Self DrivingVehiclesReport

Self-Driving Vehicles (SDVs)
& Geo-Information
REPORT | SELF-DRIVING VEHICLES (SDVS) & GEO-INFORMATION
CONTENTS
I. PREFACE
Background and Scope4
II. SELF-DRIVING VEHICLES 6
Introduction
Direction and Trends
Current Scenario: Innovations in the Industry
III. SELF-DRIVING VEHICLES ECOSYSTEM 12
Action Points for the Ecosystem
IV. GEOSPATIAL AND SELF-DRIVING VEHICLES 17
Autonomous Vehicles as Data Producers
Geospatial Technologies - ‘Under the Bonnet’
Space Technology
Geospatial Infrastructure
Geographic Information: Data
Authenticity and Reliability
Standards and Structure
V. GOVERNMENT AND SELF-DRIVING VEHICLES 25
Establishing an Enabling and Protective Legal Framework
Establishing of Open, or Interoperable and Internationally Oriented Data Policy

and Governance
CONCLUSION
ANNEXURE28
2
List of Figures & Tables

Figure 1.1 Where in the World are Self-Driving Vehicles (SDVs) 5
Figure 1.2 Autonomous Vehicle Deployment Timeline 6
Figure 1.3 Connected and Autonomous Vehicle Technology Road Map 7
Figure 1.4 Key Forecasts 9
Figure 1.5 Economic Impacts of Autonomous Vehicles in United States and United Kingdom 10
Figure 2.1 Actions Points for the Autonomous Vehicle Ecosystem 14-15
Figure 3.1 Autonomous Vehicles as Data Producers 17
Figure 3.2 Sensor Integration in Autonomous Vehicles 18
Figure 3.3 Under the Bonnet – How the Self-Driving Car works 19
Figure 3.4 Geographic Information (Data Network) 22
Figure 3.5 Trends on Autonomous Vehicles 23
Figure 4.1 Establishment of an Enabling and Protective Legal Framework 25
Table 1.1 Current Scenario: Innovations in the Industry 8-9

Table 1.2 Industry Directions 11
Abbreviations
ADAS Advanced Driver Assistance System OEMs Original Equipment Manufacturers
AMS Amsterdam Institute for Advanced Metropolitan OGC Open Geospatial Consortium
Solutions RTK Real Time Kinematics
AV Autonomous Vehicles SDVs Self-Driving Vehicles
ERTICO European Intelligent Transport Organization UK United Kingdom
GNSS Global Navigation Satellite System UN United Nations
GPS Global Positioning Software US United States
HD High Definition V2V Vehicle to Vehicle
IT Information Technology V2I Vehicle to Infrastructure
LiDAR Laser Illuminating Detection and Ranging V2N Vehicle to Network
MIT Massachusetts Institute of Technology V2P Vehicle to Pedestrian
NHTSA National Highway Safety Administration
3
PREFACE
Transportation systems globally are on the verge of a major transformation. An innovation that is expected to disrupt
the entire automobile industry, the driverless revolution has just begun. With the onset of sophisticated technological
advances which combine artificial intelligence and robotics capabilities, interest in ‘Autonomous vehicles’, ‘Self-
driving vehicles’ or ‘driverless’ cars has been surging. Critical to the growth of the ‘Self-driving vehicles’ industry is
the geospatial industry. Equipped with satellite navigation, sensor compliments and other positioning technologies,
autonomous vehicles along with geospatial industry brings in a paradigm shift in the driving experience while improving
safety and efficiency simultaneously.
The data information required by self-driving vehicles needs to be rich, diverse, and accurate especially concerning
spatial precision and frequency of updates than what is currently in use for day-to-day navigation. Geospatial content
providers, therefore, will have a crucial role to play to drive an entirely new business model as compared to the
conventional ones. These cars will not only ‘need’ geospatial content to build a thriving eco-system but will also
‘provide’ content for further use.
There is an extreme reliance of the autonomous vehicles on geospatial content. Understanding that maps are going
to be of fundamental nature when it comes to autonomous vehicles, the density of information needed for driverless
cars is going to be much higher and harder to collect. A whole array of sophisticated content providing geospatial
technologies will exist simultaneously for the successful navigation of self–driving vehicles.
BACKGROUND & SCOPE

Compared to 10 years ago, today, the idea of self-driving vehicles has started to become less far-fetched. Major
automakers plan to get their autonomous vehicles on the road by 2025 which is less than a decade away. The questions
are many and the answers too little. In this report, we take an in-depth study to understand the crucial role the
automobile industry, geospatial industry and the government together play in bringing evolution in the autonomous
vehicles industry. The report has been produced as a knowledge initiative to answer the ‘mystery’ questions around
the role of geospatial information in self-driving vehicles, the role of the autonomous vehicle ecosystem and the
government,
The study, therefore, is largely directed at all the stakeholders of the driverless car ecosystem and the reader would
gain insights into:
• The concept of self-driving vehicles
• Current market trends and benefits and costs associated with driverless cars
• Action points of the self-driving cars ecosystem
• Role of the geospatial industry
• The defining role of the government
The report, in summary, examines the core of the driverless car ecosystem i.e. the data and how the other defining factors
of the self-driving cars ecosystem form an inter-linkage to transform the complete landscape of the transportation
industry in the near future.
4
WHERE IN THE WORLD ARE SELF-DRIVING CARS?

WATERLOO, ON, CANADA
AUSTIN, TX, USA ROTTERDAM, THE NETHERLANDS
WAGENINGEN, THE NETHERLANDS

FORT COLLINS, CO, USA
BAERUM, NORWAY
KIRKLAND, WA, USA
HELSINKI, FINLAND
SAN FRANCISCO
CA, USA
MOUNTAIN VIEW
CA, USA
SION, SWITZERLAND
PHOENIX, AZ, USA
TOKYO, JAPAN
BOSTON, MA, USA
SINGAPORE
PITTSBURGH, PA, USA
SHANGHAI, CHINA
ANN ARBOR, MI, USA
Figure 1.1 Where in the World are Self-Driving Vehicles (SDVs)
5
SELF-DRIVING VEHICLES
In today’s context, most of the cars have been digitised to Level 0 : No-Automation
provide the driver with easier operation and better infor- At the ‘no-automation’ level, the driver is entirely re-
mation such as real-time traffic information, performance sponsible for the vehicle and its control system – brake,
data assessment like speed, music streaming from the steering, motive power, navigation – at all times.
cloud, etc. All in all, the cars today are a technological
marvel and there is still so much more to come. Trans- Level 1: Function Specific Automation
forming the auto industry will be tomorrow’s car that will Automation at the ‘function specific level’ involves auto-
be a step change from what is now in the offering. mation in few i.e. one or two specific control functions.
Automation at this level includes electronic stability control
Self-driving vehicles or autonomous vehicles have been or pre-charged brakes. In such a situation, the automation
a long held dream, a dream that has repeatedly failed to feature in vehicles assists with braking while the driver
materialise. ‘Cars that drive themselves’ are no longer regains control of the vehicle.
a work of science fiction. A very niche and specialised
market, the self-driving vehicle industry is making rapid Level 2: Combined Function Automation
strides in integrating a lot of technologies from different The Level 2, ‘combined automation function’, involves
eco-system to build the self-driving vehicle. automation of two primary control functions designed to
work together. An example at this level could be ena-
First of all, what are self-driving vehicles? According to the bling the adaptive cruise control in combination with lane
National Highway Safety Administration (NHTSA), self-driv- centring.
ing vehicles are vehicles that can drive themselves without
any human supervision or input to control the steering, Level 3: Limited Self Driving Automation
acceleration and braking. The above definition implies that In this level of automation, vehicles have full control
autonomous technologies imbibed in self-driving vehicles, of all safety critical functions under different traffic or
enable the car to go from Point A to Point B by performing environmental conditions and in those conditions to rely
all the required functions for a vehicle to move safely with- heavily on the vehicle to monitor the complete driving
out any human on board. Even though, the common belief process. The driver is also needed for occasional control
is that the driverless vehicles are a futuristic concept, the but with sufficiently comfortable transition time.
race to bring these vehicles to our roads has already begun.
These vehicles represent a disruption that is unprecedent- Level 4: Full Self Driving Automation
ed in both magnitude and scope. At the final level of automation, the vehicle is designed
to perform all functions of driving by itself. The car will
The self-driving car or the smart car achieves autonomy be able to perform all safety-critical driving function
at five levels. These levels are: and monitor roadway conditions for an entire trip. In
AUTONOMOUS VEHICLE DEPLOYMENT TIMELINE

5-10 YEARS 10-20 YEARS BEYOND 20 YEARS
 Controlled, AV-only environments  Less restricted environments  L arge, connected AV networks, allowing
 Moderate level of automated driving  High level of automated driving multiple mobility scenarios
 Low to medium speeds  Medium to high speed  On demand mobility and fleet services
 Customizable AVs
SOURCE: UN World Urbanization Prospects, World Business Council for Sustainable Development, Factiva, Navigant Research, EY analysis
European Commission, directorate General Information Society and Media, Informal document No.: ITS-13-07
Figure 1.2 Autonomous Vehicle Deployment Timeline
6
CONNECTED AND AUTONOMOUS VEHICLE TECHNOLOGY ROAD MAP
L0 2010
2015 L1
AUTONOMY
LEVEL
L2 2020
2025 L3
L4 2030
Figure 1.3 Connected and Autonomous Vehicle Technology Road Map
such a design the driver or in this case the passenger industry. While traditional automaker companies such as
will only have to input the destination in the car but will Audi, BMW, Toyota, etc., are stepping into the self-driving
not be needed to control the vehicle at any time during vehicle industry by introducing autonomous technologies
the trip. that offer driver support feature, technology giants like
Apple and Google are entering directly into the smart car
DIRECTIONS AND TRENDS market.
Self-driving vehicles, also coined as the ‘vehicle of the
future’ or ‘smart cars’, are already moving in a forward Bringing disruption in the traditional vehicle technology
direction by taking shape in a variety of forms. Bring- value chain is the gaining traction in the very technology
ing about the next ‘automotive revolution’ is not only that makes the driverless cars run. As the technology
the automobile manufacturers but also the technology innovation is accelerating, given that the fifth genera-
giants. At present, innovation on self-driving vehicles tion of wireless technology makes it possible to stream
is happening in the digital ecosystem, the geospatial data in real time on cloud, the quality of connectivity
ecosystem, the automobile ecosystem, etc. to support in between the vehicles has improved. The evolution
the transformation of the automobile industry. Innova- of multiple, complex low cost sensors, the computing
tion in all these ecosystems is going to bring about a speeds to work the artificial intelligence component to
new level of connectivity among vehicles, new kind of steer the self-driving vehicles, all come together to drive
cars, features like safety sensors, smartphone integra- the self-driving vehicles. Technology innovation compa-
tion, etc., to progress the autonomous vehicles industry. nies are investing in new technologies and new services
Driven by both traditional automobile companies as well along with the increasing investments being made by
as technology giants, there is a rapid acceleration in the automakers like Tesla, BMW and Audi who are pushing
innovation efforts taking place in the autonomous vehicle technology to the limit.
7
CURRENT SCENARIO: INNOVATIONS IN THE INDUSTRY
All Tesla vehicles have the hardware needed for full self-driving capa-
bility. 8 surround cameras provide 360 degree visibility around the car,
Tesla
12 updated ultrasonic sensors and forward facing radar with enhanced
processing.2
General Motors has started testing self-driving Bolt electric vehicles

General Motors on some public streets in California, Arizona and Michigan after buying
autonomous tech start-up Cruise Automation for $1 Billion3
Ford invests $1billion in an Artificial Intelligence start-up with the

Ford intention of having a fully autonomous, level-4 capable vehicle for the
commercial market by 2021. 4
Fiat Chrysler Automobiles (FCA) has partnered with Waymo (previous-

Fiat Chrysler ly the Google Self Driving Project) to carry out detailed research into
autonomous vehicles5
Honda has unveiled a new self-driving concept called the Cooperative

Mobility ecosystem that will ensure vehicle to vehicle communication in
Honda connected cars and smart city infrastructure, reducing traffic congestion
and improving road safety. The Cooperative mobility ecosystem aims to
“connect the power of artificial intelligence, robotics and big data”. 6
Volvo cars has cemented its position as the leader of automotive safety
Volvo innovation by scoring full six points in the Autonomous Emergency
Braking for Pedestrians (AEB) Pedestrian test procedure.
Toyota is partnering with Stanford University and MIT to research AI and

Toyota Robotics in order to bring greater autonomy to Toyota Cars by contribut-
ing US$50 million over 5 years. 7
Hyundai already has Level 2 smart sense technology, an advanced driver

assistance system that includes smart cruise control. The automobile
Hyundai
company is targeting to release highly autonomous vehicles by 2020 and
fully autonomous vehicles by 2030.8
BMW is working in collaboration with Intel and Mobileye to put a fleet of

BMW
40 autonomous test vehicles on the roads by the second half of 2017.9
Audi is adopting Nvidia’s drive computing platform to accelerate the

Audi introduction of next generation automated vehicles, for greater driving
safety and new mobility services. 10
Mercedes’s Drive Pilot is not only relying on its camera and sensors to
Mercedes steer the car but also gathers data from HERE maps to make the car
capable of making a turn by itself. 11
8
There is a fourfold improvement in Google’s Autonomous Cars project

– Waymo. Safety related disengagements has fallen from 0.8 times per
Google
1000 miles to just 0.2 per 1000 miles. The company has already in-
creased it total driving by 50% last year. 12
Nvidia is focusing on putting a lot of compute horsepower in the vehicle

Nvidia itself so that the vehicle can do the autonomous driving by itself without
needing to be connected to the network constantly.
Intel has launched their own solution for gathering, processing and
analysing data from autonomous vehicles and helps automakers like
Intel
BMW manage that data. For autonomous driving, Intel incorporates the
endpoint, connectivity and the data centre to offer end-to-end solutions. 13
Microsoft’s cloud to ingest huge volumes of sensor and usage data from
Microsoft connected vehicles to help use automobile companies to use the data
for self-driving 14
KEY FORECASTS
DRIVE ME FUTURE
PROJECT
Volvo’s Drive me Project expects to
TRUCK
Mercedes Benz intends to introduce its Future Truck
deliver 100 self-driving cars to by 2025, complete with a ‘highway pilot’ automated
customers by 2017 system through which a truck will be able to
communicate with nearby vehicles
$87 94.7 78% COST

REDUCTION
BILLION MILLION
According to industry estimates, 94.7 million vehicles with
By switching from car ownership models to a shared driverless
by 2020, the AV market will be self-driving capabilities to be sold
model, the costs of car ownership (based on US) could fall from
worth $87billion. annually around the world by 2035 US$0.70 per mile to around US$0.15 per mile – a 78% reduction
AV MARKET
11 MILLION (2024)
Car sharing to skyrocket from 1.5 million today to
11 million by 2024
Figure 1.4 Key Forecasts
9
Economic Impact of Connected and Autonomous Vehicles 2014-2030 in the United States
Autonomous cars total savings:

$158 billion $1.3 trillion $488 billion
Fuel savings Total saving from
accident avoidance
$11 billion
Fuel savings from
avoiding congestion $507 billion
Increased productivity
from autonomous cars
Source: Predictions for U.S.

Market, Morgan Stanley
Research, 2014
$138 billion
Increased productivity
from congestion
avoidance
Economic Impact of Connected and Autonomous Vehicles 2014-2030 in United Kingdom
£51billion
Value added annually
(at 2014 prices)
25,000
Jobs in automotive
manufacturing
320,000 created
Additional jobs
created
25,000
Serious accidents
prevented
Source: Connected and

Autonomous Vehicles, The UK
Economic Opportunity, KPMG 2,500
Analysis, 2015 Lives saved
Figure 1.5 Economic Impacts of Autonomous Vehicles in United States and United Kingdom
10
TABLE 1.2
Industry Directions
The ‘Roboat’, an autonomous boat/vehicle, the result of $27 million collaboration between the Massa-
Autonomous chusetts Institute of Technology (MIT) and the Amsterdam Institute for Advanced Metropolitan Solu-
Boat1 tions (AMS) will make its maiden voyage in 2017. Fitted with a battery of sensors, Roboats will provide
information on water and air quality while simultaneously providing an eco-friendly transport option
Uber’s Otto and Ford are reshaping the Trucking industry by exploring how driverless technology is
Autonomous going to be put in use by delivery vehicles. For example: In the first commercial use of the self-driv-
Trucks
ing truck, Otto, hauled Budweiser beer across the United States.
Autonomous Tesla, Uber, BMW, Audi, Google, etc., are working on developing their own autonomous Cars. Test
Cars drives have already begun for Hyundai, Uber, Google autonomous cars.
POTENTIAL BENEFITS OF AUTONOMOUS POTENTIAL RISKS AND COSTS OF

VEHICLES AUTONOMOUS VEHICLES
Self-driving vehicles are expected to bring significant benefits The costs and risks associated with autonomous driving
for user convenience, fuel savings and pollution reduction are by at large uncertain. The potential risks and costs of
benefits. Some of the benefits are listed below as: self-driving vehicles can be briefly listed down as:
 Beneficial to non-drivers, self-driving vehicles will improve  Autonomous vehicles require additional vehicle equip-
independent mobility thus reducing the need for chauf- ment, services and maintenance and roadway infra-
feurs. The other evident advantage is that smart cars may structure which means additional costs
also lead to subsidized rates at the public transit  The self-driving vehicles will introduce new risks
 Self-driving vehicles may reduce many common accident related to software failures, system failures. System
risks and therefore crash costs and insurance premiums failures can be fatal to vehicle occupants and other
 One of the main benefit that is foreseen is the abundant road users
use of smart cars that may allow platooning (vehicle  Driverless vehicles are prone to information hacking
groups travelling close together), narrower lanes, reduced which can be used by criminals and terrorists for cre-
intersection stops, and roadway costs ating traffic management. Similarly data sharing and
 Autonomous vehicles need not worry about parking as GPS data are liable to privacy concerns
they can drop off passengers and find a parking space if  The traditional cars will be driven simultaneously
needed on its own. This reduces total parking costs. with the self-driving vehicles which may have adverse
 As autonomous vehicles are going to be driven by technol- impacts on the convenience and safety of the other
ogy mostly, this will increase fuel efficiency which in turn modes of travel
will reduce pollution emissions  An on-going debate with respect to self-driving vehi-
 Could facilitate car sharing (vehicle rental services that cles is that as these cars are adopted, jobs for drivers
substitute for personal vehicle ownership), which can will decline. Simultaneously, because these cars will
provide significant benefits run on software, the need for auto-vehicle body shops
 Autonomous vehicles also lead to an improvement in land shall also decline causing loss of employment
use. Since self-driving vehicles would be able to drop pas-  Privacy is seen as one of the major concerns of the
sengers off, the parking spaces could be used to develop naysayers of the self-driving vehicles. Because a lot of
economic purpose. Therefore, autonomous vehicles could information has to be stored in the software, individ-
improve the available urban space by 15-20%, largely uals are concerned that the computer shall collect
through the elimination of parking spaces. personal data (without permission) which risks the
 The increased use of fully driverless vehicles will bring on privacy of the common citizen
a full driverless motorway which would allow much better  The cars may not operate at high levels of safety all the
utilisation of road space, reduction in energy consumption time. For instance, a heavy rainfall or a difficult weath-
and smoothing flows across segments er condition has the potential to damage the laser sen-
 The onset of self-driving vehicles will lead to creation of sors mounted on the cars roof. Also, for instance, if the
hundreds of thousands of additional jobs in manufacturing traffic signals fail then whether the cars will be able to
and production. Though the ‘new’ workforce that will exist interpret human signals or not is a big question.
will need to update their skills according to the market  The impact on the gasoline industry too is going to be
demands tremendous given that the newly improved self-driving
 Police officers could start focusing more on serious cars would be electric.
crimes instead of writing traffic tickets and handling car
accidents both of which shall reduce dramatically once
self-driving cars enter the market
11
SELF-DRIVING VEHICLES ECOSYSTEM

Defining the ecosystem of the autonomous vehicle industry (a) Boosting software reliability and cyber security
has always been difficult for the industry has only recently (b) Securing the software that runs the car and integrating
begun evolving. The concept of ‘smart cars’ was initially at a all the components that need connection to the internet
nascent stage and to define what industries belonged to the (c) Securing applications running within the vehicle
autonomous vehicles community was difficult. However as (d) Establishing communications between vehicles and
the penetration of autonomous vehicles is accelerating, the internet enabled devices
ecosystem is also getting established simultaneously.
Deep Learning and Artifical Intelligence: Deep Learning
The past year has highlighted that autonomous vehicle and Artificial Intelligence will play a vital role in imitating
industry is working in collaboration with multiple stakehold- the human neural networks. Deep Learning uses algo-
ers at one single time. Recent developments in technol- rithms to analyse the data and solve the problems that
ogy highlight the gaining momentum of the autonomous may come in the functioning of autonomous vehicles. The
vehicles industry. The past year has highlighted the active algorithms are more accurate at object recognitions than
role that technology giants, the automobile industry, the humans and these, in turn, will make the roads available
geospatial industry and the government play as part of the to fully autonomous vehicles for they allow detection &
autonomous vehicles ecosystem. As various milestones of recognition of multiple objects, improve perception, reduce
autonomy come into focus, there begins a growing interest power consumption and enable identification and prediction
of new stakeholders in the driverless car ecosystem. Insur- of actions. Artificial Intelligence and Deep Learning systems
ance provider, energy providers, mobile service providers will also play an important role in manufacturing production
etc., are few important stakeholders that have begun to play lines and crash tests by reducing time and cost.
a well-defined role in the self-driving vehicles ecosystem.
Original Equipment Manufacturers (OEMs): The wide-
ACTION POINTS FOR THE ECOSYSTEM spread acceptance of Self-driving vehicles is accelerating
The self-driving vehicles’ ecosystem is still getting defined the development of various new technologies such as
every day. As Original Equipment Manufacturers (OEMs), advanced sensors, GPS positioning, computer vision and
technology providers, the government and transport image recognition. Simultaneously there is the develop-
authorities come together, it is important to determine the ment of Robotics and drone technology as well. Henceforth,
action points for them. These action points define the role self-driving vehicles require a standard vehicle technical
and responsibilities of the ecosystem for the self-driving architecture and the stages of development for these are:
vehicles.
(a) Multiple Sensor Technology
Software Companies: Substantial opportunities are going to (b) Sensor Fusion
be available for a range of technology providers’ esp. soft- (c) Autonomous Engine
ware companies who are involved in application develop-
ment for example, cloud and IT services, security software OEMs, therefore, need to start investing and acquiring the
and vehicle engineering. For instance, there is an unspoken right technology expertise and also consolidate ventures
agreement amongst all the stakeholders that the data so which integrate hardware and software system. OEMs need
generated by the autonomous vehicles will use cloud as to also incorporate the Advanced Driver Assistance System
a medium for storage. Transportation agencies will then (ADAS) providers or tech providers for the efficient function-
be able to use this data, use software and analytical tools ing of the autonomous vehicles.
to understand the data and then implement the findings
in designing and developing ‘smart’ roads and an efficient Governments and Transport Service Providers: Autono-
transport system. mous vehicles have the capability to drive the development
of ‘intelligent’ cities. As established, these vehicles are
Software providers can provide the right framework for the always collecting location, road and traffic data which can
proper functioning of the autonomous vehicles. Getting the be used by government authorities to assess and analyse to
right software in place is important. World’s top computing assist in urban network planning. The data so derived can
companies like Apple, Google, Cisco, etc., are already inte- also be used to develop new road revenue models. Some
grating their innovative products with in-vehicle systems. cities in Europe and United States of America have already
The role of the technology and software providers is: started using the technology to change the way they charge
12
for road usage. Because road usage requires accurate data, and in turn stimulate the banks to create new lending
it is imperative to ensure that the smart cars collect the projects for autonomous vehicles. This futuristic trend
right data. For this purpose - governments, businesses and emphasises the role of banks and financial institutions
organizations, all need to come together. to develop key relationships with service industries such
as fleet providers and energy providers to drive up the
Municipal authorities and transport service providers revenue streams.
and operators can play an enhanced role by using smart
technology to caution drivers of any potential hazards and A radically reshaping of the insurance industry is also
crashes. Smart traffic signals could also be gaining traction in question due to the potential widespread adoption of
and could potentially save up to 15% in fuel consumption. self-driving vehicles. Warren Buffet predicts that the need
The transport authorities could also use the data collected for insurance coverage by accidents is going to be less
by self-driving vehicles to schedule self-driving vehicles at likely once fully automated cars are adopted. In such a
off-peak times to enhance productivity and traffic through- scenario, new insurance models are expected to arise.
out. In many ways, the data collected by the cars could, Insurers will need to adapt to the new technologies.
therefore, be used to mitigate traffic congestions by sched- Profound changes are also expected in how insurers
uling mail service providers, supermarkets and retailers so underwrite and sell insurance products. For example, the
as to improve delivery times. insurers could move from motor insurance to product
liability insurance. In the United States insurance compa-
Global interest in self-driving vehicles is going to speed reg- nies such as State Farm have already begun to examine
ulations up especially concerning smart roads, smart traffic and evaluate ways to understand the transformations that
signals, etc. Most countries have already aggressively start- self-driving vehicles could bring.
ed moving into the autonomous vehicle space. For instance,
China is especially fast forward in defining regulations and Mobile Service Providers: According to Garter by 2018, 20%
setting standards to convert cities to driverless hubs with of all new vehicles will need to be self-aware to capture
multi-year rollout plans. Many cities of United States of systems status, positioning and surroundings in real time.
America are also already looking into establishing policy This would lead to a significant increase in data consump-
frameworks at the municipal, state and federal level. tion pattern and therefore create potentially attractive
revenue models for mobile service providers. Mobile service
The government can also play a fundamental role in the providers will have to build new capacity, products and
ecosystem by improving its public infrastructure. The public services. This shall benefit the telecommunications industry
infrastructure which is important for self-driving vehicles and the Original Equipment Manufacturers (OEMs) to take
are roads, pedestrian tracks, bridges, traffic lights, etc. The advantage by selling their products to consumers.
governments should play a defining role in establishing the
need for infrastructure planning. Rough road conditions and Energy Companies: Self-driving vehicles are disruptive,
broken traffic signals could be detrimental to sensor-driven and they have the potential to disrupt the energy market as
cars – more than they would confuse a human driver. The well. As these vehicles become globally prevalent, energy
government could either take up on itself to transform the providers will soon have to start considering options of how
public infrastructure that will support the driverless cars in to profit from them. Energy companies have to decide on:
the long run or it could implement laws – laws that would
impose more liability on the cities. This is turn would make 1: Designing innovative service stations
cities more responsible and pressurized to maintain streets 2: In-road charging systems
and traffic signals. 3: Using alternate source of energy sources to generate
power
Finance Service Providers: Banking and financial ser-
vice providers need to play a defining role in autonomous The energy companies also need to start exploring options
vehicles to reshape ownership models and introduce new of partnering with car companies to understand the energy
revenue streams. Foreseeing how autonomy is going to requirements of the car companies.
disrupt the finance industry, banks and other financial
institutions will need to take the first step and boost their Roadway Contractors: The way to make self-driving vehi-
investments in machine to machine security. Simulta- cles safe is to make the infrastructure around the vehicles
neously, the transformation of the traditional vehicle safe. Modest changes need to be made to the infrastructure
ownership models could alter the nature of car financing so that autonomous vehicles can behave as predictably as
13
ACTIONS POINTS FOR THE AUTONOMOUS VEHICLE ECOSYSTEM
Cyber Security
 Safety and security assurance
Telematics and Infotainment

 Monitoring and managing
autonomous pods
 Collation and analysis of data
Energy Companies
 Designing innovative service stations
 In-road charging systems
 Alternative sources of energy sources
to generate power
Mobile Service Providers

 Build new capacity, products and
services
 Quality assurance
 Network application programming
interface
Finance Service Providers

 Reshape ownership models
 Introduction of new revenue streams
 Boost investments in machine to machine
security
 Radical re-shaping of the insurance industry
Transport Service Providers

 Transport providers to use smart technology
 Traffic scheduling
Figure 2.1 Actions Points for the
Autonomous Vehicle Ecosystem
14
Software Companies
 Boosting software reliability
 Securing the software that runs the car
 Securing applications running within the vehicle
Establishing vehicle to vehicle communication &

internet enabled devices
Deep Learning and Artificial Intelligence

 To imitate the human neural networks
 Using and analysing algorithms to analyse data
Original Equipment Manufacturers

 Integrate software and hardware
 Autonomous capability
 Investing and acquiring right technology
Geospatial Content Providers

 Continuous stream of location and
situational awareness data
 Provide high resolution - real, accurate
and authenticated data
Government
 Establishment of an enabling and protective legal framework
 Establishment of open, or interoperable and internationally
oriented data policy and governance
 Government plays a defining role by maintaining public
infrastructure such as roads, street lights, etc.
Roadway Contractors
 Road infrastructure should be such that it supports human as well as
robot drivers
 Developing smart roads that include special metal mesh that driverless
vehicle sensors can navigate through
15
possible. Currently, the system of roads is built with human can seek support from the local city government to network
drivers in mind. The contractors responsible for public roadside infrastructure to enhance the vehicle to infrastruc-
infrastructure such as roads and traffic lights, etc., need to ture communication such as traffic lights, cameras, road
start considering options for designing systems that are ef- sensors and parking meters so that they can communicate
ficient, safe and effective transportation system. It has been with each other and the dream of smart cars on intelligent
found that the vehicles ability to communicate with smart roads become a reality.
road infrastructure is much more important than it seems.
To develop the synergy of smart infrastructure the following Telematics & Infotainment: Telematics play a key role
have to be taken care of: in monitoring and managing a fleet of autonomous pods
• The road lines need to be painted every year with through collation & analysis of data and tracking malfunc-
reflexive paints so that the on-board sensors and tioning’s of the pods through software and hardware anal-
cameras catch on to the road lines. ysis. Telematics service provider may use a pay-per-use
• Developing smart roads line that include a special and premium subscription model to be sold on contractual
metal mesh that driverless vehicle sensors can navi- basis for maintenance, diagnostics, infotainment and con-
gate through tent streaming for autonomous transportation.
• Building roadside sensors along the streets and high-
ways so that cars can navigate them would allow the Cyber Security: As vehicles become connected and auto-
vehicles to ‘see’ activity far ahead on their routes mated, concerns are going to rise over cyber security. As
• Road infrastructure needs to be such that it sup- these cars will provide real-time accurate data – the fear
ports human as well as robot drivers such that roads that these systems can be hacked even by terrorists for
understand the technology used in the driverless massive damage is a major concern. Due to the recent suc-
vehicles cessful hacking of Tesla Model-S, the autonomous vehicle
• Building several traffic signals at different intersec- ecosystem is venturing out to find solutions to this problem.
tions while ensuring that roads have clear land mark- It is crucial that it be decided early on by the government,
ings such that driverless cars can navigate safely the technology providers and the automobile makers on
how to resolve the problem.
For instance, the United States National Highway Traffic
Safety Administration (NHTSA) establishes that yellow turn Now that we have established the action points for the
signals are significantly safer than red one because they are self-driving vehicles ecosystem, let us dive deep into the
easily recognizable. Similarly, well-maintained roads are core of what drives the self-driving cars - the geospatial
also critical for driving predictability. Also, the contractors component
16
GEOSPATIAL AND SELF-DRIVING VEHICLES

The self-driving vehicle industry has been growing at real-time data for efficient driving. However, there is one
an accelerating rate and the geospatial industry plays a other important aspect to the self-driving vehicles. They
crucial role for this emerging technology relies heavily not only consume data but they act as ‘super sensors’ or
on geographic data. The technology requirements of ‘on-the move data infrastructure’ producing vast amount
the self-driving vehicles industry is largely based on of data by themselves.
geospatial technologies such as sensors, LiDAR, Radar,
navigation, etc., the use of which can have a potentially Autonomous vehicle is expected to exponentially increase
disruptive impact on the geographic information industry. the amount of data being produced. For instance, the
The idea of self-driving vehicles on road is only becoming driverless car from Google is a true data creator pro-
a reality because of the advancements being made in po- ducing data on where to drive and how fast to drive. The
sitioning and sensor technologies. Thus, the whole array self-driving vehicles are capable of producing 0.75GB
of geospatial technologies is at the core of the self-driv- data per second which means they would create 2
ing vehicles. petabytes of data a year. It is foreseen that by 2020, an
average autonomous car may be able to process 4000 GB
AUTONOMOUS VEHICLES – AS DATA PRODUCERS or 4TB of data per day while the average internet user
When it comes to self-driving vehicles, data is the oil. shall process only 1.5 GB. This means that one single
Data is at the centre of the functioning of the autono- autonomous car, loaded with LiDAR, sensors, etc., and
mous vehicles and without data, no driverless car can shall be able to generate data approximately produced by
work. Automobiles are dependent on data and connectiv- 2666 internet users in a day. This also means that in 2020
ity and therefore data has the potential to change the way when there are 3 million autonomous cars worldwide –
driving is thought about. Autonomous cars are expect- automated driving shall be representative of the data of 3
ed to engage a plethora of sensors, consuming lot of billion people.
Autonomous car data vs. human data

In 2020, the average autonomous car may process 4,000 gigabytes of data per day,
while the average internet user will process 1.5 gigabytes.
THE COMING FLOOD OF DATA IN SONAR

AUTONOMOUS VECHICLES ~10-100 KB
PER SECOND
RADAR GPS
~10-100 KB ~50 KB
PER SECOND PER SECOND
CAMERAS LIDAR
~20-40 KB ~10-70 MB
PER SECOND PER SECOND
AUTONOMOUS VEHICLES
4000 GB
PER DAY... EACH DAY
1 autonomous car = 2,666 internet user
Source: Intel
Figure 3.1 Autonomous Vehicles as Data Producers
17
Analysing the trends, self-driving vehicles shall send and the on-board processor to apply the brakes or to move out
receive information which will inflate the amount of data of the way. Modern self-driving vehicles rely on both LiDAR
produced overall. This data may be location data, weath- and RADAR to validate the data that is generated on what is
er data, traffic data, navigation data, etc. Therefore, the seen and how motion is predicted.
data captured by the autonomous vehicle can be used for
actionable items to derive full value. As it is understood, LiDAR is currently seen as a premium option as Original
the driving force behind the technology is Big Data- au- Equipment Manufacturers (OEMs) figure out the cost
tonomous vehicles are actually an extension of the same structure of their self-driving vehicles. RADAR, on the
leading to almost unlimited data production. Countless other hand is a proven technology becoming more efficient
sensors and cameras mounted on top of autonomous for autonomous cars. At present, the technology is being
vehicles map the surroundings – spot lane markings, introduced to fit smaller, low power and efficient sensors
roadway edges, traffic signs and lights and identify which are suitable for OEMs cost reduction strategy
pedestrians. The driverless vehicles are also fitted with
other high tech gear such as altimeters and accelerom- High-Powered Cameras
eters to give more accurate positioning than would be All autonomous cars utilize high-powered cameras,
possible on using GNSS. All this data that is produced by but the actual camera technology and setup on each
the autonomous vehicle during the drive could be stored self-driving car varies. Some vehicles may have a single
in a centralized infrastructure from where it can be dis- camera embedded in the windshield while others might
seminated to the other stakeholders of data users after require several cameras mounted to the vehicle’s exteri-
thorough analysis and processing. For instance, the data or to give a composite picture of the surrounding world.
collected by self-driving vehicles may have to be shared A master of classification and texture interpretation,
with municipalities and the government. Road conditions, cameras are the cheapest and most available sensor.
potholes, traffic light situations, etc., can be dealt with They use and create the maximum amount of data thus
on an immediate basis because of the real-time data that making processing a computationally and algorithmically
the sensors in the car generate. In such a scenario, the complex task. Because cameras can sense colour, they
self-driving vehicles may also be looked at as an impor- are best useful for scene representation.
tant stakeholder of the ‘technology infrastructure’ of the
geospatial industry. The data, thus, created shall be a GPS (Global Positioning Software)
powerful tool for touching the public sector goals. Self-driving cars need a unique and detailed mapping sys-
tem, and Global Positioning Software (GPS) is a vital part of
Below are the important geospatial technologies incor- the autonomous car’s large scale navigation. GPS software
porated inside the car that are primary to the success of
the driverless cars and establish autonomous vehicles as
data creators.
GEOSPATIAL TECHNOLOGIES:
‘UNDER THE BONNET’ GNSS/GPS
Advanced Multiple Sensors
The most important part of self-driving vehicles is geodata
and this data about the environment is gathered by using
multiple sensors. These sensors are being used for map-
ping, localization and for avoiding obstacles. The main sen-
sor used to gather geo-data and information is the LiDAR Lidar/Radar PROCESSOR Ultrasonic
Sensors
– Laser Illuminating Detection and Ranging, with ranges up
to 100 meters. The LiDAR is used to build 3D maps and to
allow the car to foresee any potential hazard by bouncing
laser beam of surfaces surrounding the car to determine
the distance and profile of the subject accurately. While Camaras
LiDAR is being used to accurately map the surroundings, it
is RADAR that is used to map and monitor the speed of the
surrounding vehicles to avoid potential accidents, detours,
traffic delays and any other obstacles by sending a signal to Figure 3.2 Sensor Integration in Autonomous Vehicles
18
Figure 3.3 Under the Bonnet – How the Self-Driving Car works (The Economist)
is considered to be important because it defines the ‘mis- self-driving cars for drifting around tight corners and are
sion’ of the autonomous vehicles by helping the car navi- used extensively by parking aid systems.
gate successfully from the starting point to the end point of
the drive by looking at all road and by choosing the path. A Processors
fully autonomous vehicle requires an accurate localization All self-driving vehicles will have increased computing
solution which can only be provided through high-precision requirements for the data generated from the cameras,
Global Navigation Satellite System (GNSS) technology or and the sensors will need to be processed in real –time.
GPS. This technology ensures that accurate and reliable The processor is also necessary to model the behaviour-
data is available to ensure for a vehicle to stay in its lane or al dynamics of other drivers, pedestrians and objects
at least at a safe distance from other vehicles. around the vehicle. It is important to note that actions
such as steering, accelerating and hitting the brakes are
Ultrasonic Sensors all controlled by the processed information.
Ultrasonic sensors provide 360-degree visibility around
the car for up to 250 meters of range. They allow for All the aforementioned technologies mentioned above
detection of both hard and soft objects very close to the work in conjunction with each other to successfully and
vehicles. Ultrasonic sensors are also beneficial to the safely manoeuvre the vehicle to its final destination.
19
SPACE TECHNOLOGY public mobility. GNSS infrastructure is needed to convert

Space technology is defined as the technology devel- the normal GPS signals into highly accurate navigation
oped by space science or the aerospace industry for use devices. Additionally, GNSS infrastructure can also be
in spaceflight, satellites, or space exploration. Just like used to create highly precise reference frames for the
other technology domain, space technology too has a vehicles by ensuring that high accuracy maps are tied
vital role to play in self-driving vehicles and is gaining down to real ground points. This helps the driverless
importance slowly. Space technology can tackle the big- cars to navigate by themselves. This also ensures that
gest problem that autonomous vehicles continue to face each vehicle on the road, irrespective of whether they
that is deciding the right moment to hand over control are driverless or not, is in a universal frame of reference
to humans. At present, Renault-Nissan, committed to irrespective of who the automaker is and what is the
self-driving vehicles is deploying the Seamless Autono- sensor portfolio in use. GNSS infrastructure, therefore,
mous Mobility system which uses artificial intelligence is imperative to the success of the autonomous vehicle
and is derived from NASA technology. Collaborating for it efficiently combines location information to actual
with NASA’s Ames Research centre, the Nissan LEAF physical objects and a corresponding digital map i.e. to a
vehicle employs cameras, sensors and cellular data single reference frame.
networking and robotics software developed for Ames’
K-10 and K-REX planetary rovers to operate autono- For instance, WEpod, a completely automated vehicle
mously demonstrating the transfer of space technology currently in a testing phase in the Netherlands is utilising
to the autonomous vehicle industry. a combination of robust GNSS infrastructure, digital maps,
radars, cameras, laser scanners and ultrasonic sensors.
GEOSPATIAL INFRASTRUCTURE The localisation system used multi-constellation GNSS
The food for self-driving vehicles is data. As is evident, to with network-based-real-time kinematics (RTK). Similarly
prepare food one needs a well established and an operational the InDrive Project is developing a close to market solution
kitchen - similarly, to make use of available data, there is need that relies heavily on accurate and high-integrity satellite
of geospatial infrastructure unique to autonomous vehicles. navigation based on European GNSS.
Self-driving vehicles need a unique and detailed mapping
system so that data is accessible to all stakeholders of the Data Infrastructure
ecosystem as easily as possible. In such a scenario, geospatial Autonomous vehicles will be used as geospatial data col-
infrastructure is of paramount importance for smart mobility. lecting machines and, therefore, not surprisingly, these
vehicles will be collecting a lot of data of their surround-
GNSS Infrastructure ings. Autonomous vehicles need to be aware of what is
Information technology is at the heart of the driverless happening around them i.e. in their environment at all
vehicle ecosystem. GNSS infrastructure is at the core times, and that is why localisation of data is a fundamen-
of localisation of data and positioning of self-driving tal concept. It is, however, important to consider that the
vehicles. data collected by these vehicles is only useful if it can be
critically dissected in real time and processed accurate-
Global Navigation Satellite System (GNSS) is a constel- ly. Real-time data as gathered by autonomous vehicles
lation of satellites that provide autonomous geo-spa- could range from traffic patterns to delivery tracking, etc.
tial positioning with global coverage. They allow small This data is continuously gathered and will need a central
electronic receivers to determine location in terms of repository for storage and further distribution. More so,
longitude, latitude and elevation. The advantage to having the geographic information so collected will need to be
a GNSS system is accuracy, redundancy and availability processed to create accurate roadway maps which would
at all times. An integrated GNSS system is complimen- too require an efficient infrastructure. Simultaneously, a
tary and interoperable with other automotive technolo- robust data transfer pipe shall be the need of the hour
gies. By integrating sensor data and connectivity based to ensure that data that is collected and processed flows
information operators may reduce the need for expensive seamlessly. Therefore, the geospatial intelligence of
sensors and save money on infrastructure requirements. driverless cars shall rely heavily on a well-established
Therefore, GNSS is viewed as a fundamental enabling geospatial data infrastructure.
technology for autonomous vehicles – being at core of
localisation of data and positioning. Challenges in Establishing Geospatial Infrastructure
While geospatial infrastructure is imperative to the
An available and reliable GNSS is required to increase success of self-driving vehicles, it comes with numerous
the safety, enhance the traffic flow and to provide better challenges. The biggest problem that is foreseen is that it
20
is not going to be easy to map every terrain and link it to generated from in-car sensors such as LiDAR and Radar
the same reference frame which could take a lot of time is processed in software using algorithms to further
and involve high costs. make decisions on steering, braking, speed and route
guidance.
Also, there is a lack of uniform roll out of cellular tech-
nology which acts as the biggest roadblock for estab- For example, the self-driving car from Google is already
lishing an efficient geospatial infrastructure. While the a data creator in the truest sense. All the sensors in the
United States and Europe are steadily moving towards car enable the car to drive without a driver generating
the newer 5G standards, the rest of the world is still nearly 1GB data every single second. If one driverless car
stuck in the older standards of 2G and 3G. Since most of collects approximately 1GB data (or more!) every single
the geospatial data generated are from data-heavy sen- second, it can be only left to the imagination how much
sors, the need to have good cellular or Wi-Fi technology data can these in-car sensors generate over a period of
globally is paramount for conformity of self-driving time and with an influx of many such cars.
vehicles.
Base Map Data for Navigation
GEOGRAPHIC INFORMATION: DATA The base map is at the centre of the self-driving vehicles
Data is at the core of the success of the self-driving for efficient and easy navigation for even though sensors
vehicles (SDVs) industry. The on-going debate at pres- on the car detect things in real time, prior information
ent focusses on data and the legal quandaries that is necessary to evaluate what exits. High precision base
one may face with respect to data. It is imperative maps are being made by leveraging aerial imagery,
to realise that the geodata that is crucial to location sensors, mobile driven LiDAR, Aerial LiDAR data specif-
information is going to an important aspect of the au- ically for self-driving vehicle models and markets. The
tonomous vehicle technology. Location information is base map cannot be static and needs to be updated every
critical because it shall determine the potential of the second. Metre resolution maps are good enough for GPS
self-driving vehicles. Geodata, at present, is available navigation, but the accuracy of base maps need to be
for the surrounding environment and the more the defined in the absolute ranges of ‘centimetre’ or even ‘mil-
data is available, the more ‘automated’ and conven- limetre’ for the SDVs. It is these high-precision maps that
ient the decisions would become for these self-op- can smoothen the transition to a new technology. A highly
erating vehicles. Self-driving vehicles use geo-data accurate base map is needed to assert the stationary
to analyse their own location. The cars are fitted with physical assets related to roadways such as road lanes,
sensors to monitor positional awareness, proximity to road edges, dividers, traffic signals, poles and all critical
pedestrians, traffic signals and much more. The data data required for safe navigation of road by the self-driving
is, therefore, collected at four levels which are: vehicles (SDVs).
ii In-car sensor data
ii Base map data for navigation For example, HERE is using satellite and aerial imagery
ii Vehicle to Vehicle (V2V) and Vehicle to Infrastructure for base HD maps. These maps incorporate data from GPS
(V2I) devices and from sensor systems that have been outfit-
ii Commercial datasets (also includes social networking ted in the cars to collect datasets to make the base map.
sites) HERE has driven 1.2 miles in 30 countries on 6 continents
in the last 15 months.
In a way, it is Big Data that actually controls self-driving
vehicles. Let’s see how! Connected Data – Vehicle to Vehicle (V2V) and Vehicle to
Infrastructure (V2I) Communication
In-car Sensor Data The concept of V2I and V2V can be defined as the initia-
According to INVENT – a company dedicated to develop- tion of the term ‘Internet of Cars’, a play off on the term
ing inter-vehicular networking, computing, and sensing ‘internet of things’. In the connected car discussion,
technologies, autonomous vehicles or ‘smart cars’ of Vehicle to Vehicle (V2V) speaks of all the communication
the future is an extensive data collection system of the happening between the vehicles on the road. V2V is still
environment. These vehicles have embedded cameras, at a very nascent stage but could be the beginning of
sensors, computers (processors), short range wireless road safety. In this one vehicle can communicate with
network interfaces and GPS receivers that can sync with another directly which would allow cars to maintain an
a vast network that is continuously collecting data about internal map of the surrounding vehicles. The idea is to
the surroundings and that too in real-time. The data employ a small radio transmitter and receiver on each
21
V
V2
V2
Vehicle
P
V2N
V2V Vehicle - to - Vehicle
Vehicle Pedestrian
V2P Vehicle - to - Pedestrian
V2N Vehicle - to - Network
V2N
V2I Vehicle - to - Infrastructure
Network
Infrastructure
V2
I
Figure 3.4 Geographic Information (Data Network)
vehicle that broadcasts information about location, speed data such as geographic locations, housing, and demo-
and direction to other vehicles within several yards. This graphics to give a 360-degree view to the customers.
will help provide driver warnings to guide the driver
about when it is safe to change lanes, speed and merge AUTHENTICITY AND RELIABILITY
thereby helping electronic safety systems work safely. Self-driving vehicles or now commonly known as intelli-
gent vehicles are a complex system in which data plays a
Once V2V is successfully established the next step would crucial role. Data in such a case needs to be reliable for
be to develop the Vehicle to Infrastructure (V2I). The idea a safe and secure autonomous vehicle that can by itself
behind V2I is an integrated data network between the vehi- analyse complex environment in real time. Passenger ve-
cle and the roadside infrastructure such as traffic signals, hicles are complex issue in general and the involvement
roadway sensors, pedestrians (V2P), etc. It is predicted of data to majorly drive autonomy can be a challenge.
that the first V2I systems will be developed and employed The geographic information so required needs to be of
by 2020. high definition, industrial class reliability, authenticat-
ed by the supreme authority and true to real time. The
Commercial Data Sets whole technology is that is why based on data-centric
At present, many private companies are collecting connectivity.
data for use in self-driving vehicles. Social networking
sites such as Facebook, Twitter, etc., also collect data Data is already central to a thousand of applications such
pertaining to ‘location’, traffic, etc., which can be made as defence, power, medical, communications, construc-
available for the real time functioning of the autono- tion, transportation, etc. Therefore, data, a strategic
mous vehicles. asset integrates all the other components of the auton-
omous vehicles and helps them to operate better. It is
An interesting data source for autonomous vehicles is understood that the autonomous system cannot stop,
telematics or weather data that can also to some degree even if for a few micro-seconds. Henceforth, it is im-
be used to evaluate and monitor driving behaviour to imperative that the data – which is a food to the whole of
prove the driving experience. In such a scenario, all the the autonomous system in the self-driving vehicles – be
data can be combined with the socially available public reliable, flexible, real time and secure.
22
To ensure data integrity, reliability and accuracy, ad- and structure of the data to be used. A central unifica-
vanced data encryption techniques that are used in the tion process is needed worldwide with respect to data for
current financial and banking system could be leveraged self-driving vehicles. It is not fundamentally possible to use
for autonomous driving and associated data flow. En- different formats and structures for datasets because that
hanced integrity techniques that are often found in safety complicates the autonomous driving further which is not in
of life applications such as Aviation can be leveraged the best interest of the autonomous vehicle ecosystem.
to form primary data exchange standards that provide
security and integrity. Data encryption is a method that While setting a data unification standard, it is impor-
aims to reduce risks, however, not acting as a substitute tant to consider the standards laid down by established
for data protection controls. Autonomous vehicles should standard development organizations like Open Geospa-
be able to leverage data collected by the smart gateways tial Consortium, Industrial IOT, OMG, etc. Road maps
and the smart sensors and cameras on-board the car need to be set which shall lead to the development and
in encrypted format i.e. the binary vectors and submit implementation of the standards needed. As per the
the encryption of the vectors to the data servers/infra- technology ecosystem, best standards need to be defined
structure. This shall make the encrypted data unread- which is, for instance, standards for change detection
able, thus ensuring the accurate and reliable data to be and standards for interfacing. These standards need to
safeguarded till the software managing the algorithm is be available in a format used by all autonomous vehicles
presented the appropriate credentials and keys to unlock is the same or easy to interpret. Lessons can be learnt
the encrypted data. In such a scenario, data remains ac- from Europe, China and Japan. While in Japan, three
curate and reliable for the original binary vectors remain major automakers are struggling to agree on standards,
uncorrupted. The only drawback that exists in encryption in China, standards are being set for vehicle to vehicle to
of data is the loss or corruptions of the authentication communication. China also aims to establish a national
credentials would result in loss of the entire system or data standard by 2020 which would speed up the imple-
if managed by an application, could result only to loss of mentation of self-driving cars in the auto market. Safety
data managed by the applications, per say. However, in
the long run, encryption of data would ensure authentic
and reliable data for driverless cars. It is also important
to note that encrypted data will ensure safety from cyber
hacking. TRENDICATORS
Understanding the importance of data, automakers are
750 MB/SEC Data gathered by

taking a step forward to control maps and navigation
and that explains the acquisition of HERE by the German fully functioning AV
consortium of automakers consisting of Audi, BMW,
Daimler and others. While the autonomous vehicle shall
have sensors and cameras to collect real time data – the
% of global vechicle sales attributed to AVs
data in the infotainment system needs to be pre-installed
and therefore needs to be reliable especially in necessary
times. For instance, in a bad weather, the camera or sen- 4% 41% 75%
sors may only be able to deliver information to up to 300
meters ahead and for more information it is important to
see further. In such a case, the pre-installed data comes 2025: US$7,000 - US$10,000
into play and helps the autonomous vehicle to be ‘aware’ Estimated price consumers
of its surroundings. In conclusion, the data so available
will pay for AV technologies: 2030: US$5,000
2035: US$3,000
needs to be reliable, consistent and accurate according
to the set standards and structure.
85%
CAGR of AV sales for three largest markets
(North America, Western Europe and Asia-Pacific)
STANDARDS AND STRUCTURE between 2020 and 2035
As discussed, while reliable data is important, this data
needs to be authenticated on the basis of set standards Source: Factiva, Navigant Research and CNET
and structure. As the autonomous technology comes to the
forefront, it is necessary for the government to get involved
so as to analyse and address the issues of standards Figure 3.5 Trends on Autonomous Vehicles
23
and insurance are also two important facets of setting initiate communication will all the vehicles on the road to
standards. Self-driving cars have unique safety demands live route conditions in real time.
and therefore with the emergence of the new interna-
tional standards. While ISO 26262, sets the standards for In summary, the world is moving towards making fully
developing such systems – it is imperative to set stand- autonomous vehicles a reality and setting data standards
ards for safety requirements. and defining a structure of data is important. There-
fore in this context, open data standards, interoperable
For instance, to bring the promise of automated vehicles standards, policy standards and technical standards – all
to reality, an European Intelligent Transport Organization, need to be defined beforehand. If the data is available in
ERTICO, is defining a standardised interface known as a uniform format and structure – the benefits of auton-
SENSORIS to share information between the in-vehicle omous driving shall be far more than the risks and the
sensors and a dedicated cloud. The cloud is also expect- costs associated with it.
ed to warrant broad access, delivery and processing of
vehicle sensor data will enable location based services The government too has a significant role to play in set-
which are key for automated driving. Developed by the ting up standards and structure for geo-information data.
location cloud company, HERE, SENSORIS standards In our next chapter, we identify the crucial role that the
is expected to be universal and will enable driverless government plays in establishing an enabling and legal
vehicles to prepare for changing conditions well before framework for driverless cars and for open and interop-
the vehicle can see them. A universal data language will erable data policy.
24
GOVERNMENT AND SELF-DRIVING VEHICLES
Autonomous vehicles embody great potential and can yield and regulations and specific standards pertaining to
a wide array of benefits for citizens, organisations and the manufacturing, vehicle design. The government
governments. However, cashing in on those will require nu- also needs to establish the standards and legislations
merous framework conditions to be met, which are typically necessary for road management and traffic signal
‘governmental’. Accordingly, there is no doubt that gov- management to ensure safety on the roads. Policies
ernments, both at the national and international level, will need to be set for setting infrastructure for data and
have a major role to play. The role of the government will communications networks necessary to maintain safety.
emerge and re-emerge in different forms and shapes over To establish the necessary guidelines to ensure safety,
time and may, in fact, differ from continent to continent, it is imperative that the governments take a proactive
country to country and even from region to region. These approach.
roles are closely linked to a wide variety of factors, including
technical know-how, technical and data infrastructure, the (b) Privacy / Data Sharing
economic structure of the main enabling technologies, size Autonomous vehicles are referred to as a dynamic data
and character of demand for driverless cars, legal constitu- collection system, and because apart from making use of
ency, policy objectives, just to mention a few. the available data, they will be gathering a large volume
of data to operate, there is a growing concern about
Interestingly, the world of geo-community can play data ownership, collection and use. It is here where the
a small but crucial role in the sound development of government should create guidelines for the driverless
driverless cars. We can define and cluster the universal car industry to be transparent with consumers about data
key requirements that are conditional for the prosperous ownership and sharing. Simultaneously, the government
development of autonomous vehicles as: needs to provide clear rules and regulations with respect
to the private information that will be collected by these
(a) Establishment of an enabling and protective legal smart cars of its passengers. Because privacy is a concern,
framework the government has to play a defined role by developing
(b) Establishment of open, or interoperable, and interna- policies to further its cause.
tionally oriented data policy and governance
ESTABLISHMENT OF AN ENABLING AND

PROTECTIVE LEGAL FRAMEWORK
It has been accepted that autonomous cars are going to
be arriving soon and smart vehicle technology is advanc- Privacy / Data
ing with or without legislative and agency actions at the Sharing
federal level. Governments worldwide are realising the Safety
need for changes in legislation required specifically for
self-driving vehicles. These legislations are not only need-
ed for the vehicles but for the entire automobile ecosys-
tem. At present, changes in legislation are currently only
occurring for testing driverless cars. However, change can
still be expected to define the ownership, importation and
use of autonomous cars in the near future. Liability
There is a wide range of issues that necessarily need to

be addressed by the governments through legislation,
Cyber
which include:
Security
(a) Safety
It is necessary that the government leads the charge in
the establishment of data safety standards for auton- Figure 4.1 Establishment of an Enabling and Protective
omous vehicles. Also important is to define the rules Legal Framework
25
(c) Cyber Security produced the self-driving software? The answers to these
Autonomous vehicles are seen to be an easy target for questions need to be established beforehand for the realm
terrorists. Since the technology is going to be digitally of self-driving cars to grow.
driven, any cyber-attack carries the risk of coordinated
traffic disruptions and collisions. Also, because the car is ESTABLISHMENT OF OPEN, OR INTEROPER-
generating humongous amount of data every second, this ABLE, AND INTERNATIONALLY ORIENTED
data in the wrong hands could cause much bigger securi- DATA POLICY AND GOVERNANCE
ty threats around the globe than imaginable. It is vital that For the rapid and prosperous development of the
the governments develop a framework to improve critical self-driving vehicles, data is crucial. Data being both the
infrastructure and develops policy frameworks to encom- fuel and output of the autonomous vehicles– must be
pass the risks associated with the autonomous vehicles. available, accessible and re-usable. Clearly, this applies
to its own data – in particular (systems of) reference
(d) Liability data, which should feature high-quality standards and
To define Liability is going to be a major issue in auton- free and open access. The data is foreseen to bridge the
omous vehicles. How do you divide the blame between information gap across the autonomous vehicle ecosys-
a human driver and a car’s automated system? Is it the tem, thereby enhancing the sharing of benchmarks and
software? Or maybe it was the hardware? Or perhaps it standards that shall raise the efficiency. It is up to the
was due to the hardware and software interacting in un- government to realise that the open government data
expected ways. According to an article in the San Diego is crucial and the benefits so derived need to be under-
Union-Tribune, “if the issue of liability is not solved, stood and be conferred to the government and stake-
it could delay or even wipe out the vision of driverless holders of the self-driving vehicles.
cars gaining widespread consumer use.” Therefore,
an important question that continues to persist is who While flawless data exchange is crucial, the autonomous
is liable when an autonomous car crashes. Complex vehicle will lead to the establishment and elaboration of
systems in the self-driving vehicles are like any other data standards. The role of government may well be just
technology system is not immune to software failure. A to define the functional requirements i.e. in relation to
sense of responsibility aka a sense of liability must be the legal provisions set (road safety, liability, etc.) - and
established with respect to these vehicles. While a few follow the market forces. In such a scenario, the gov-
of the automobile manufacturers are willing to assume ernment would rather not step into the actual market.
full responsibility in case of a highly automated system However, the government will need to take a proactive
failing – the other manufacturers are taking a cautious role to monitor the process preventing de facto monopo-
view. Therefore, liabilities need to underline the assur- lies from arising once single standards emerge.
ance and resilience predominantly in the stakeholders
that deliver highly automated driving functions. While specific proprietary data and standards may
become an ‘essential facility’, concepts of ‘universal
A developing area of policy and legal framework, as the access” and ‘must carry’ should be applied (legislatively
use of autonomous car technologies increases, the incre- also), ensuring a level playing field and the need for en-
mental shifts in the liability and the responsibility of driving suring fair competition. In such a scenario, where need-
also increases. In the case of an accident – who does the ed, agile procedures will need to be established upfront
plaintiff sue is an important question. In a typical case, the so that market deficiencies and abuse of dominance can
plaintiff would assign the blame to the driver or the car be recognised early and addressed efficiently.
manufacturer, but in the event of self-driving vehicles, it is
tedious. The plaintiff may have to consider the operator of It is, therefore, becoming imperative that standard pol-
the self-driving vehicle, the car manufacturer, the soft- icies are formulated for both standards and data which
ware provider or the autonomous car technology creator shall create a predictable and inclusive ecosystem.
or the automobile manufacturer. Also, it is necessary to Appreciating the international context of the autono-
question that if software in the vehicle stops working or if it mous vehicles, governments, in particular, those of the
misinterprets a worn down sign and an accident occurs as power players in the value chain (car producers, platform
a result of it – who is going to be held liable. Will it be the owners, analytics companies) should aim to establish
department of transportation or the government for the an international coalition going beyond the short-term
poorly maintained signage or will it be the company that national economic interests.
26
Therefore, in this respect, the role that the geospa- CONCLUSION

tial community could play will be pivotal as (real-time) In conclusion, 25 years from now, roads globally shall be
location data will be crucial in realising the potential filled with millions of autonomous vehicles. According to
of the self-driving cars. Location data is the linking pin research conducted on auto industry, it is forecasted that
that could solve the international vacuum that appears 94.7 million vehicles with self-driving capabilities will
to exist. Appreciating and leveraging on this position, the be sold annually globally by 2035. As more automobile
community should help to set the policy agenda, set roles makers race to launch the fully automated self-driving
and responsibilities and suggest priority areas for action. vehicles, legislators and decision makers, are taking
Also, being ‘in the middle’ it should aim to pull together initiatives to include autonomous driving in their legal and
industry leaders and policy makers and key data and policy frameworks. The technology is evolving – geospatial
infrastructure holders, including those from the public industry is realising their role in the autonomous vehicle
sector like road authorities. Short term actions could domain and is exploring viable opportunities to contribute.
involve connecting to neighbouring communities like ITS
communities. Data, as discussed, is a strategic asset for the autono-
mous vehicles and therefore the data ecosystem is also
Open government spatial data for autonomous vehicles reaching a defining maturity. Data from in-car technol-
will bring significant social, environmental and eco- ogies such as vehicle technical sensors, environment
nomic benefits. The data shall bridge the information sensors, location/navigation platforms, high definition
gap across the entire eco-system, thereby enhancing satellite imagery receptors, etc., is the fundamental
the sharing of benchmarks and standards that shall enabler of self-driving vehicles. The role of geographic
raise efficiency. At the forefront of the benefits of open, data is critical to the proper functioning of the smart car.
interoperable data for self-driving vehicles is, therefore, Not surprisingly, data along with software platforms and
transparency, commercial value and participatory gov- infrastructure technologies (cloud, big data, etc.) will
ernance. transform the entire landscape of future mobility.
27
ANNEXURE
(A) Participants to the session ‘Autonomous Vehicles’ (AV) that was held at the Geospatial World Forum Conference
in Rotterdam (The Netherlands) on 25 May 2016
Moderated by: Rob van de Velde and Arnoud de Boer, Geonovum
NAME ORGANIZATION
Denise McKenzie Open Geospatial Consortium, UK
Kumar Navalur Digital Globe, USA
Andy Wilson Ordnance Survey of Great Britain, UK
Sandeep Singhal Google, India
Chris Gibson Trimble, USA
Michiel Beck Ministry of Infrastructure and Environment, The Netherlands
Robert Voute CGI Nederland, The Netherlands
Marije de Vreeze Connekt, The Netherlands
Marc De Vries Geonovum, The Netherlands
Frans Jorna Hiemstra, The Netherlands
Theo Thewessen Geodan, The Netherlands
Hans Nobbe Rijkswaterstaat, The Netherlands
Harald Kraaij Kadaster, The Netherlands
Hans Nouwens Smart Data City, The Netherlands
Noud Hooyman Ministry of Infrastructure and Environment, The Netherlands
Heide, Jene van der Kadaster, The Netherlands
Rob Bieling Mapcreator, The Netherlands
Rob Huibers Andes, The Netherlands
Taner Kodanaz Digital Globe, USA
Jean Pierre Krause Msc ETH & MA HSG, Switzerland
Anamika Das Geospatial Media and Communications, India
28
(B) Participants to the session ‘Autonomous Vehicles’ (AV) that was held at the Geospatial World Forum Conference
in Hyderabad (India) on 24 January 2017
Moderated by: Rob van de Velde and Marc de Vries, Geonovum
NAME ORGANIZATION
John Renard Cyient EMEA, UK
Kumar Navulur Digital Globe, USA
Sundara Ramalingam Nagalingam Nvidia, India
Rajendra Tamhane Genesys International, India
Vijay Kumar TCS, India
Bengt Kjellson National Land Survey, Sweden
Ivan Deloatch Federal Geographic Data Committee, USA
Alessandro Annoni Joint Research Centre, European Commission
Rob Beck NEO BV, The Netherlands
Alan Smart ACIL Allen Consulting, Australia
Andreas Gerster FARO, Germany
Dan Kruimel AAM, Australia
Anand Murthy Symtronics Automation Private Limited, India
Ashok Gupta National Technical Research Organisation, India
Paomesh Menon Genesys International, India
Kariappa SECON, India
Arya Bhattacharya Mahindra Ecole Centrale, India
Vinod Mishra MapmyIndia, India
Shivalik Prasad MapmyIndia, India
Louis Nastro Applanix Corporation (Trimble), Canada
Briad Ysseloyk Applanix Corporation (Trimble), Canada
Ad Bastiaansen ILOC Group, The Netherlands
N.D. Gholba Indian Air Force
Vivek Saxena Defence Terrain Research Laboratory (DTRL) of Defence

Research and Development Organization (DRDO), India
Pankaj Dahiya National Remote Sensing Centre, India
Robin Jiss National Technical Research Organization, India
George Lukes Institute for Defence Analyses (IDA), USA
Akshay Bandiwdekar Swift Navigation, USA
29
Rob van de Velde Geonovum, The Netherlands
Marc de Vries Geonovum, The Netherlands
Anamika Das Geospatial Media and Communications, India
Ananya Narain Geospatial Media and Communications, India
Meenal Dhande Geospatial Media and Communications, India
REFERENCES
1 http://edition.cnn.com/2016/10/12/design/self-driving-boats-mit/
2 https://www.gsa.europa.eu/newsroom/news/driving-towards-autonomous-vehicle
3 Parts of this chapter are largely based on the outcomes of the Autonomous Vehicles’ sessions held at the 2016 and 2017 Geospatial
World Forum Conferences held in Rotterdam and Hyderabad respectively. Names of the participants to these meetings are listed in
Annex [ ].
4 https://www.tesla.com/blog/all-tesla-cars-being-produced-now-have-full-self-driving-hardware
5 http://autoweek.com/article/car-news/gm-wants-test-autonomous-vehicles-public-streets
6 http://www.itpro.co.uk/strategy/27302/driverless-cars-ford-to-invest-1-billion-into-ai-startup
7 http://www.carsguide.com.au/car-news/google-and-fiat-partner-up-for-autonomous-car-research-50870
8 http://readwrite.com/2017/01/07/honda-ces-2017-tl4/
9 http://www.networkworld.com/article/3167005/car-tech/toyota-funds-ai-research-to-build-autonomous-cars.html
10 http://www.forbes.com/sites/elaineramirez/2017/02/07/how-south-korea-plans-to-put-driverless-cars-on-the-road-by-2020/#c-
06fa0d4c877
11 http://www.wired.co.uk/article/bmw-intel-driverless-tech-ces-2017
12 https://www.macrumors.com/2017/01/05/audi-and-nvidia-autonomous-car-2020/
13 http://www.carscoops.com/2017/01/heres-how-mercedes-upcoming-autonomous.html
14 http://www.caradvice.com.au/519047/googles-waymo-autonomous-cars-getting-safer/
15 http://www.forbes.com/sites/moorinsights/2017/01/23/chipmakers-get-serious-about-autonomous-driving-at-ces-2017/#-
428fea67888d
16 https://blogs.microsoft.com/blog/2017/01/05/microsoft-connected-vehicle-platform-helps-automakers-trans-
form-cars/#sm.0001h3xbfv1c8pf2nq8i1jskyrxd0
30
ABOUT GEONOVUM
Geonovum is the National Spatial Data Infrastructure (NSDI) executive committee in the Netherlands. The organiza-
tion devotes itself to making the government perform better with spatial data, by developing and managing spatial
data standards.Goenovum is a public organisation, supported by the Ministry of Infrastructure and Environment, the
Ministry of Economic Affairs and The Dutch Cadastre and the Geological Survey of the Netherlands.
ABOUT GEOSPATIAL MEDIA AND

COMMUNICATIONS
Geospatial Media and Communications, with its vision of Making a Difference through Geospatial Knowledge in World
Economy and Society, works to build the geospatial industry in all its facets. It is a catalyst organisation pursuing
business objectives towards promotion and facilitation of growth of Geospatial Industry through creating awareness,
policy advocacy, business development and by connecting stakeholders and communities worldwide. Since 1997 Ge-
ospatial Media has invested its energies and resources in developing geospatial market globally and has provided a
leadership role in promoting geospatial tools to several stakeholders with a thrust on prospective industries. Head-
quartered in India, it has regional offices in USA, UAE, Brazil, South Africa, Malaysia and The Netherlands.
Geospatial Media achieves its objectives by publishing content on geospatial technologies, trends, policies and ap-
plications. It also undertakes policy advocacy, business consulting and produces industry reports on market behav-
iour, requirements, challenges and prospects of geospatial information and applications for society and economy.
In addition, it is one of the few professional organisations that organises many national, regional and international
conferences on the domain.
ACKNOWLEDGEMENTS
Our appreciations to the efforts put in by the GEONOVUM Team:
Rob van de Velde
Marc de Vries
Arnoud de Boer
Our appreciations to the efforts put in by the Geospatial Media and Communications’ Research Team:
Anamika Das
Ananya Narain
for preparing a comprehensive report &
To our Graphic Team for creating visual concepts

Pradeep Singh
Subhash Kumar
Manoj Kumar
31

Self DrivingVehiclesReport

Uploaded by

Copyright:

Available Formats

Self DrivingVehiclesReport

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Self DrivingVehiclesReport

Uploaded by

Copyright:

Available Formats

Self-Driving Vehicles (SDVs)

Background and Scope4

II. SELF-DRIVING VEHICLES 6

Direction and Trends

Current Scenario: Innovations in the Industry

III. SELF-DRIVING VEHICLES ECOSYSTEM 12

Action Points for the Ecosystem

IV. GEOSPATIAL AND SELF-DRIVING VEHICLES  17

Autonomous Vehicles as Data Producers

Geospatial Technologies - ‘Under the Bonnet’

Geographic Information: Data

Authenticity and Reliability

Standards and Structure

V. GOVERNMENT AND SELF-DRIVING VEHICLES 25

Establishing an Enabling and Protective Legal Framework

Establishing of Open, or Interoperable and Internationally Oriented Data Policy

List of Figures & Tables

Table 1.1 Current Scenario: Innovations in the Industry 8-9

BACKGROUND & SCOPE

• The concept of self-driving vehicles

• Action points of the self-driving cars ecosystem

• Role of the geospatial industry

• The defining role of the government

WHERE IN THE WORLD ARE SELF-DRIVING CARS?

AUSTIN, TX, USA ROTTERDAM, THE NETHERLANDS

WAGENINGEN, THE NETHERLANDS

Figure 1.1 Where in the World are Self-Driving Vehicles (SDVs)

AUTONOMOUS VEHICLE DEPLOYMENT TIMELINE

Figure 1.2 Autonomous Vehicle Deployment Timeline

CONNECTED AND AUTONOMOUS VEHICLE TECHNOLOGY ROAD MAP

Figure 1.3 Connected and Autonomous Vehicle Technology Road Map

CURRENT SCENARIO: INNOVATIONS IN THE INDUSTRY

General Motors has started testing self-driving Bolt electric vehicles

Ford invests $1billion in an Artificial Intelligence start-up with the

Fiat Chrysler Automobiles (FCA) has partnered with Waymo (previous-

Honda has unveiled a new self-driving concept called the Cooperative

Toyota is partnering with Stanford University and MIT to research AI and

Hyundai already has Level 2 smart sense technology, an advanced driver

BMW is working in collaboration with Intel and Mobileye to put a fleet of

Audi is adopting Nvidia’s drive computing platform to accelerate the

There is a fourfold improvement in Google’s Autonomous Cars project

Nvidia is focusing on putting a lot of compute horsepower in the vehicle

$87 94.7 78% COST

Autonomous cars total savings:

Source: Predictions for U.S.

Economic Impact of Connected and Autonomous Vehicles 2014-2030 in United Kingdom

Source: Connected and

POTENTIAL BENEFITS OF AUTONOMOUS POTENTIAL RISKS AND COSTS OF

SELF-DRIVING VEHICLES ECOSYSTEM

ACTIONS POINTS FOR THE AUTONOMOUS VEHICLE ECOSYSTEM

Telematics and Infotainment

Mobile Service Providers

Finance Service Providers

Transport Service Providers

Deep Learning and Artificial Intelligence

Original Equipment Manufacturers

Geospatial Content Providers

GEOSPATIAL AND SELF-DRIVING VEHICLES

Autonomous car data vs. human data

THE COMING FLOOD OF DATA IN SONAR

Background and Scope4

II. SELF-DRIVING VEHICLES 6

III. SELF-DRIVING VEHICLES ECOSYSTEM 12

IV. GEOSPATIAL AND SELF-DRIVING VEHICLES 17

V. GOVERNMENT AND SELF-DRIVING VEHICLES 25

Establishing of Open, or Interoperable and Internationally Oriented Data Policy

Table 1.1 Current Scenario: Innovations in the Industry 8-9

• The concept of self-driving vehicles

• Action points of the self-driving cars ecosystem

• Role of the geospatial industry

• The defining role of the government