Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous...

20
i Design of Communication Systems for Future Autonomous Vehicles Oliver Johnson u5570104 Executive Summary Systems engineering techniques were used to design communications for future autonomous vehicles (driverless cars). It was identified that our current transport system needed a paradigm shift to increase convenience, increase safety and reduce traffic congestion. Idea generation showed a number of ways in which the autonomous vehicle communications could be designed to improve the current transport system. Requirements mapping showed that incorporating vehicle-to-vehicle communication could affect the design requirements relating to a general autonomous vehicle system. A system interface was used to select the most appropriate method of vehicle-to-vehicle communication, which was dedicated short range communications. The final recommendations were develop common communication standards, use short range communication, transmit only the most pertinent information and develop an efficient mobile application.

Transcript of Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous...

Page 1: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

i

Design of Communication Systems for Future Autonomous Vehicles

Oliver Johnson

u5570104

Executive Summary

Systems engineering techniques were used to design communications for future autonomous

vehicles (driverless cars). It was identified that our current transport system needed a paradigm

shift to increase convenience, increase safety and reduce traffic congestion. Idea generation

showed a number of ways in which the autonomous vehicle communications could be designed

to improve the current transport system. Requirements mapping showed that incorporating

vehicle-to-vehicle communication could affect the design requirements relating to a general

autonomous vehicle system. A system interface was used to select the most appropriate method

of vehicle-to-vehicle communication, which was dedicated short range communications. The

final recommendations were develop common communication standards, use short range

communication, transmit only the most pertinent information and develop an efficient mobile

application.

Page 2: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

ii

Summary of Systems Techniques Used

Needs and Opportunities

Desktop Research Section 1.1

Problem Framing to support concept generation

Section 3.1

Problem Scoping

Stakeholder Analysis Section 2.1

Influence-Interest Grid

Section 2.1

Idea Generation

Concept Generation

Section 3.1

Requirements Analysis

Technical Performance Measures Section 3.4

House of Quality

Section 3.4

Logic and Functional Analysis

Logical Flow Section 3.7

Enhanced Functional Flow

Section 4.3

System Architecture

Subsystem Interface

Section 4.1

Testing, Validation and Evaluation

Weighted Evaluation

Section 5.1

Design Communication

Graphical Abstract Front page

Analysis Section 6.1

Page 3: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

iii

Table of Contents

1 The paradigm change needed for transport systems ................................................. 1

1.1 The Need for Effective Autonomous Vehicles ........................................................ 1

1.2 The Opportunity to Design Communication Systems ............................................. 1

2 Where are we heading? ................................................................................................ 2

2.1 Interaction between stakeholders ............................................................................. 2

2.2 Common international standards ............................................................................. 2

3 Development of Solution............................................................................................... 3

3.1 Concept Classification Tree ..................................................................................... 3

3.2 Possible Avenues of Design .................................................................................... 3

3.3 System Boundary Chart ........................................................................................... 4

3.4 Analysis of customer requirements and design requirements ................................. 4

3.5 Outcomes from the House of Quality ...................................................................... 6

3.6 Concept Refinement ................................................................................................ 7

3.7 Design of Mobile application – Logical Flow ......................................................... 7

4 Solution Proposal – Dedicated Short Range Communications ................................ 9

4.1 System Interface – Analysis of Different Options for Communications ................. 9

4.2 The Need for Low Latency .................................................................................... 10

4.3 Enhanced Functional Flow Block Diagram for DSRC ......................................... 10

5 Evaluation .................................................................................................................... 11

5.1 Weighted Evaluation ............................................................................................. 11

5.2 Discussion .............................................................................................................. 12

6 Design Communication .............................................................................................. 13

6.1 Graphical Abstract ............................................................................................. 13

7 Summary of Recommendations ................................................................................. 13

7.1 Develop common communication standards ......................................................... 13

7.2 Short range communication (DSRC 5.9 GHz) ...................................................... 13

7.3 Only transmit safety warnings and route information ........................................... 13

7.4 Develop efficient smartphone as part of the AV system ....................................... 14

8 Further Work .............................................................................................................. 14

9 Reflection ..................................................................................................................... 14

10 Bibliography: ............................................................................................................... 16

Page 4: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

1

1 The paradigm change needed for transport systems

1.1 The Need for Effective Autonomous Vehicles

Autonomous vehicles are often stated to be able to have the ability to ease road congestion,

reduce travel time, provide safer roads and increase convenience (KPMG 2013, Fagnant &

Kockelman 2013). Road congestion will be able to be reduced by decreasing the number of

cars on the road and decreasing the distance between cars on the road. AVs also have the ability

to increase the use of ridesharing, further decreasing road congestion. Accident rates are able

to be reduced, as they will eliminate human error, which accounts most vehicle crashes.

1.2 The Opportunity to Design Communication Systems

The design of appropriate communication systems represents an opportunity to further increase

the safety and convenience of autonomous vehicles. The aspects of the communication systems

for autonomous vehicles that are to be considered are the inter-vehicle communication and the

development of effective ways for the user to interact with the AVs and the company providing

them.

In a future autonomous vehicle network, there may be many different types of vehicles

occupying the road. There would be the traditional (human driver) vehicles, personally owned

autonomous vehicles (AVs) and shared autonomous vehicles (shared AVs) participating in a

sharing scheme where users can rent a vehicle from a pool of cars whenever is required

(Fagnant & Kockelman 2013). In both the personally owned AV case and shared AV case, the

vehicles would need to be connected to a company that would be able to set up routing and

coordinate information sent to the user. The prospect of many companies interacting in the

autonomous vehicle setting may cause difficulties for vehicle to vehicle communication, and

the general operation of autonomous vehicles.

In particular, the ‘client’ in the project was identified to be people in the general public. In

particular, there is interest in the use of driverless vehicles to increase mobility for people who

are unable to drive (Hull & Hymowitz 2016). In many ways, the client’s needs would be met

by designing appropriate communications systems for the car.

Page 5: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

2

2 Where are we heading?

2.1 Interaction between stakeholders

A stakeholder map was overlayed onto an influence-interest grid to demonstrate the

relationships between the key stakeholders in the future autonomous vehicle system. Two

blocks are shown for the Other AV Companies, as there are two options as to how they are

perceived by the AV Company in question. The first view (left, greyed out) represents the case

where other AV companies are regarded as direct competitors, leading to little collaboration.

The second case, of communication and cooperation between the other AV companies, is

shown on the right. In this case the other AV companies are viewed as high interest. A higher

functioning autonomous vehicle system is able to be developed through cooperation between

different AV companies, as superior safety can be introduced into the system through the

development of international standards.

Figure 1. Stakeholder Analysis: Two options for the interest value of the Other AV Companies

are shown, where the option on the right is preferred.

2.2 Common international standards

By viewing other AV companies as collaborators, it is possible that common international

standards for communications between vehicles could be developed. Without common

protocols, the types of information being sent to neighbouring AVs would be ill-defined.

Common standards (possibly through ISO) are an essential requirement to ensure reliable and

effective inter-vehicle communication.

Page 6: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

3

3 Development of Solution

3.1 Concept Classification Tree

Ideas were classified in the concept generation tree as follows for the design of

communications systems for autonomous vehicles. The problem was framed as a “how might

we” statement.

Figure 2. Concept Classification

3.2 Possible Avenues of Design

The concept classification tree identified possible avenues to pursue in regard to the design. It

is noted that from the stakeholder map, it was recommended that international standards be

developed for communication. This is especially important for the vehicle-to-vehicle

communication aspect of the design. There were two possible design concepts to consider for

the vehicle-to-vehicle communication aspect of the design. These were the long range and short

range options for communication, and it was recommended that these both be taken to the next

stage of the design process for further analysis. The governing thought behind vehicle-to-

vehicle communication is that connected vehicles are predicted to increase road safety, as

warnings are able to be sent in case of imminent hazards and potential crashes (Karagiannis et

al. 2011; Biswas et al. 2006). From Figure 2, it is evident that the technical aspects and the

Page 7: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

4

human aspects of AVs can largely be separated. However, some ideas of the short range

communication were able to be discarded, such as infrared and Li-Fi, as they both require line

of sight to work, and therefore would not give the kind of range required for a vehicle

communication system.

The mobile application and entertainment concepts are also promising, as they could allow the

communication system to be used dually for the marketing of the autonomous vehicle to

potential customers through convenience and entertainment possibilities.

3.3 System Boundary Chart

A system boundary chart was produced for the inter-vehicle communications, to set the scope

of the portfolio. It is noted that many of the other aspects of autonomous vehicles have been

excluded, as while they are worthy of study, they do not directly relate to the design of

communication systems.

Internal External Excluded

Vehicle

AV Company

Methods of Inter-vehicle

communication

Increasing ease of use for

the customer

Government

Other AV companies

Passenger

Pricing of AV trips

Carbon emissions

Details of the artificial

intelligence in AVs

Wait time

Physical design of AV

Details of AV sensors

Table 1. System Boundary Chart

The system boundary chart demonstrates that the focus of the portfolio will be on the

communication aspects of the autonomous vehicle. Technical aspects such as the artificial

intelligence, physical design and sensors of the AV have been excluded from the scope.

3.4 Analysis of customer requirements and design requirements

The customer requirements of a passenger for an autonomous vehicle system were

determined and ranked through a pairwise analysis technique. The ranks of the customer

requirements are shown in Table 2: Technical Performance Measures. Howard and Dai

(2013) found that people found the safety benefits of AVs most attractive. The design

requirements in the technical performance measures table were specifically chosen to be

related to the communications aspects of the AV, rather than the AV system as a whole.

Page 8: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

5

Customer

Requirement Rank Design Requirements Metric Direction

Safe transport 1

Number of accidents Accidents/km ↓

Hazard rate Hazards/sec ↓

Update frequency Hz ●

Amount of data sent MB/sec ↑

Reliability 2

Latency Milliseconds ↓

Hazard rate Events/second ↓

Data rate utilised MB/sec ↓

Fast Transport 3

Latency Milliseconds ↓

Number of accidents Accidents/km ↓

Processing power required Calculations/sec ↓

Reduce Traffic

Congestion 4

Latency Milliseconds ↓

Hazard Rate Hazards/sec ↓

Table 2. Technical Performance Measures: Mapping of customer requirements to design

attributes.

According to the pairwise analysis, the most important customer requirement is safety, which

is unsurprising, given that safety is often heralded as the main benefit that driverless cars will

provide (Fagnant & Kockelman 2013; Litman 2015). However, since driverless cars are still in

the prototype phase, it is unclear how great the benefits will be. The design attributes were

compared in the following house of quality (Table 3).

Page 9: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

6

Table 3. House of Quality (HoQ) for autonomous vehicle communications systems

3.5 Outcomes from the House of Quality

Processing power required vs. amount of data sent to other vehicles

The amount of data that is sent to neighbouring vehicles is a crucial design attribute for the

vehicle-to vehicle system. If a low amount of data were to be sent, only the most important

safety information would be included. For example, AVs would only alert other vehicles if

there is an imminent collision. A high amount of data would correspond to sending all of the

sensory information to nearby AVs to complement their own sensors and increase their field

of view. From the house of quality, it is thought that having a very high amount of data sent

between vehicles would increase the processing rate required. The AV may not be

computationally powerful enough to process many other nearby AVs sensory information, and

would result in minor safety improvements.

Hazard rate vs. accident rate

Page 10: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

7

The house of quality demonstrated that the safety and the hazard rate of the vehicle were highly

related. Thus the best way of improving the safety may be to consider the reliability and hazard

rate as an avenue for improvement.

Amount of data sent to other vehicles vs. data rate needed/utilised

Designing a vehicle-to-vehicle communications system that requires high amounts of data

transfer between vehicles would also require a high rate for the hardware of the chosen

communications system. A communication system that only uses a low data rate is desirable

as cost will be reduced and wireless communication systems may not be able to cope with the

high volumes of sensory data transmitted.

Latency

Latency is identified through the House of Quality as positively correlated with most other

design requirements and all of the customer requirements. It can therefore be recommended

that latency should be reduced in the design of the vehicle-to-vehicle communication system.

3.6 Concept Refinement

After completing the requirements analysis, the concepts were ready to be refined in order to

develop a proposal for the communications systems of vehicles.

In concept generation it was suggested that Wi-Fi, Bluetooth and Cellular could be investigated

for the development of a communications system for autonomous vehicles. However, several

of these ideas can be ruled out. Wi-Fi does have advantages in that it has a high data rate and

low latency. However, its mobility support is very low (Malla & Sahu 2013), so it is

inappropriate to be used for vehicle-to-vehicle communications. Bluetooth offers a high

number of connectivity options, but its maximum range is only 100m (Bandiwadekar & Dale

2015). The aim of vehicular communication is to provide a range of up to 1km (Biswas et al.

2006), so Bluetooth is undesirable.

3.7 Design of Mobile application – Logical Flow

The design of the mobile application, where the user interacts with the phone to order the AV,

forms a part of the overall communication system. A mobile application is almost certainly

something that future AV companies will include in their system. Moreover, effective design

of the phone application would lead to higher customer adoption rates for the AV system. There

are multiple ways in which the mobile application could work, but the system that is being

considered is one where customers rent from a pool of AVs whenever they need them.

Page 11: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

8

Figure 3: Logical Flow for typical ride in a shared AV with mobile application. The column

categories are the main parts of the commute. The squares represent processes, the diamonds

represent decisions and the rounded off squares represent terminals.

It is noted that the passenger’s phone does not communicate directly with the AV. This will be

shown more clearly in the system interface. This is because the inter-vehicle communication

system is set up solely for the sending of data between other AVs, and interference with

personal devices would be undesirable. Instead, the user’s phone communicates with the AV

company, which then coordinates the movements of the AV.

This logical flow demonstrates that a mobile application in conjunction with the rest of the

shared AV system is essential. One aspect to note is that although there are many steps in this

procedure, many of them happen very quickly, or while the passenger is in the shared AV,

implying that the total journey time is not excessive. Comparing to a traditional transport

system, the AV logical flow does not include any steps such as parking cars or storing bicycles,

contributing to a further decrease in wait time. The decision whether to ride-share or not is an

interesting addition, as potentially implementing this could decrease traffic significantly. It

would be interesting to determine at what price decrease users would be most likely to slightly

increase their wait time by including other passengers in their car.

In the first column of the logical flow, there is a question to the user asking if the wait time is

acceptable. If the wait time is too high, then users would be turned off using the service. It is

recommended that there is some kind of efficient search algorithm to find the nearest AV to

the customer.

Page 12: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

9

4 Solution Proposal – Dedicated Short Range Communications

A communications system solution was proposed to further increase the safety of

autonomous vehicles. The chosen system is a dedicated short range communications (DSRC)

system, which is a specialised communications designed for vehicles (Sichitiu & Kihl 2008).

4.1 System Interface – Analysis of Different Options for Communications

It was shown from the house of quality (sect. 3.4) that allowing AVs to share sensors would

decrease the number of accidents, a crucial design requirement. Sharing sensors would allow

nearby AVs to in know each other’s planned route, increase the field of view of the AV and

alert AVs to possible hazards on the road. Figure 4 shows the chosen method as a system

interface, investigating the dynamics of multiple AV companies communicating and sharing

data in order to further increase the safety of AVs.

Figure 4: System Interface: Short range AV communication achieved by DSRC

The subsystem interface of an autonomous vehicle system, with short range communication is

shown in Figure 4, where information is transmitted directly between AVs. The two main

subsystems are the AV and the Central Control System. The user interacting with the car

through the mobile application is also displayed.

The second option considered for communication between AVs was utilising the cellular

connection between the AV and the AV provider. The central control system searches for

nearby AVs via GPS and relays their sensory information to the AV in question. This achieves

similar outcomes to the system shown in Figure 4, except via the central control system.

Page 13: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

10

4.2 The Need for Low Latency

A further option that was considered was to use a cellular connection to distribute data between

vehicles. The problem with the cellular network option is the latency is too high. Average

latency for 4G-LTE networks was 176 milliseconds (Westwood 2013). Karagiannis et al.

(2011) identified the critical latency for collision warning to be 100 ms, but a value less than

this is highly desirable. Therefore 4G networks do not offer the performance required for

transmitting safety information. In addition, the bandwidth for 4G networks may not be high

enough for sending high volumes of sensory data.

However, the low range option would allow for low latency communication between

neighbouring AVs. Latency for dedicated short-range communications in vehicle-to-vehicle

communication has been shown to be less than 20 milliseconds (Yang 2004), which is well

below the 100 millisecond threshold. However, a limitation is that as the AV drives, it would

continuously have to disconnect and reconnect with differing AVs and for security purposes,

each AV would have to authenticate the connection between AVs. However, via a cellular

connection, the AV company would keep a continuous cellular connection to all of its vehicles,

so there would be no need to reconnect and disconnect from AVs. The network would also

require high mobility support, hence why Wi-Fi would be a poor choice.

From problem scoping (sect 2.1), it was determined that we should consider the other

‘competitor’ AV companies as high interest, as we are seeking to utilise their sensory

information. However, without proper cooperation, companies may not develop compatible

communication systems. A recommendation would be to develop an international standard

(through ISO) for communication between AVs.

In fact, the frequency band of 5.9 GHz has already been allocated for dedicated short range

communications in vehicles in Europe and North America (Karagiannis et al. 2011). While this

is a start, when autonomous vehicles are finally commercial, there will need to be negotiation

between providers concerning how much and which information to send neighbouring

vehicles.

Overall, the system interfaces have demonstrated that the low range communication option

(Figure 4) would be more effective, due to its simplicity, and the reduced computational power

and bandwidth required for the central AV control system.

4.3 Enhanced Functional Flow Block Diagram for DSRC

An enhanced Functional Flow Block Diagram (eFFBD) differs from a conventional FFBD, in

that an eFFBD specifies data flow (Seidner & Roux 2008). A conventional FFBD displays

the functions in a system, and the order in which they are performed (Long 2002). However,

an eFFBD displays both the functions and the flow of data in between functions in the

system, resulting in a more complete representation of the system (Seidner et al. 2007; Long

2002).

Page 14: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

11

Figure 5: eFFBD for Hazard transmission system. Key: square boxes – functions, rounded

boxes – data flow.

The eFFBD demonstrates the central role that data plays in the design of the DSRC system.

In particular, the control system must sift through many different forms of data every second

to make a decision about a possible safety hazard. It also demonstrates that there are many

ways in which disruptions to this data flow could occur, leading to reliability and security

issues.

It is an important recommendation that the communications between autonomous vehicles

must be secure. Malicious attacks have the potential to be deadly, in part because there is no

human driver on standby to intervene (Gerla et al. 2014). Interferers could pose as other

driverless vehicles with intent to cause the AVs to malfunction. Because the communication

range in DSRC is relatively large (up to 1km), potential for threat is larger than a short range,

line of sight communications (Petit & Shladover 2015). However, there must be allowance

within the system for organisations to intervene, should the AV become out of control or

suffer from an attack. A possibility would be to use two-factor authentication (for example,

the AV company and a regulatory body) to gain access to the AV in such a situation (Lee et

al. 2014; Gerla et al. 2014).

5 Evaluation

5.1 Weighted Evaluation

A weighted evaluation was carried out comparing the current Canberra transport, the original

AV system design and the design described earlier with sensor sharing. The customer

Page 15: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

12

requirements were taken from requirements analysis, where a pairwise analysis was

undertaken.

The weighted evaluation was selected to be most appropriate for the evaluation, since it

considered the importance of all of the customer requirements. An evaluation meets the

benchmark matrix was not included, since there is not real benchmark that is trying to be

achieved, it is just improvement that is sought. Perhaps the only relevant benchmark that could

be included is that the AV system is equal for an improvement on each criteria, comparing to

the current Canberra transport system.

Rank Importance

Current

Transport

Original

AV Design

Proposed

AV Design

Customer

Requirements R I x R R I x R I I x R

Safe Transport 1 5 3 15 4 20 5 25

Reliable 2 4 4 16 4 16 4 16

Fast Transport 3 3 3 9 3 9 3 9

Reduce Traffic 4 2 3 6 4 8 4 8

Enjoyable 5 1 4 4 4 4 5 5

Total 50 Total 57 Total 63

Table 4. Ranking matrix

5.2 Discussion

It has been estimated that with 50% driverless car penetration, the crash rate will reduce by

around 37.5% (Fagnant & Kockelman 2013), providing the evidence to increase the rating of

safe transport from 3 to 4 comparing Canberra’s current transport system with a shared

autonomous vehicle system. Fagnant and Kockeleman (2013) also estimated that traffic

congestion will decrease by around 35% at the same market penetration, providing the evidence

for the traffic reduction. There is much uncertainty in these statistics, as no current autonomous

vehicle system exists currently, and the data is based on conditions in the United States.

The only difference in rating when comparing the original AV design and the AV design with

sensor sharing is that safety has a rating of 4 and 5 respectively. This is because it is expected

sharing the sensors should allow the driverless cars to have a larger field view, noticing possible

accidents much earlier and thus avoiding them. In particular, vehicle-to-vehicle

communications seek to reduce the rate of intersection collisions, which account for around 50

percent of urban area crashes (Transportation Research Board 2003). Assuming that

autonomous vehicles with sensor sharing can reduce this by a conservative 30%, this

calculation provides the justification to raise the safety from 4 to 5 stars.

Overall, the evaluation matrix shows that both AV system designs are an improvement on the

current transport system, and the proposed design with sensor sharing has a slightly higher

rating than the original AV system design. However, it is acknowledged that there is much

uncertainty in the calculations for the ratings.

Page 16: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

13

6 Design Communication

6.1 Graphical Abstract

A ‘graphical abstract’ (on front page) was used in order to communicate the design to a general

audience. In the graphical abstract, the red car notices the pedestrian hazard with its camera.

Since the pedestrian is out of the field of view of the black car, it relays information wirelessly

to it, alerting it that there is a possible hazard nearby. This is basic principle of the

recommended system as outlined in the system interface.

A 2-D image was chosen to visually communicate the design due do its simplicity. However,

the design could equally be displayed in a variety of other forms such as a video or 3D image.

For example, a video could include a possible crash about to occur, but then with the sharing

of sensors between adjacent cars, the crash is avoided. Visual representations of the data

moving to different vehicles could be included to demonstrate how the accident was avoided

by the sharing of data. The video could be framed as an advertisement for this kind of

technology.

A technical drawing of the electronics and wireless receivers/transmitters could also be used,

which would useful to communicate the design to an engineering/technical audience, but would

be less useful for the communication to a general audience. Focussing on the results is more

important for general audience, as this is shown on the front page.

7 Summary of Recommendations

7.1 Develop common communication standards

By considering other AV companies as high interest and collaborators in the quest to increase

vehicle safety through vehicle-to-vehicle communications, international standards can be

developed to ensure compatibility between autonomous vehicles of different manufacturers.

7.2 Short range communication (DSRC 5.9 GHz)

Dedicated short range communications (DSRC) were recommended to enable sensory

information to be transmitted between nearby autonomous vehicles. It was found that using

longer range options, such as cellular networks, would offer latencies too great for the

transmission of safety information.

7.3 Only transmit safety warnings and route information

Whilst transmitting all of the sensory data to nearby autonomous vehicles would increase their

field of view, the data rate and processing required for this would be too great, and would

decrease reliability. It is therefore recommended that only information pertaining to critical

accident warnings and route information, such as indication, changing lanes and moving

through intersections, be transmitted between vehicles. The transmission of high volumes of

Page 17: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

14

sensory information would also difficult for other AV companies to process unfamiliar

information.

7.4 Develop efficient smartphone as part of the AV system

Well-designed phone applications were also identified through the logical flow as crucial for

reliability and passenger convenience.

8 Further Work

8.1 Conduct further research into allowing common communication methods between

providers

8.2 Conduct experiments for dedicated short range communication to test performance.

8.3 Market autonomous travel to consumers through entertainment possibilities. By

designing appropriate communications systems, passengers can be entertained through

the sharing of content between autonomous vehicles, increasing their appeal to

prospective future customers. This would operate much like a BitTorrent scheme except

for the car environment (Gerla & Kleinrock 2011; Lee et al. 2007).

9 Reflection

If I were to undertake my portfolio differently next time, I would like to carry out some

simulation and testing of my ideas. This was not possible in the time that I had available, and

I did not possess sufficient technical knowledge of communications systems to implement such

a system. I feel I have gained three sets of skills from completing this portfolio. Not only have

I learnt to apply the systems tools, but I also feel more comfortable in understanding the

research areas of autonomous vehicles and vehicular networking, which I think will be

beneficial for my future career. Even though this course is not content based, I have

inadvertently become very familiar with the theory of autonomous vehicles, and have become

very excited about their eventual release. The third set of skills I gained was an understanding

of how to use the systems tools. I think that after completing the group project, applying the

systems tools was much easier, and I could see their usefulness more easily. I was initially

sceptical about how the tools seem to apply well to the problem, however I think the test of a

good systems engineer is to be able to apply the tools in places outside their usual scope.

Perhaps more innovative insights can be gathered from using the analysis techniques unusually.

I found that when I was writing my portfolio, nearly all of it had to be rewritten, as I could not

use my TCs directly in the portfolio because of a lack of flow. However I did not think that the

research and work I did for my TCs was wasted, as I gained a wider understanding of the

problem. My original idea was to transmit all sensory information to neighbouring AVs, but

after progressing through the design process and researching the vehicular communications

field, it was clear to me that this was not going to work.

Page 18: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

15

My peer reviews gave some good feedback. An interesting aspect that I noticed was the two of

the reviewers liked the logic and function section the most, whereas I thought it was fairly weak

and generic, and was originally planning to change it before the final submission. I followed

their advice, and kept the logical flow diagram about the mobile application in the portfolio, as

I realised apart from conveying my design, it gave the readers more background into how an

AV system might function.

My first reviewer gave comments that were very blunt and short, but still useful. The advice to

include more evidence was used in the final report, as in some places I had just made some

general statements. I sought to change this for the final portfolio submission. However, I think

that when this reviewer phrases their sentences, critique of my work could be phrased in terms

of “this could be improved by” rather than “is extremely weak”.

As several of the reviewers noticed, my conclusion on my original report was too short and not

very convincing. In my final report, I decided to expand on the recommendations part to make

it very clear to the reader the important aspects of my final design. The other reviewer stated

that I could omit the conclusion if I made the recommendations section detailed enough. I

decided to follow this advice of the second reviewer, and I think that it increased the clarity of

my portfolio, as before the conclusion and the summary of recommendations displayed the

same information.

The idea of investigating vehicle-to-vehicle communications was one that I had been pondering

for several months. Going through the systems design process helped me understand the

problem more fully, and I think I will be able to adopt this way of thinking for other problems

I encounter, whether they be engineering or non-engineering.

Page 19: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

16

10 Bibliography:

Bandiwadekar, AS & Dale, MP 2015, “Implementation of Programmable Logic Controller

with Wireless Bluetooth Connectivity”, International Journal of Applied Research, vol. 1 no.

8, pp. 246-253.

Bansal, P, Kockelman, KM & Singh, A 2016 “Assessing public opinions of and interest in new

vehicle technologies: An Austin perspective”, Transportation Research Part C: Emerging

Technologies, vol. 67, pp. 1-14.

Biswas, S, Tatchikou, R & Dion, F 2006, “Vehicle-to-vehicle wireless communication

protocols for enhancing highway traffic safety”, IEEE Communications Magazine, vol. 44, no.

1, pp. 74-82.

Fagnant, DJ & Kockelman, KM 2013, “Preparing a Nation for Autonomous Vehicles:

Opportunities, Barriers and Policy Recommendations”, Eno Centre for Transportation.

Federal Highway Administration 2009, National Household Travel Survey. U.S. Department

of Transportation, Washington, DC.

Gerla, M, Lee, E-K, Pau, G & Lee, U 2014 “Internet of Vehicles: From Intelligent Grid to

Autonomous Cars and Vehicular Clouds”, 2014 IEEE World Forum on Internet of Things, pp.

241-246.

Gerla, M & Kleinrock, L 2011, “Vehicular networks and the future of the mobile internet”,

Computer Networks, vol. 55, no. 2, pp. 457-469.

Hull, D & Hymowitz, C 2016, “Google Thinks Self-Driving Cars Will be Great for Stranded

Seniors”, Bloomberg Businessweek, 2 March.

IHS Automotive 2014, Self-Driving Cars Moving into the Industry’s Driver’s Seat, IHS Inc

Karagiannis, G, Altintas, O, Ekici, E, Heijenk, G, Jarupan, B, Lin, K & Weil, T 2011,

“Vehicular Networking: A Survey and Tutorial on Requirements, Architectures, Challenges,

Standards and Solutions,” IEEE Communications Surveys & Tutorials, vol. 13, no. 4, pp. 584-

616.

Howard, D & Dai, D, “Public Perceptions of Self-driving Cars: The Case of Berkeley,

California, 93rd Annual Meeting of the Transportation Research Board.

Lee, KC, Lee, S-H, Cheung, R Lee, U Gerla, M 2007, “First Experience with CarTorrent in a

Vehicular Ad Hoc Network Testbed”, 2007 Mobile Networking for Vehicular Environments,

11 May 2007, Anchorage, AK, pp.109-114.

Page 20: Design of Communication Systems for Future Autonomous Vehiclesu3951377/student... · for autonomous vehicles that are to be considered are the inter-vehicle communication and the

17

Litman, T 2015, Autonomous Vehicle Implementation Predictions, Victorian Transport Policy

Institute.

KPMG 2013, Self-Driving Cars: Are We Ready?, KPMG LPP.

Lee, E-K, Lim, J, Joy, J, Gerla, M & Gadh, R 2014 “Multi-factor authentication and

authorization using attribute based identification,” UCLA CSD, Tech. Rep. 140003, 2014.

Long, J 2002, “Relationship between Common Graphical Representations in System

Engineering”, Vitech Corporation.

Malla, AM & Sahu, RK 2013, “A Review on Vehicle to Vehicle Communication Protocols in

VANETs”, International Journal of Advanced Research in Computer Science and Software

Engineering, vol. 3, no. 2, pp. 409-414.

Petit, J & Shladover, SE, “Potential Cyberattacks on Automated Vehicles”, IEEE Transactions

in Intelligent Transport Systems, vol. 16, no. 2, pp. 546-556.

Poczter, SL & Jankovic, LM 2014, “The Google Car: Driving Toward A better Future?”,

Journal of Business Case Studies, vol. 10, no. 1, pp. 7-14..

Seidner, C, Lerat, JP, Roux, OH 2007, “Usability of formal verification on EFFBD models:

Applying Petri nets to Systems Engineering issues”, INCOSE International Symposium, vol.

17, no. 1, pp. 848-859.

Seidner, C & Roux, OH 2008, “Formal Methods for Systems Engineering Behavior Models”,

IEEE Transactions on Industrial Informatics, vol. 4, no. 4, pp. 280-291.

Sichitiu, ML & Kihl, M 2008, “Inter-Vehicle Communications Systems : A Survey”, IEEE

Communications Surveys & Tutorials, vol. 10, no. 2, pp. 88-105.

Transportation Research Board 2003, “Unsignalized Intersection Collisions”, National

Cooperative Highway Research Program, Washington, DC.

Westwood, S 2013, “Global State of LTE Report”, OpenSignal.

Yang, X, Liu, L,Vaidya, NH & Zhao, F 2004, “A vehicle-to-vehicle communication protocol

for cooperative collision warning,” Mobile and Ubiquitous Systems: Networking and Services,

2004. MOBIQUITOUS 2004. The First Annual International Conference on, pp. 114-123.