estimated as potentially falling to short timeframe (by 2030). The level of costs and
€180 per car and €200 per HGV. benefits are less than an order of magnitude smaller
than those from autonomous vehicles and can be
considered indicative of the benefits of “low hanging
fruit” on a first stage towards the fuller use of
automation and information systems. Nevertheless,
they show that high benefit-cost ratios approaching
six are plausible even when based on marginal
changes to existing systems enabled by a disruptive
technology.
Policy paper on sustainable infrastructure August 2019 19You can also read
