ISO 9001:2015

Study Paper
on
AI in Automotives

Study Paper on “ AI in Automotives ”, Smart Network Division, TEC

 Content
S/N Topics Page No.
Abstract 1

1.0 Introduction 1

2.0 Artificial Intelligence 1

3.0 Essential elements of AI 2
4.0 Autonomous Cars 4

4.1 Definition 4
4.2 Levels of Automation 4
5.0 AI in Automotive 5
5.1 Vehicle Communication Architecture 6
5.2 AI in ITS 7
6.0 Security & Privacy 20
7.0 Standards in AI 22

7.1 International 22
7.2 National 23
8.0 Use cases in AI for Automotives 23
8.1 International 23
8.2 National 24
9.0 Challenges with Autonomous Cars 24
9.1 Cyber Threats to Autonomous Cars 24
9.2 Ethical issues 24
9.3 Others 25
Conclusion 27
Glossary 28
References 29

ii

Study Paper on “ AI in Automotives ”, Smart Network Division, TEC

Abstract
The purpose of this study paper is to bring into the limelight, the Artificial Intelligence in the
automotive industry. AI is a broad term using which artificial things can be made intelligent.
AI provides intelligence to automotives to achieve more and more ease in driving. In this era,
focus is on driverless cars, which is due to the influence of artificial intelligence in the
industry. Vehicles need to connect and communicate over road regarding various factors such
as traffic, weather conditions, etc. The paper covers the necessary protocols and architecture
of vehicle communication. Security & Privacy issues in smart automotive system lie in wide
ranges across WiFi, LAN; tracking of vehicle, etc. All this works as per the Standardization
in AI for Automotives: International and National. FGAI4AD lays down standards for
dealing with accident situations. FG-AI4AD supports suspension operations and services
enabled by AI programs through independent and assisted driving. Lastly, it focuses on
challenges with such cars and cyber threats which are liable to occur. Also, the ethical issues
over which debate is ongoing: who is liable for a mishap? have been covered here.

1.0 Introduction
In this era of smart technology and rising expectations and demands, there lies the need of
enriching the automotive sector. The paper “AI in Automotives” enlightens the latest
technologies being worked upon in automotives. Also it highlights the ongoing HPC (High
Performance Computing) concept. It explains the significance, usage and applications of AI,
various protocols, to the status of the technologies going on in various countries and various
car models.
According to ITU News Magazine titled ‘Technology driving tomorrow’s cars’, By the shift
of artificial intelligence capabilities — in support of automated driving and digital assistants
— shifting from the cloud to in - vehicle systems in the form of more powerful processors,
improved vehicle networks and on-board storage, we are leading to technological
advancements in automotives. Cars are getting smarter at understanding what humans are
doing and helping them move and arrive at their destinations accurately and safely. HERE is
working to integrate location information related to navigation with sensor based contextual
information compiled from Audi, BMW and Daimler vehicles to help drivers avoid obstacles
and road hazards along their path.
AI has the capability to sense various characteristics of the driver and the car. It can sense the
speed and position of the vehicle, attention of the driver (if he deviates from actual path),
weather situation, traffic levels, etc. If the driver deviates from original and expected track, it
senses that there is some problem and stops the car. This is how useful is AI in Automotives.

2.0 Artificial Intelligence

Artificial intelligence (AI) refers to the imitation of human intelligence in machines designed
to think like humans and imitate their actions. The term can also be applied to any machine
that displays features that resonate with the human mind such as learning and solving
problems. New paradigm shift has been created by deep learning and machine learning. Fig.1
shows how these terms are related through a Venn diagram.

1
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Artificial Intelligence

Machine Learning

Deep Learning

Fig. 1 Artificial intelligence

The automotive industry has a long running track record to make use of the latest
technologies to bring the efficient, innovative and safe vehicles to the market. These
technologies nowadays include artificial intelligence and high performance computing- two
keys of automotive success.
For many years, HPC has been used to manufacture good products in the modeling, design
and simulation applications; today this focus is broadening to substantial investments in AI.
AI usage is moving ahead to drive semi-autonomous and autonomous cars.
These advanced capabilities, coupled with rising consumer expectations have pushed the
automotive industry into a period of digital transformation. These new technologies have
helped reduce the costs and give consumers more of what they want.
There are three main protocols for vehicle connectivity and communication:
 V2V- Vehicle to vehicle communication: This is a smart technology which enables
vehicle data to exchange from one vehicle to another. This is for short range
communication, approx. 300 meters, sharing the details on speed and position of
vehicles.
 V2I- Vehicle to infrastructure communication: This captures data from surroundings,
such as traffic congestion, bridge clearance levels, weather forecast, and then
wirelessly transmits it to other drivers so that they can remain safe.
 V2X- Vehicle to everything communication: This encompasses the above two, V2V
and V2I. This makes the automobile smarter and gives the power to communicate
with the traffic system, including cars and infrastructure.

3.0 Essential elements of Artificial Intelligence

3.1 Machine Learning
Machine learning is the use of artificial intelligence (AI) that gives systems the ability to
automatically learn and develop from experience without being clearly planned. Machine
learning focuses on the development of computer programs that can access data and use it for
self-study.

2
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
The learning process begins with looking at or data, for example, direct experience, or
instruction, to look at patterns in data and make better decisions in the future based on the
examples we provide. The main purpose is to allow computers to read automatically without
human intervention or assistance and to correct actions accordingly.
3.2 Deep learning

Deep learning is an artificial intelligence (AI) activity that mimics the functioning of the
human brain in data processing and decision-making patterns. Deep learning is a type of
machine learning in artificial intelligence with readable networks that can be directed from
random or unlabeled data. Also known as deep neural learning or a deep neural network.

Deep learning has emerged in the digital age, which has brought about the explosion of data
by all means and in all regions of the world. This data, known simply as big data, is taken
from sources such as social media, online search engines, commerce platforms, and online
movie theaters, among others. This enormous data is easily accessible and can be shared with
fintech applications.

However, the details, commonly constructed, are so vast that it can take decades for people to
understand and extract relevant information. Companies are recognizing the incredible power
that can come from exposing this wealth of information and are becoming increasingly
accustomed to AI automated support systems.

3.3 Big data

Big data is a combination of systematic and informal data collected by organizations that can
be mapped and used in machine learning projects, prediction modeling and other advanced
analytics applications.
Systems that operate and store large amounts of data have become commonplace in the
structure of data management in organizations. Big data is usually seen in 3Vs: Volume: large
data volumes in multiple locations, Variety: different types of data stored in large data
systems and Velocity: the speed at which data is processed, collected and processed. These
qualities were first identified by Doug Laney, who then became an analyst at Meta Group
Inc., in 2001. Recently, some Vs have been added to various definitions of big data, including
authenticity, value and variability.
Big data comes from many different sources, such as business transactions, customer
information, medical records, online click logs, mobile applications, social networks,
scientific research repositories, digital data and real-time data sensors used in Internet of
Things (IoT) environments. Data can be left untouched in large data systems or processed
using data mining tools or data preparation software and is therefore suitable for use in
certain analytics.

3.4 Blockchain
The blockchain is a growing list of records, called blocks, connected using cryptography.
Each block contains a cryptographic hash of the previous block, time stamp, and transaction
data. Blockchain has been trending these days because of the world's most popular currency -
Bitcoin. Many governments and leading banks have decided to bring their standard

3
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
transactions based on the Blockchain concept. The use and power of this framework is
enormous and is considered to change the way work is performed in various domains.

4.0 Autonomous cars

4.1 Definition
Autonomous cars are self-driving cars. Drivers do not need to worry about steering here. This
depends on the degree of independence that will be easily provided in the model. These
vehicles operate like traditional vehicles, without the need for human passenger control or
any human passenger required to be in the vehicle.
4.2 Levels of automation
There are six levels of automation:

Fig. 2 Levels of automation

Source: 1

Level 0: No automation means manual control, like traditional car

Level 1: Driver assistance means that it features a single automated system, e.g., it monitors
speed through cruise control.
Level 2: Partial automation means vehicle can perform steering and acceleration on its own.
Driver can still monitor and take control at any time.
Level 3: Conditional automation means Environmental detection capabilities. The vehicle can
take on most driving tasks, but human override is still required.
Level 4: High automation means the vehicle performs all driving tasks under specific
circumstances. Geofencing is required. Human override is still an option.
Level 5: Full automation means the vehicle performs all driving tasks under all conditions.
Zero human interaction required.

4
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
5.0 AI in Automotives
Autonomous vehicles rely on sensors, actuators, complex algorithms, machine learning
systems, and powerful software processors.

Independent vehicles create and maintain a map of the environment based on a variety of
sensors located in different parts of the vehicle. Radar sensors monitor the position of nearby
vehicles. Video cameras detect traffic lights, learn traffic signals, track other vehicles, and
detect pedestrians. Lidar sensors (light detection and lightning) jump light edges in areas
close to the car to measure distances, find road edges, and indicate line markings. Ultrasonic
sensors on wheels receive curbs and other motors during parking.

Sophisticated software then processes all of these sensors, paves the way, and sends
instructions to car actuators, speed controls, braking, and driving. Strong codes, avoidance
techniques, predictable modeling, and object recognition help software to follow traffic rules
and navigate.

Fig. 3 shows the components in Advanced Driver Assistance System. ADAS not only assists
with parking, locking of car doors (using mobile phones) but also collects information on the
car, driver, driving habits and passenger. Based on this information, ADAS makes an
informed decision.

Fig. 3 Components of autonomous cars

Source: 2

5
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
5.1 Vehicle Communication Architecture
Advanced vehicular communications and services in V2X (Vehicle to Everything) & Internet
of Vehicles (IoV):
 The V2X, which stands for 'everything in the world', is the umbrella term for a car
communications system, where data from sensors and other sources travels with high
bandwidth, low-latency, high reliability links, which opens up a completely independent
driving mode.
 There are several V2X features, including vehicle-to-car (V2V), car-to-infrastructure
(V2I), pedestrian (V2P), and vehicle communication (V2N). In this multicultural
ecosystem, cars will communicate with other vehicles, in infrastructure such as robots or
parking lots, pedestrians on smartphones, and on mobile network operators. Different
operating cases will have different set of requirements, a communication system that must
be handled efficiently and costly. IoV global network of vehicles – enabled by various
Wireless Access Technologies (WAT) involves Internet and includes heterogeneous
access networks
 IoV–can be seen a special use case of Internet of Things (IoT)
 IoV Target domains:
o Vehicles driving and safety (basic function – in Vehicular Ad hoc Network
(VANET))
o Novel domains:
Traffic management, automobile production repair and vehicle insurance, road
infrastructure construction and repair, logistics and transportation, etc.
 Commercial , objectives, architecture
o Business oriented architecture
o High opportunities for various apps (safety, traffic optimization and efficiency,
infotainment, etc.)
 Collaboration capabilities:
o collaboration between heterogeneous nets, reliable Internet service
 Communication types:
o includes all V2X types of communications
 Processing power and decision capabilities:
o high capabilities – (cloud based), big data, data mining, …
 Compatibility with any personal devices
 Scalability:
o Scalable (and it integrates various access: VANET, WiFi, 4G/LTE, 5G...)
 Connectivity:
o “always-connected”-feature is possible; one can use the best network type
 Network/environment awareness:
o global network awareness is possible (cloud-assisted)
 Cloud Computing/Edge computing (CC/EC) compatibility:
o the main operations can be based on CC/Edge computing services

Fig. 4 & 5 show the architecture of IoV and V2X respectively. In IoV, data is obtained from
outside sources (vehicle, infrastructure, sensor, etc.). This data is collected, cleaned and
stored in database, in the nearby edge/device/vehicle (in case of less latency).
Data is saved in the cloud as per its requirement after proper feature extraction, as the main
data center with AI functions working thereby also. Latency time is more in this case.

6
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 4 Data Flow in IoV
Source:3

Fig. 5 Vehicle communication architecture

Source: 4

5.2 AI in ITS
According to the U.S. Department of Transportation, "Intelligent Transportation Systems
(ITS) uses a variety of technologies to monitor, evaluate, and manage transportation systems
to enhance efficiency and safety." Putting the ideas of the science fiction style model aside
for now, this explanation can be simplified in the following concepts of what makes
intelligent transport: management, efficiency, and safety. In other words, smart transport uses

7
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
new and emerging technologies to make traveling around the city much easier, more
expensive (city and individual), and safer.
What emerging technologies support these new opportunities? Mainly an increase in IoT
devices and 5G communication technologies. The first offers inexpensive sensors and
controls that can be installed on almost any controlled and remote-controlled body machine.
The latter provides the high-speed communication required for managing and controlling
travel systems in real time with minimal latency. Fig. 6 & 7 support this concept.

Fig. 6 Application of AI technologies in smart transportation

Source: Artificial Intelligence across Industries IEC

Fig. 7 Enabling autonomous vehicles with AI technologies

Source: Artificial Intelligence across Industries IEC

8
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
5.2.1 Elements of ITS
5.2.1.1 AI in Advanced vehicle control systems
Advanced vehicle control systems are oriented to complementing major portions of the
driving task.
a) Longitudinal collision avoidance: AI helps to prevent head-on, rear-end or backing
collision between vehicles, vehicles to objects or pedestrians. The overall system
structure for multiple-vehicle collision avoidance and impact mitigation could be
depicted in Fig 8 It includes remote sensor detection of relative distance and speed
detection/estimation, V2V (Dedicated Short Range Communication), control, and
decision making. An algorithm for frontal-rear collision avoidance and impact
minimization should take the following factors into account:
(i) Multiple-vehicle collision scenario,
(ii) Mechanical definition of impact,
(iii) Vehicle mass,
(iv) Relative distance and relative speed at braking,
(v) Relative distance and speed at collision (time instant of impact),
(vi) Deceleration capability (braking capability) of each vehicle in the coupled group,
(vii) Constraint to the first vehicle in the front end of the coupled group,
(viii) Constraint to the last vehicle in the rear end of the coupled group.

Fig. 8 Longitudinal collision avoidance

Source: 6
b) Lateral collision avoidance: AI helps prevent collisions when vehicles leave their lane
of travel.

9
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 9 Lateral collision avoidance
Source: 7
c) Intersection collision avoidance: AI helps prevent collisions at intersections.

Fig. 10 Intersection collision avoidance

Source: 8
d) Vision enhancement systems: AI improves driver’s ability to see the roadway and
objects on or along the roadway.

Fig. 11 Vision enhancement system

Source: 9
e) Pre-crash restraint deployment: AI anticipates an imminent collision and activates
passenger safety systems before the collision occurs earlier than is currently feasible.

10
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 12 Pre-crash restraint deployment
Source: 10
f) Automated road systems

Fig. 13 Automated road systems

Source: 11
g) Safety readiness: AI provides warnings about the condition of the driver, the vehicle
and the roadway.
Automotive safety Integrity Level (ASIL) refers to the level of vehicle safety. It is a risk
classification system defined by ISO 26262 standards for road safety performance.
The standard defines operational safety as "the absence of unreasonable risk due to accidents
caused by malfunction of electrical or electronic systems." ASILs introduce safety
requirements - depending on the possibilities and acceptance of risks - so that car parts
comply with ISO 26262.

Fig. 14 Safety readiness

Source: 12

11
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
5.2.1.2 AI in Advanced traffic management systems
a) Traffic network monitoring and control: AI manages the movement of traffic on
streets and highways.

Fig. 15 Traffic network monitoring and control

Source: 13
b) Travel demand management: AI supports policies and regulations designed to
mitigate the environmental and social impacts of traffic congestion.

Fig. 16 Travel demand management

Source: 14
c) Incident detection and management: AI helps public and private organizations quickly
identify incidents and implement a response to minimize their effects on traffic.

12
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 17 Incident detection and management
Source: 15
d) Emissions testing and mitigation: AI provides information for monitoring air quality
and developing air quality improvement strategies.

Fig. 18 Emissions testing and mitigation

Source: 16
e) Parking management: AI provides information of parking lots or manages the entry
and exit of vehicles.

13
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 19 Parking management
Source: 18
5.2.1.3 AI in Advanced traveller information systems
a) Pre-trip travel information: AI provides information for selecting the best
transportation mode, departure time and route.

Fig. 20 Pre trip travel information

Source: 39
b) En-route driver information: AI provides driver advisory and in-vehicle signing for
convenience and safety during travel.

14
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 21 En-route driver information
Source: 20
c) En-route transit information: AI provides information to travellers using public
transportation after the start of the trip.

Fig. 22 En-route transit information

Source: 30
d) Route guidance: AI provides travellers with simple instruction on how to best reach
their destinations.

Fig. 23 Route guidance

Source: 16

15
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 e) Ride matching and reservation: AI makes ride sharing easier and more convenient.

Fig. 24 Ride matching and reservation

Source: 17
5.2.1.4 AI in Advanced public transportation systems
a) Public transportation management: AI automates operations, planning and
management functions of public transit systems.

Fig. 25 Public transport management

Source: 22
b) Personalized public transportation: AI offers flexibly routed transit vehicles for more
convenient service to customers.

16
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 26 Personalized public transportation
Source: 23

5.2.1.5 AI in Advanced fleet management systems

a) Vehicle administration: AI provides electronic purchasing of credentials and
automated mileage and fuel reporting and auditing.
b) Safety monitoring and tracking: AI senses the safety status of a commercial vehicle,
cargo and driver.

Fig. 27 Safety monitoring and tracking

Source: 20

c) Fleet management: AI facilitates the fleet management system.

d) Vehicle preclearance: AI facilitates domestic and international border clearance,
minimizing stops.
e) Automated roadside safety inspections: AI facilitates roadside inspections.
f) Hazardous material incident response: AI provides immediate description of
hazardous materials to emergency responders.

17
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 28 Road safety and fleet management
Source: 25

5.2.1.6 AI in Emergency management systems

a) Emergency notification and personal security: AI provides immediate notification of
an incident and an immediate request for assistance.
b) Public travel security: AI creates a secure environment for public transportation
operators.
c) Emergency vehicle management: AI reduces the time it takes emergency vehicles to
respond to an incident.

18
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 29 Emergency management system
Source: 29
5.2.1.7 AI in Electronic payment services
a) Electronic payment services: AI allows travellers to pay for transportation services
electronically based on short-range vehicle-to-infrastructure communication.

19
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Fig. 30 Electronic payment service
Source: 27
It also allows travellers to pay for transportation services electronically based on
GNSS and wide area communication.
5.2.1.8 AI in Pedestrian supporting systems
a) Pedestrians’ route guidance: AI helps pedestrians to find appropriate directions to go
to destinations.
b) Vehicle-pedestrian accident avoidance: AI detects dangerous situations, and to
provide necessary alarm both for pedestrians and drives.

Fig. 31 Pedestrian supporting system

Source: 21

6.0 Security & Privacy for “AI in Automotives”

As the development of self-driving technology continues, the hope of driverless cars
operating on roads is at hand. Industry experts predict that driverless cars will be
commercially available in the next five to ten years. However, the use of these technologies
raises important privacy and safety issues that need to be checked and repaired before these
vehicles can be fully sold.

20
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 Security risks of LTE-V2X/DSRC
• Dynamic changes of network topology make it difficult to detect
attacks.
• Lack of security key update management mechanism to manage the
legality and timeliness of the vehicle identity
• The mechanism of isolation and punishment has not yet been established to
the vehicle nodes which are untrusted or out of control.

Security risks of cellular mobile comm. (2G/3G/4G)

• Based on the cellular mobile communication systems, vehicular system can
provide remote WiFi hotspots, which brings potential an attack portal
• Through the cellular mobile communication system, the sound and data are
transmitted between vehicle devices by means of microphones. Once an
attacker breaks through the cellular mobile communication system, it will
cause the abnormality of automobile system

Security risks of WiFi

• It could be a springboard for attackers to launch an attack on the
vehicle.
• Through WIFI, devices can access to the car's internal network, so attackers
can get the internal data of vehicles
• By setting up pseudo AP, attackers can access the vehicle communication data
by cheating the users.

Security risks of satellite comm. (GPS/BEIDOU, etc.)

• There are security defects in the satellite communication module equipped on
on-board system.
• Navigation location data spoofing, location data replay, etc.

Security risks of wireless LAN communication (Bluetooth, Zigbee, etc.)

• Attacks mainly concentrate in three aspects: authentication process for security
specifications, simple matching process, and data encryption and decryption

Privacy issues
• Location tracking
• Leakage of passenger and owner data
• Stealing of sensor data, e.g., voice recognition system and control system

21
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
7.0 Standardization in AI for Automotives
7.1 International
Global SDOs ITU and IEC / ISO have developed appropriate standards, and created more AI-
focused focus groups, many of which were highlighted during the March 2019 GSC-22
conference on AI.

ISO/IEC
 The ISO / IEC JTC1 SC42 Artificial Intelligence team operates at all levels of the
operating environment (ISO / IEC NP TR 24030), ethics and management (ISO / IEC
AWI TR 24368, ISO / IEC AWI 38507), terminology 22989), big data (ISO / IEC
AWI TR 20547-1) and selection (ISO / IEC NP TR 24027), big data and AI
architecture (ISO / IEC DIS 20547-3.2, ISO / IEC WD 23053), risk management (ISO
/ IEC AWI 23894), reliability and robustness (ISO / IEC PDTR 24028, ISO / IEC NP
TR 24029-1).
 ISO/TC 184 Automation systems and integration team

ITU-T
Since 2019, road injuries have been the leading cause of death among children and young
adults aged 5-29 (in addition to HIV and TB). AI can play a key role in reducing 1.3 million
road deaths and 25 million injuries (SDG 3.6) occurring annually, and also promotes safe,
affordable and sustainable security systems (SDG 11.2). However, widespread AI
distribution, which is socially acceptable, depends on technology to achieve public trust.
FG-AI4AD supports suspension operations and services enabled by AI programs through
independent and assisted driving. FG-AI4AD will focus on the evaluation of AI behavior
responsible for dynamic driving activity in accordance with the 1949 and 1968 UNECE
Global Road Safety Forum. To build public trust it is important that AI performance on our
roadway integrates, or exceeds, the performance of a competent and caring driver. FG aims to
create international synergies in the definition of a minimum performance limit for these AI
programs (such as AI as Drive).

IEEE
 The IEEE P7000 addresses specific issues at the intersection of technical and ethical
approaches, program design, system definition, data confidentiality, bias, terminology
and ontologies, KPIs for face recognition.
 IEEE ECPAIS (Ethics Certification Program for Autonomous and Intelligent
Systems) aims to create certification data and marking methods that improve
transparency, accountability and reduce algorithmic bias in Autonomous and
Intelligent Systems (A / IS).
 IEEE A-IS is the IEEE Global Initiative on Ethics of Autonomous and Intelligent
Systems and aims to ensure that all stakeholders involved in the design and
development of independent and intelligent systems are educated, trained and
empowered to prioritize ethical considerations.

22
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
Compliance with the EU ethical guidelines
The EC HLEG Reliable AI Guidelines (see the Trusted AI Code of Conduct Guidelines,
published 8 April 2019) include the seven required categories. Within the current structure of
ETSI, each technical body will need to find ways to apply the guidelines in their AI-related
work, each focusing on the same “learning curve”. ETSI should develop AI approach based
on sharing and collaboration between groups.
As a first step, ETSI TBs should agree to analyze the technology proposed by HLEG to
evaluate reliable AI, whenever AI capabilities are demonstrated at ETSI levels.
To ensure the prioritization of ethical issues relating to the use of AI, ETSI and its members
must consider their commitment to the ETSI code of conduct, to develop standards compliant
with the Universal Declaration of Human Rights and EC HLEG guidelines and EU guidelines
on ethics in artificial intelligence: Context and implementation as it evolves.
SAE
SAE was founded in 1905 by Henry Ford, Andrew L. Riker, Edward Birdsall and John
Wilkinson. The organization has contributed to many well-known people including Thomas
Edison, Charles Kettering, Glenn Martin and Orville Wright. Another first SAE member,
Elmer Sperry, coined the word automotive, derived from Greek cars, meaning independence,
and Latin motus, meaning movement.
SAE International is a professional organization and organization that develops standards in
the engineering industry, focusing on transport sectors such as automotive, aerospace and
commercial vehicles. The organization was originally founded as the Automotive Engineers
Association.
SAE is best known for its horsepower ratings and standards in the aerospace industry. While
none of their recommendations or standards is legally binding, they are generally accepted by
state-owned enterprises and organizations.
US SAE offices in Troy, Michigan and Warrendale, Pennsylvania. The organization works
with 138,000 individual members worldwide through major initiatives that include
consolidation, promotion of STEM (science, technology, engineering and mathematics),
technology development and certification. SAE also hosts conferences, releases publications
and sponsors of co-building competitions.
7.2 National
The standards making process for AI in Automotives is in nascent state/ under progress.

8.0 Use Cases of AI for Automated cars

8.1 International
Country City Car Company Autonomy level
USA California Tesla 4,5
Germany Stuttgart Daimler 2

23
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 EU Munich BMW AG 5
US Michigan Ford 4
US CA AutoX 4
China Beijing Baidu 4
Singapore Bendemeer LTA 3,4,5
Tbl 1: Implementation of AI in various car models worldwide

8.2 National
Although the Govt of India has denied manufacturing driverless cars citing job loss as the
reason, a few models are being constructed which have their genre in foreign countries.
City Company Model Autonomy Level
Gurugram Maruti - -
New Delhi Tata Elxsi 1,2,3
New Delhi Tata Syntiant 1,2,3
Halol, Gujarat MG Hector MG Gloster 1
Tbl 2: Implementation of AI in various car models in India

9.0 Challenges with Autonomous Cars

Fully independent vehicles (level 5) are being tested in several international packages, but
none are currently available to the general public. We are still far from that. Challenges range
from technical and legal to environmental and philosophical. Here are some of the unknown
ones.
9.1 Cyber Security Threats to Autonomous Cars
 Remote Hijack- The hacker can hijack the car anytime and anywhere.
 Private Information Access- The hacker can steal all the information once he gets
access, leaving the product unable to use.
9.2 Ethical issues
 Artificial vs. Emotional Intelligence
Human drivers rely on subtle clues and non-verbal communication - such as eye contact with
pedestrians or learning the facial expressions and body language of other drivers - to make
instant judging calls and guess the behavior. Will driverless cars be able to duplicate this
connection? Will they have life-saving emotions as human drivers?
 Accident Liability
Who has to deal with accidents caused by driverless vehicles? Manufacturer? A human
passenger? Recent plans suggest that a fully standard Level 5 car will not have a dashboard or
24
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
steering wheel, so a human passenger would not have the opportunity to control the car in an
emergency.
If at times, any mishap occurs, such as accident, some component failure, unpredictable
stopping of car, who should be held responsible for the same. There are four agents in this
system:
o Driver
o Manufacturer (OEM)
o Owner
o Retail

This is a hot issue going on in-hand with the development of driverless cars. Technically
moving, falters occur due to driver’s carelessness. So, if, in case, component failure occurs, it
is the driver who is responsible for the situation and he should handle the same.
FGAI4AD, as discussed above, focuses on the behavioral evaluation of AI when used on
roads. Because road injuries are leading cause of deaths of children. FGAI4AD standardizes
the activities for use of AI in automotives. Using its methodology, components, structure and
various paradigms it covers, we can reach to the conclusion of this ethical issue.
9.3 Others
 Lidar and Radar
Lidar is expensive and still trying to find the right balance between distance and solution. If
more private cars were to drive on the same road, would their lidar signs interfere with each
other? And if more radio waves are available, will the frequency range be sufficient to
support mass production of private vehicles?
 Weather Conditions
What happens when a driverless car travels in heavy rain? When there is a layer of snow on
the road, the road dividers disappear. How will cameras and sensors track line marking if
marking is hidden by water, oil, ice or debris?
 Traffic Conditions and Laws
Will private cars have a problem with the tunnel or bridges? What are they going to do with
the huge traffic that will bounce? Will private cars be returned to a specific line? Will they be
given carpool road access? And what about legacy ships that share the road for the next 20 or
30 years?
 State vs. Federal Regulation
The regulatory process in the U.S. has recently changed from state directives to state-owned
driverless car licenses. Some states have even proposed a tax on driverless cars to prevent the
rise of "zombie" cars traveling around without passengers. Lawmakers also drafted bills
suggesting that all driverless cars should be zero-emission vehicles and be fitted with a shock
button. But will the laws vary from government to state? Will you be able to cross state lines
in a private car?

25
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
What solutions does US have for autonomous cars?
Today's cars have 100 million lines of code. Tomorrow’s private cars will have more than
300 million lines of code, so cyber security is a growing problem. Synopsys, a US leader in
app security testing and software design analysis, helps automotive customers build security
in their software throughout the life cycle of development and in every transaction.
Synopsys also offers a comprehensive portfolio of automated IP, certified by ISO 26262 and
ASIL B & D readiness, to help its customers have excellent chip applications for applications
such as ADAS, infotainment, and standard MCUs. Embedded vision processor solutions help
customers integrate skills such as object recognition and facial recognition, night vision, and
flexible cruise control.

26
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
Conclusion
Artificial intelligence is transforming the automotive industry. Significant investments have
been made in technology leaders such as Google, Tesla, Uber, and major car companies in
the region. It is not an opportunity for big companies but also a start. It will not only help
modernize the automotive area, but also provide safety to the passengers. This document has
served as an introduction to the integration of AI, IoT, and automotive industry. Vehicle
communication has been discussed in detail and is being greatly stressed by OEMs to develop
a safe and standard product. Although security, privacy issues and few challenges are here
but AI in automotive sector is achieving limelight and will grow in upcoming years.

27
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
Glossary
ADAS Advanced Driver Assistance System
AV Autonomous Vehicles
DSRC Dedicated Short Range Communication
ETSI European Telecommunications Standards Institute
GPS Global Positioning System
HLEG High Level Expert Group
IEC International Electrotechnical Commission
IEEE Institute of Electrical and Electronics Engineers
ISO International Organization for Standardization
ITU International Telecommunications Union
LAN Local Area Network
LTE Long Term Evolution
MCU Micro Controller Unit
OEM Original Equipment Manufacturer
SDG Sustainable Development Goals
SAE Society of Automotive Engineers
TB Technical Body
UNECE United Nations Economic Commission for Europe
V2R Vehicle to Roadside Unit
V2S Vehicle to Sensor
VANET Vehicular Ad hoc Network

28
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
References
1. https://blog.netapp.com/artificial-intelligence-in-the-automotive-industry/
2. https://www.infopulse.com/files/images/four-essential-control-points-in-adas.jpg
3. https://media.springernature.com/original/springer-
static/image/chp%3A10.1007%2F978-3-030-18963-
1_2/MediaObjects/477461_1_En_2_Fig17_HTML.png
4. https://www.mdpi.com/futureinternet/futureinternet1100070/article_deploy/html/imag
es/futureinternet-11-00070-g001.png
5. Artificial Intelligence across Industries IEC
6. https://www.researchgate.net/profile/Sofiane-
Ouni/publication/325222342/figure/fig5/AS:627622287536135@1526648222641/Act
ivity-Diagram-for-Ambulance-Vehicle.png
7. https://www.researchgate.net/figure/Observation-for-minimum-lateral-distance-from-
lead-vehicle-during-emergency-collision_fig7_328690797
8. https://www.researchgate.net/figure/Collision-avoidance-with-the-use-of-an-
intersections-infrastructure-is-possible-in-time_fig1_282365837
9. https://upload.wikimedia.org/wikipedia/commons/thumb/8/87/Audi_A8_2013_%281
1209949525%29.jpg/1200px-Audi_A8_2013_%2811209949525%29.jpg
10. https://global.toyota/pages/release/305812/0226_0.jpg
11. https://cdn.usharama.edu.in/blog/AHS-I/AHS-01.jpg
12. https://i2.wp.com/semiengineering.com/wp-
content/uploads/2019/08/Synopsys_functional-safety-afterthought-airbag-
fig1.jpg?fit=1224%2C416&ssl=1
13. https://filecenter.deltaww.com/solutions/images/02/0201/Smart_Traffic_EN.jpg
14. https://www.tn.gov/content/tn/tdot/long-range-planning-home/air-quality-
planning/transportation-demand-management-
tdm/jcr%3Acontent/content/tn_textandimage/image.img.jpg/1508345836119.jpg
15. http://www.isifast.org/uploads/6/1/7/3/61737407/6786148_orig.png
16. https://cdn.guidingtech.com/imager/assets/193369/Here-Vs-Google-Maps-
20_7c4a12eb7455b3a1ce1ef1cadcf29289.png?1524143632
17. https://s1.cdn.autoevolution.com/images/news/gallery/how-portable-emissions-
measurement-systems-work_3.jpg
18. https://www.researchgate.net/profile/Robert_Heinrich3/publication/314522008/figure/
fig5/AS:656952749199361@1533641149440/Component-model-of-the-evolved-
parking-lot-management-system.png
19. https://ars.els-cdn.com/content/image/1-s2.0-S0967070X11000503-gr2.jpg
20. https://files.readme.io/ee77696-vehicle-changed-flow.svg
21. https://www.researchgate.net/publication/307914301/figure/fig1/AS:40437358793932
8@1473421584389/a-Safety-route-guidance-system-using-participatory-sensing-
bAn-example-of-generated.png
22. https://cdn-uym.ibb.gov.tr/cdn/uym/documents/Resim3_EN.JPG
23. https://i0.wp.com/yatebts.com/wp-content/uploads/2015/08/iot_pub_trans_2015-8-
27_version2-2_large.png?fit=840%2C480&ssl=1
24. https://www.aubreydaniels.com/sites/default/files/xVehicleSafety-
Figure1.jpg.pagespeed.ic.vmzrTeiX-4.jpg
25. https://roadsafety.piarc.org/sites/safety/files/public/wysiwyg/import/planning-design-
operation-0/F10.6-Sequence-of-road-safety-checks-during-the-design-stages.png
26. “Security threats and requirementsanalysis for IOV” by Liang Wei

29
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC
 27. https://image.slidesharecdn.com/mis10-electronicpaymentsystem-110216032512-
phpapp01/95/mis-10-electronic-payment-system-22-728.jpg?cb=1297826774
28. https://www.researchgate.net/profile/Xiao_Yun_Lu/publication/261162167/figure/fig
1/AS:296850914594818@1447786181306/Control-System-for-Multiple-Vehicle-
Collision-Avoidance-and-Impact-Mitigation.png
29. https://www.yelowsoft.com/wp-content/uploads/2018/10/taxi-booking-software-fleet-
operation.png
30. https://ops.fhwa.dot.gov/publications/manag_demand_tis/images/works.jpg
31. RECOMMENDATION ITU-R M.1890-1 Operational radiocommunication
objectives and requirements for advanced Intelligent Transport Systems (Question
ITU-R 205-5/5) (2015-2019)
32. “Vehicles to Everything Communications and Services on 5G Technology” by Eugen
Borcoci
33. https://argus-sec.com/etsi-intelligent-transport/
34. https://whatis.techtarget.com/definition/SAE-International
35. https://www.digi.com/blog/post/introduction-to-smart-transportation-benefits
36. https://www.hindawi.com/journals/mpe/2014/192175/

30
Study Paper on “ AI in Automotives ”, Smart Network Division, TEC