Security and Privacy Issues for Autonomous Vehicles Networks using 5G Networks
_______________
1) Introduction
Fully autonomous vehicles(AVs) are rapidly getting closer to becoming a reality. Today's
smart cars are already equipped with functions such as Lane Keeping and Road Departure
Assists as well as Adaptive Cruise control wherein the car steers or brakes to maintain its
position on the road independently of the driver. The number of vehicles equipped with such
technology is predicted to rapidly increase, as indicated by figure 1. In fact, large companies
such as Tesla [1] and Honda [2] are attempting to extend these features to create a fully
automated vehicle that does not require a driver at all. It is only a matter of time before
such technology becomes a part of everyday traffic.
Fig 1. Predicting the Increase in the Number of AVs [3]
Given this imminent future, it is important to consider the possible challenges of implementing
and integrating such vehicles into today's traffic. This includes exploring efficient
network architectures and analyzing possible security and privacy threats. One of the most
popular two proposed architectures is enabled by blockchain using 5G Networks. This paper
will explore the pros of said architecture as well as attempt to provide solutions for the cons.
1.1 What are Autonomous Vehicles and why should we be interested?
Oxford dictionary defines an autonomous vehicle as a vehicle that can drive itself without
input from a human being. This could be applied to any vehicle including cars, trains, buses
and more.
​
Implementing this idea can provide a plethora of advantages. While it will improve the
comfort of transportation in general, it will also reduce dangers such as the possibility of
traffic accidents. In 2019, the cause of nearly 90% of car accidents was human error [4]. Using
driverless cars would remove the possibility of human error; elements such as distracted
and drowsy driving will no longer be a problem. Furthermore, the vehicle's ability to rapidly
and accurately analyze real-time information regarding traffic will allow it to, theoretically,
make better decisions. Additionally, since vehicular flow is coordinated, traffic congestion
and parking will become less of a problem.
​
For the public transportation sector, autonomous vehicles can increase efficiency and
open up endless possibilities as indicated by figure 2. Buses and trains will no longer experience
delays and less resources will be required to manage their operation. They could
be organised using the feeder system that automatically sends public transport to nearby
stations on-demand or the trunk system that operates over a larger radius and travels over
fixed paths [5] . Additionally, systems such as Robo-Taxis could be implemented wherein
owners of autonomous vehicles can rent out their cars as taxis for others to hire - a service
similar to Uber. If a person possessed multiple AVs they could rent out all of them simultaneously as they would only need to be controlled and monitored digitally.
Fig 2. Uses of AVs in the public transport system [7]
While the benefits that fully autonomous vehicles can provide seem to be endless, its
implementation is yet to be perfected. More specifically, the overarching concerns generally
fall under the category of security threats and network efficiency. Making a majority of
traffic digitally controlled introduces opportunities for cyber-attacks from malicious entities.
For instance, a hacker could intentionally cause an accident by overriding an autonomous
vehicle's mechanics or feeding it false information. If done with public transportation there
could be many casualties. Also, since the vehicles rely on traffic data, it will need to be
connected to a network that collects, stores, relays it. This network will have an extremely
large amount of nodes including roadside sensors, network towers and the vehicles themselves
among other devices. Also, given the complexity of the computations and the size of the
network, extremely large amounts of data will need to be transmitted at high speeds. To
support such a rapid system an appropriately efficient network is required. The following
sections will expand on these problems and detail solutions to them.
1.2 What is an Autonomous vehicle network (AVN)?
An autonomous vehicle network brings an upgrade to an autonomous vehicle. In AVN, an
autonomous vehicle will have an onboard control unit (OBU) that can communicate with
other road users such as people, vehicles, infrastructures via a network. The vehicle-to-everything (V2X) communications can be divided into vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), vehicle-to-pedestrian (V2P).
​
In figure 3, we can see data about traffic conditions, like traffic jams or traffic accidents. This data
is sent to the V2X network with the goal that drivers can know the data on street conditions
ahead of time and make an early move. Besides, emergency rescue can benefit from V2X
interchanges by sending warnings to the vehicles ahead to move to different routes. The
V2X network can also be used for finding food and drink close by, for example, discovering
corner stores and eateries.
​
The execution of vehicular networks faces numerous difficulties, particularly in security
and privacy issues. Without a legitimate verification convention, hackers may infuse unauthenticated messages to acquire individual advantages. For instance, a hacker communicates simulated traffic messages to have less traffic congestion by misdirecting vehicles in dffierent routes. Because of the wireless component of V2X networks, the on-the-air messages can be threatened by third-party influence. Therefore, security components should be a part of
V2X communications.
Fig 3. Uses of AVs in the public transport system [8]
In figure 4, a trust authority (TA), RSUs, Base Stations (BSs), and OBUs are significant
elements in vehicular organizations [9]. OBUs can speak with other OBUs and roadside
infrastructures within their correspondence range. Both BS and RSU are access points of the AVNs, accepting requests from the TA and gathering messages from OBUs. All roadside frameworks are associated with the TA through the wired link.
Fig 4. Uses of AVs in the public transport system [9]
The TA is the most important part of the AVNs. Every vehicle should enlist at the
TA with its identity data to join the network. On the off chance that a vehicle is found
acting suspiciously, a report will be created and conveyed to a close-by AP. The report
will be additionally sent to the TA. At that point, the TA can check this report and make
authoritative moves to check if the report is reliable, for example, renounce the suspicious
vehicle. When the TA chooses to repudiate a vehicle, it will add the character of the vehicle
into the certicate revocation list (CRL) and send an update solicitation to APs. APs
acknowledge this update and broadcast the refreshed CRL. OBUs inside the concurrence
range get the message and update their CRLs. APs and OBUs will dispose of the message
wherein the joined authentication is in the most recent CRLs. The essential objective of
APs is to transfer messages in vehicular networks. Because of the way APs work they are
more vulnerable to hackers.
2) Why 5G?
5G is the fifth generation standard for cellular broadband networks. It will be the successor to 4G and has been deploying throughout the world since 2019. It boasts features such as high bandwidth, low latency, and can allow download speeds up to 10 gigabits per second. The high bandwidth will allow large amounts of data to be delivered quickly and efficiently. Low latency will mean quicker response times so data being transferred back and forth will be fast. Because of 5G's ultrafast and low latency features, it will not only serve cellular technologies but will also likely serve laptop and desktop computers as well as enable many new IoT devices such as autonomous vehicles. There will be more devices connected to the internet than ever before and 5G has very fast speeds to enable this which makes 5G a prime candidate for expanding the IoT, especially in the realm of autonomous vehicle networks. Figure 5 shows a comparison of 4G and 5G. 4G would not provide the necessary speeds and low latency required for an autonomous vehicle network like 5G could.
Fig 5. The Comparison Between 4G and 5G [10]
2.1 5G Architecture for Autonomous Vehicles
5G architecture for autonomous vehicles expands on 5G architecture itself. Figure 6 is a visualization of what 5G autonomous vehicle architecture looks like.
5G vehicle network architecture contains three stratums, the vehicle, network and application stratum. Within the vehicle stratum, 5G enabled vehicles can communicate with each other using Dedicated Short Range Communication (DSRC) LTE-V-Direct or device-to-device (D2D) millimeter-waves. Vehicles can also collect information from traffic signals, pedestrians, and information from Road Side Units (RSUs) and Base Stations (BSs) contained in the network stratum.
Fig 6. An overview of 5G autonomous vehicle network architecture [11]
The network stratum is made up of these RSUs, BSs, and all the entities of the 3GPP core such as trusted authority, service providers and cloud. 5G enabled vehicles within this network are able to access the cloud and these services through the 3GPP core network. There are three functions within this network: control and management function (CMF), data function (DF), and the security and privacy function (SPF). CMF provides access and mobility management, policy control and authentication and authorisation services. DF provides packet forwarding. SPF provides security and privacy services while vehicles access the core network. There is also a trusted third party (TTP), that manages certificates and the identities of vehicles throughout many applications.
In the application stratum, there are many types of vehicle-to-everything (V2X) services such as vehicle-to-vehicle (V2V), vehicle-to-network (V2N), vehicle-to-infrastructure (V2I) and vehicle-to-pedestrian (V2P). V2V communication allows vehicles to exchange messages wirelessly with each other about their whereabouts in relation to each other. This can help vehicles avoid collisions and determine the threat of a crash with another vehicle. V2N allows vehicles to communicate with and gather information from the network stratum. For example, a vehicle could request updates about traffic, accidents and events occurring later on in the vehicle's journey through information the network provides. V2I enables communication between vehicles and road infrastructure. Through this, vehicles can communicate with road markings, signs and traffic signals to gather information about the surrounding road. Finally, V2P is how vehicles would communicate with pedestrians and non-vehicle road users such as cyclists. V2P allows vehicles to avoid these vulnerable road users. [11]
2.2 Why does 5G bring so many challenges?
Although 5G can uphold the stringent requirements for low latency and high speeds to enable a network of autonomous vehicles, the use of 5G poses many security challenges. There are a few reasons for this. The first reason is because 5G has higher bandwidths and enables many more IoT devices. Higher bandwidth and more IoT devices are great for consumers, but hackers will also take advantage of them. Attacks from hackers will be more powerful and more devices could be used to make an attack. For this reason, it is very important that security measures are strong in autonomous vehicle networks so hackers do not take advantage and potentially cause loss of life.
The second reason is that 5G is complex and heterogeneous meaning that security measures will have to be applied to different levels. In the case of autonomous vehicle networks, security measures will have to be in place among all of the stratums and all of the V2X applications. Because these different stratums and application types are so complex and different, security measures will have to scale with that as well.
The third reason is because of distributed edge clouds. Distributed edge clouds allow applications and data hosting to be held at the network edge rather than a centralized server. Edge cloud nodes are often prone to spoofing, eavesdropping and other attacks. This would be no different with 5G autonomous vehicle networks as a lot of information will be stored in the 3GPP core cloud [12].
3) Requirements for Autonomous Vehicular Networks
Given the remaining security and privacy vulnerabilities in 5G-enabled AVN, the next sections will begin evaluating solutions that cover these loopholes. When doing so, the following 4 major security and privacy requirements for autonomous vehicle networks will be referenced:
1. Confidentiality: The ability to ensure that communication is secure. Transmitted messages should only be accessible to the intended sender and receiver to prevent malicious entities from stealing data such as GPS locations and misusing it.
2. Authenticity: Should be able to ensure that the data received is from an authorized source. This safeguards from impersonation attacks aiming to give false information or gather private and protected data.
3. Integrity: The content of the data transmitted should be tampered with in any way before its delivery. This is especially important as any fault in information such as those used by the sensors can lead to dangerous consequences on the road.
4. Transparency: While some data need to be protected from unauthorized individuals it is equally necessary for information, such as data regarding upcoming traffic and more, to be available and accessible among each vehicle node. For example, while a vehicle's GPS coordinates may need to be kept private from other entities, it needs to be available to a central system that will help guide the automobile safely along its route.
Overall, an optimal solution to improve the security and privacy of the AVN must be able to fulfil all these minimum requirements. The following sections will explore blockchain as
a possible solution.
4) Blockchain Design
A Blockchain (BC) is a distributed database containing a chain of immutable logical blocks. BCs keep track of every transaction made by every member of the network like a distributed ledger. A single block stores data from many transactions and is chained to the previous
block in the chain by the hash of the previous block, as shown in figure 7. In the context of AVs, transactions could be any observation or activity made by the AV, for example, related to traffic, weather, or obstacle identification information. [13] BCs store transactions securely as they are resistant to data modification, hacking, and failure. There are many proposed models of BC implementation for AVNs, but this report will aim to cover the uniting concepts between them. This section will first describe the BC at a high level, then the contents of a single block. Next, the main categories of BC will be presented. Finally, several key features of BC will be discussed.
4.1 Blockchain Security
BCs are valued for meeting high-security demands of data traceability, data transparency, resilience to failure, and data immutability by building trust among members without a central authority. [14] BCs ensures that data is traceable to its source since each transaction record includes information about participating members. Furthermore, since each member in the BC network has a full copy of the BC, every record in the BC is immutable and can be easily verified against other copies. Data is also transparent due to the BC being stored in each member node and important information, such as a traffic accident alert, can be accessed by all nodes. Lastly, BCs are highly resistant to failure as the loss of nodes does not cause data to be lost. To summarize, BCNs are resilient and enforce data traceability, transparency, immutability by distributing copies of the BC to every node and verifying data validity by group agreement.
4.2 Block Description
Each block in the BC is logically divided into a header and body as shown in figure 8. The body of the block simply stores data from many transactions, but the header is much more complex. A typical header contains the hash of the current block, the hash of the previous block, a value called Nounce, a timestamp, and the Merkle root. [13] Nounce is an abbreviation of number used once and is the solution to a cryptographic puzzle required to append a new block onto the BC. The Merkle root is the root of a Merkle tree: an efficient data structure where each node is labelled with the hash of its children's labels, or data in the case of leaf nodes. Having the Merkle root allows efficient verification that a transaction is contained within the block.
4.3 Categories of BC
There are three categories of BC: public, private, and consortium. These are primarily differentiated by how one can become a member of the network, but other differences are highlighted in figure 9. First, public BCs are permissionless to join and members are anonymous. Anyone can participate in a transaction and anyone can compete to create a block.
Public BCs are used in most models of BC-enabled AVNs since AVs on the road will need to communicate even if they are not associated with each other and have nothing in common.
​
In sharp contrast, private BC members must be authorized and authenticated to join. These are normally used by a single organization where members are known. Unlike public BCs, private BCs have a central authority that can override a transaction or alter data. This means that data is not completely immutable. In a private BC, only pre-approved members have the ability to create blocks.
Lastly, consortium BCs are almost identical to private BCs. The difference is that multiple organizations participate, meaning that members can be approved by any of the participating organizations and multiple organizations will have to agree before data can be altered.
4.4 Smart Contracts
Smart contracts are a feature of BC that allows for a degree of secure automation of transactions. Smart contracts are small programs stored in a BC that can automatically manage transactions under certain programmed conditions. They are written in programming languages specifically designed for it. For example, in the cryptocurrency Etherium, smart contracts can be implemented in Solidity. In the context of AVs, a smart contract could help an AV communicate with roadside infrastructure (like V2I), or help identify riders.
4.5 Consensus Algorithms
Consensus algorithms are one of the key mechanisms that help enforce reliability in a BCN is how the AVN can decide to add new blocks in a decentralized manner. In general, block creators must commit some resource, such as computation time or financial stake, and compete to add the block and receive a reward, usually transaction fees. The two most well-known consensus algorithms are Proof of Work (PoW) and Proof of Stake (PoS). In PoW, miners commit computation time by competing to solve a cryptographic puzzle to which the Nounce is the solution. In PoS, forgers or minters commit financial stake as collateral, and the winner is randomly chosen with a probability based on the size of collateral. If the block is found to be fraudulent, the stake will be lost. In both cases, the first to complete the puzzle receives the transaction fees as a reward.
5) Security Threats
AVNs use wireless communication to convey messages between different objectives which is why they are vulnerable to commonly known network attacks. We will be focusing on the top network attacks that can compromise the network that is Sybil attack, message modification, bogus message, denial-of-service (DoS), packet-drop attack, grey-hole, replay, location, impersonate and eavesdropping [15].
5.1 Sybil Attack
Sybil attack is when networks are invaded using various genuine/dummy identities. This assault is difficult to identify. Hackers can send various messages with various identities to mislead different vehicles without being recognized. Indeed, even if a dummy identity has been recognized, the hacker may get away without being identified.
5.2 Message Modification Attack
In a message modification attack, a hacker changes bundle identities to guide a message to a
broiler vehicle or adjust the information on a compromised vehicle. The design is to acquire
data about the objective and no information is changed. Nonetheless, passive attacks are
regularly preliminary activities for active attacks.
5.3 Bogus Messages
In Bogus messages represent fake or spam messages produced and disseminated by hackers. A hacker can be a third-party or a validated vehicle. The hacker can communicate fake messages to misinform other drivers' choices to get benefits. For instance, the hacker sends
a phoney message alert to cause different drivers to pick different routes so that a couple of vehicles would be left on their objective route.
5.4 DoS Attack
Dos attacks can be set o by both inside and outside users. DoS happens when hackers infuse an incredible volume of messages into the network forcefully to make network assets inaccessible to legitimate vehicles. This assault can be dispatched in an appropriate approach to frame the disseminated DoS attack, which can seriously endanger the accessibility
of vehicular networks.
5.5 Packet-drop Attack
The packet-drop attack is hard to identify. This attack happens when dropping every single packet like a black hole as opposed to sending them which causes a large amount of data loss.
5.6 Grey Hole Attack
The grey hole attack is like a packet drop attack, all things considered, it drops packets
specically. The packet dropping conduct making this attack harder to recognize and foresee
than the packet drop attack.
5.7 Replay Attack
A replay attack happens when hackers can maliciously or deceitfully communicate messages.
This assault in vehicular organizations can cause other authentic vehicles to have
an off-base assessment of the current trac condition. Besides, it can initiate the DoS attack.
5.8 Location Attack
The location attack happens when hackers follow genuine vehicles' locations by continuously
observing and examining messages sent by the vehicle. This attack should be possible in
any event, when hackers continue to change their nom de plumes.
5.9 Impersonate Attack
The impersonate attack happens when an attacker sends messages to a target vehicle using
the hacked vehicle. Normally, the hacker needs to hack real vehicles first. Once successful,
hackers can compromise the network and blame the hacked vehicle.
5.10 Eavesdropping
Eavesdropping happens when hackers gather all conceivable data from the network. Unique
in relation to the past referenced assaults, snooping is a detached attack, and it has no
impact on the network. Nonetheless, it disregards the privacy of the clients [9].
6) Applications of Blockchain in 5G V2X Networks
To alleviate some of the security and privacy concerns, we can apply blockchain to different aspects of 5G autonomous vehicle networks to mitigate many of the privacy and security concerns. This section will discuss two such applications, secure group setup with privacy preservation and intelligent sensing and tracking using blockchain.
6.1 Secure Group Setup with Privacy Preservation
While driving on the road, autonomous vehicles may form platoons or groups where they will drive together for a short or extended period of time. Platoons allow vehicles to brake and accelerate simultaneously as well as help to increase road capacity. A blockchain solution
can be applied to V2V communications while autonomous vehicles form platoons. Lai et. al [11] proposed Secure group set up with privacy preservation that can mitigate privacy concerns as well as combat Sybil attacks, replay attacks and bogus messages by preventing malicious
cars from joining the platoon in the first place.
Figure 10 gives us a visual for what this blockchain looks like. In this application, Blockchain helps vehicles make decisions about what vehicles join the platoon. When a vehicle attempts to join a platoon, vehicles send messages to each other and generate a rating for that vehicle. These ratings could be based on previous experiences with said vehicle, the validity of its identity, and the authenticity of messages received from said vehicle. They then upload these ratings to nearby RSUs or BSs. The RSU or BS will calculate a vehicle's trust value based on the ratings the vehicles sent and place it into a block and add it to the blockchain. The blockchain essentially keeps track of vehicles' trust values. Vehicles in a platoon can use this blockchain to decide if a vehicle is trustworthy enough to join the group.
​
We can layer privacy preservation to work in conjunction with this solution. Vehicles try to join platoons based on shared attributes such as similar routes, interests and passenger's social attributes. However, it can be considered a privacy concern if vehicles and passengers know what the others' shared attributes are. To combat this, we can use multi-party private set intersections. Instead of directly sharing attributes with other vehicles, they can simply be aware of how many attributes they have in common. Vehicles can set a threshold for how many attributes they want to have in common with vehicles joining a platoon. The intersect of their attributes is taken and they only cooperate if they have enough attributes in common.
Through this technique of secure group setup with privacy preservation, we can ensure that many security and privacy concerns in these V2V applications are mitigated.
6.2 Intelligent Sensing and Tracking using Blockchain
Blockchain can be applied to solve other privacy and security concerns in V2N and V2I applications as well. D. Reebadiya et al [8] proposed an Intelligent Sensing and Tracking Scheme for autonomous vehicle networks depicted in figure 11. This scheme would prevent
hackers from uploading fraudulent and unverified data to the network and prevent hackers from manipulating good, verified data.
This scheme works in 4 layers. The real-time infrastructure deployment or data layer generates information using real-time sensors on vehicles, road infrastructure and pedestrians. For example, cameras could generate data about traffic accidents or lawbreaking vehicles. Wearable sensors on pedestrians can generate data about where pedestrians are or if a pedestrian is at risk of being hurt. On-vehicle sensors will generate data about other vehicles that are close in proximity to them. This data on its own is insecure and to minimise latency we can use mobile edge servers and blockchain.
The mobile edge server layer computes and analyzes the data and makes decisions about it very quickly. For example, if a bad traffic accident occurred the mobile edge layer can compute and analyze the data generated and make the decision to call autonomous emergency
vehicles.
Fig 11. A high level visualisation of intelligent sensing and tracking using Blockchain [8]
As it stands, this information could be manipulated and information channels could be hacked. Blockchain can be applied to make this data secure and hard to manipulate or change. The blockchain layer will store sensitive information into immutable blocks encrypted with cryptographic hashing. Due to the nature of blockchain, this information will be immutable, transparent and traceable and very difficult to tamper with. Storing information in this way would remove the need for third-party verification and mean that data is integrity protected and more trustworthy. Blockchain's consensus mechanism further ensures the validity and reliability of data uploaded to it. Autonomous vehicles will perform fixed algorithms on the data from this blockchain such as Proof of Work, Proof of Burn and Proof of Attorney to validate and verify new entries in the chain since blockchain is consensus-based technology. The cloud computing layer delivers many essential services in this scheme including storage, databases, and network and software analytics which allows these resources to be flexible and scalable.
7) Challenges and Potential Solutions of BC-Enabled AVN Implementation
Implementing BC for an AVN comes with many new challenges and security threats that
must be addressed. Researchers have proposed many implementation models and solutions
to address these problems, but more research is required as there is not yet a standardized
AVN model.
7.1 Causes
AVNs are a unique form of network that are affected by environmental factors not found in other networks. With member nodes constantly entering and leaving the network, and actively moving over large geographic areas, the network topology is very dynamic. [14] This
makes implementing a BC-enabled AVN difficult as these environmental factors introduce new concerns that have not yet been well researched.
7.2 System Data Throughput
Due to the many sensors on each vehicle, AVs will generate a huge amount of real-time data that will be difficult to transfer over the network. With data coming from cameras, actuators, radar, lidar and other components, the network will need a large capacity to process the number of transactions. Since blocks have a limited size, the time to generate and add blocks to the BC increases due to waiting for previous blocks to be added. [14]
Using 5G networks will help increase throughput and reduce latency, but further improvements can be made. MEC servers can be leveraged to outsource computation and cache content. MEC servers are a type of edge computing that would reside in RSUs and RANs. Furthermore, the sharing of unuseful data can be reduced by forming social groups:
an AV Social Network. [16] Working with the assumption that AVs in close proximity and with similar interests will want similar information, the overall network congestion can be reduced by only sharing information with the social group.
Additionally, consensus algorithms like PoW cause delay in adding blocks and are expensive to compute. AVs may not have sufficient computing capacity which will increase latency in the system. Considering the scarce resources of nodes, a lightweight BC model is needed and a consensus mechanism must be designed specifically for AVNs.
7.3 Storage Scalability
Due to the huge amount of data generated and the rapid increase of nodes added to the network, it is impractical to store a complete copy of the BC in each node. [13] While storing a copy of the BC in each node makes data readily accessible and transparent, BC does not scale well for data storage. Storing a copy of the BC in each node also raises the issue of data privacy as all past and present data is visible to all nodes in the network.
Sharding the database is an option to reduce the storage load on each node. [16] Sharding involves dividing and further distributing the BC among nodes at the cost of some reliability. This concept can also be applied to computing as well, with only a portion of the total number of nodes working on consensus. When there are a large number of nodes, the detriments of sharding is reduced.
Another interesting solution to mitigate the storage cost of an entire BC copy at each node, is to use a private BC model with two types of nodes. [16] AVs would act as ordinary nodes that only relay, exchange, and accept data. These lightweight nodes would only store
the metadata, i.e. block header, of each block rather than the data as well. RSUs and other RSI would act as full nodes that are capable of competing to add blocks and complete consensus.
7.4 Trust among AVs
Trust is extremely important among AVs and RSIs in an AVN however in a public BC, malicious nodes can enter the network and falsify data or messages for their own benet. As an example, a malicious node could broadcast that an accident occurred ahead to clear
traffic for itself. The question of how to only add trusted nodes to the BCN must be further researched. Reputation models are a promising solution and are discussed in section 6.1, but must be decentralized to scale well. Lastly, AI may be used in the future to detect malicious
nodes.
7.5 Incentive to Share Data
With scarce onboard computing resources and energy, it is not always beneficial for an AV to share data with the network. [16] An incentive mechanism must be introduced to encourage AVs to share data and benet the network. This is important as potentially vital information such as poor weather must be distributed to affected nodes.
8) Conclusion
Fully autonomous vehicles are an up-and-coming technology that will soon be incorporated into everyday traffic. While still imperfect, the most popular architecture for implementing the autonomous vehicular network is one that is 5G-enabled. Our project explored security and privacy concerns in 5G AVNs and how we can best solve them. In this paper, we
discussed what 5G is and the architecture of 5G-enabled autonomous vehicles. We looked at why 5G has so many privacy and security concerns that we had to look out for. Requirements for 5G-enabled vehicle networks were laid out and we used this as a guideline for justifying our solution. We evaluated blockchain as a possible solution and, when compared with the security and privacy requirements, it proved to be suitable for 5G AVNs. We proved this by
looking at some applications and implementations of blockchain in 5G AVNs and how other solutions can combine with blockchain to respond to security threats even more efficiently. We also looked at problems with implementing blockchain-based solutions and how we can
potentially solve those challenges as well.
_______________
Appendix
Team Member Statements
Below are the statements from each member outlining their contributions to the project.
Trishna
Trishna is a 4th year undergraduate student pursuing a computer science major and a minor in data science. For this project she created, maintained and regularly updated the website throughout the semester. For the project demo she researched the general structure of the AVN and justified why it would be beneficial, in terms of improving security and privacy, to integrate 5G and blockchain into it. For this report she worked on the introduction sections - explaining what are autonomous vehicles and why we should be interested in them - as well as the requirements section - describing what are the key requirements the proposed solutions should meet in order to be considered suitable for AVN. She also contributed to the conclusion.
Megan
Megan is in her fourth year of studies in computer science at the University of Victoria. For this project, she researched and discovered many of the security and privacy issues in 5G autonomous vehicle networks. She also looked into 5G-enabled vehicle architecture, requirements of such networks and why 5G autonomous vehicle networks have so many security and privacy concerns. She also researched and analyzed some ways of resolving privacy and security concerns in such networks. She also dived deep into why blockchain solutions work really well for such concerns and what other solutions can be combined with blockchain to make it even stronger. Sections she wrote in this report include:
-
Applications and Implementations of Blockchain in 5G
-
Secure group set up with privacy preservation
-
Intelligent Sensing and Tracking using Blockchain
-
What is 5G?
-
5G Architecture for Autonomous Vehicles
-
Why does 5G bring so many challenges?
-
Part of the Conclusion
In the project demo she spoke about:
-
The Introduction and what the project was about
-
Applications and Implementations of Blockchain based solutions
For the project proposal she wrote the “Approach” section and for the project update she wrote the “Changes” and some of the “Progress” section. She ensured deadlines were met and that project requirements had been completed and feedback had been given to those who responded to us.
Hiran
Hiran is a fourth-year undergraduate student in Computer Science with network option with Business Minor at the University of Victoria. He proposed the theme of autonomous vehicles using 5G networks. He mainly researched security threats related to autonomous vehicle networks. Did an in-depth analysis of possible solutions for the corresponding security threats. He took incharge of the presentation for the project. He covered the Autonomous Vehicle Network and security threats sections.
Owen
Third year undergraduate Computer Science (Network option) student at the University of Victoria, contemplating graduate school (wink wink) versus entering the workforce quickly. Throughout the project, Owen researched every broad aspect of AVNs and 5G as well as blockchain, once the group decided to dig deeper into one of the potential security solutions for AVNs on the request of Dr. Pan. In the project report and project presentation, Owen covered the “Blockchain Design” and “Challenges and Potential Solutions of BC-Enabled AVN Implementation” sections. Owen also ported the report document to LaTeX.
References
[1] Al Jazeera. Musk (again) says Tesla is close to fully self-driving vehicles. Automotive Industry News | Al Jazeera. July 2020. url: https://www.aljazeera.com/economy/2020/7/9/musk-again-says-tesla-is-close-to-fully-self-driving-vehicles (visited on 04/13/2021).
[2] R. Inc. Honda's part self-driving Legend a big step for autonomous tech: The Guardian. Reuters | Business | The Guardian. Mar. 2021. url: https://www.theguardian.pe.ca / business / reuters / honda - to - sell - limited - batch - of - level - 3 - self -driving-cars-559493/ (visited on 04/13/2021).
[3] F. Richter. Infographic: Vehicles Are Ready for Autonomy - But Are We? Statista Infographics. Nov. 2019. url: https://www.statista.com/chart/20091/annual-increase - in - number - of - vehicles - equipped - with - hardware - for - fully -autonomous-driving/ (visited on 04/13/2021).
[4] CANADA ROAD SAFETY WEEK 2019 FACTS AND STATS. Canadian Association of Chiefs of Police. May 2019.
[5] Andrés Fielbaum. Strategic Public Transport Design Using Autonomous Vehicles and Other New Technologies. In: International Journal of Intelligent Transportation Systems Research 18.2 (Aug. 2019), pp. 183191. doi: 10.1007/s13177-019-00190-5. url: https://doi.org/10.1007%2Fs13177-019-00190-5.
[6] M. Brown. Tesla's Elon Musk explains how self-driving will go beyond robo-taxi . Inverse. Oct. 2020. url: https://www.inverse.com/innovation/hail-tesla-robo-taxi (visited on 04/13/2021).
[7] N. Michell. Autonomous vehicles could increase private car traffic, says new report. Cities Today - Connecting the world's urban leaders. Feb. 2017. url: https://cities- today.com/autonomous-vehicles-increase-private-car-traffic-says-new- report/ (visited on 04/13/2021).
[8] Dakshita Reebadiya et al. Blockchain-based Secure and Intelligent Sensing Scheme for Autonomous Vehicles Activity Tracking Beyond 5G Networks. In: Peer-to-Peer Networking and Applications (2021), pp. 118. doi: 10.1007/s12083-021-01073-x.
[9] J. Huang et al. Recent Advances and Challenges in Security and Privacy for V2X Communications. In: IEEE Open Journal of Vehicular Technology 1 (2020), pp. 244 266. doi: 10.1109/OJVT.2020.2999885.
[10] S. Rahmadika, K. Lee, and K. Rhee. Blockchain-Enabled 5G Autonomous Vehicular Networks. In: 2019 International Conference on Sustainable Engineering and Creative Computing (ICSECC). 2019, pp. 275280. doi: 10.1109/ICSECC.2019.8907054.
[11] C. Lai et al. Security and Privacy Challenges in 5G-Enabled Vehicular Networks. In: IEEE Network 34.2 (2020), pp. 3745. doi: 10.1109/MNET.001.1900220.
[12] ImageObject and Organization. Five 5G security challenges that service providers need
to tackle. Infradata. July 2020. url: https://www.infradata.com/news-blog/5g- security-challenges-for-service-providers/ (visited on 03/13/2021).
[13] Rajesh Gupta et al. Blockchain-based Security Attack Resilience Schemes for Autonomous Vehicles in Industry 4.0: A Systematic Review. In: Computers & Electrical Engineering 86 (Sept. 2019). doi: 10.1016/j.compeleceng.2020.106717.
[14] Rakesh Shrestha et al. Evolution of V2X Communication and Integration of Blockchain for Security Enhancements. In: Electronics 9.9 (2020). issn: 2079-9292. doi: 10.3390/ electronics9091338. url: https://www.mdpi.com/2079-9292/9/9/1338.
[15] Keping Yu et al. Deep learning-based trac safety solution for a mixture of autonomous and manual vehicles in a 5G-enabled intelligent transportation system. In: IEEE Transactions on Intelligent Transportation Systems (2020). doi: 10.1109/TITS.
2020.3042504.
[16] Y. Wang et al. Challenges and Solutions in Autonomous Driving: A Blockchain Approach
. In: IEEE Network 34.4 (2020), pp. 218226. doi: 10 . 1109 / MNET . 001 .
1900504.