Blockchain Based Intelligent Incentive Enabled Information Sharing Scheme in Future Generation IoV Networks

With the advancement of vehicle’s traffic information processing and communication capability in Internet of Vehicle (IoV) network widely used for the vehicle-to- infrastructure transportation communication. Firstly, the protection of the user’s (vehicles) privacy at the time of information sharing and secondly, users lack the motivation to share the traffic information with roadside units (RSUs) are two major concerns in the IoV networks. In this regard, we propose a novel adaptive neuro-fuzzy based payment using blockchain (ANFPB) transportation communication scheme that not only motivates users to take participate in the information sharing problems with the payment mechanism but also allows users to anonymously share the traffic information with RSU in the IoV network. Meanwhile, a smart contract is presented to generate pseudonyms to share the traffic information anonymously in a non-trustful IoV network. Also, an algorithm ANFPB is presented for the evaluation of payment based on location, timeline, and quality of information shared by the vehicles. Finally, the extensive simulation analysis shows that the proposed ANFPB is more efficient in terms of preserving privacy and computational costs as compared to state-of-the-art schemes.


Introduction
In the recent decades, more number of smart vehicles connected to the next generation internet which terms as 6G Internet of vehicles [1][2]. This 6G IoV have decentralized network configuration with three layer architecture. The bottom layer corresponds to vehicles-to-vehicles communication. The middle layer responsible to vehicles-to-infrastructure (V2I) (RSUs) edge (fog) communications that collects the information's from the vehicles to maintain the traffic rules and avoid road hazards. The upper layer is cloud server that provides huge data storage for longer term to deep analytics in all spatial regions. 6G IoV network is mainly designed in such a way that it provides driving comfort and safety for the user's terms as infotainment application and safety applications respectively [3][4][5][6]. Non-safety infotainment applications include the real-time traffic information while driving, music on the road, video on demand, and the internet of the wheel. Whereas safety applications applicable to various domains such as alleviating highway turbulence, manoeuvre control, cooperative cruise control, traffic congestion, offenses, broken pavement, dangerous driving and accidents, weather conditions, or any looting or terrorist attack [7][8]. It requires V2I communication to frequently accessible the safety applications by the users in 6G IoV network. Privacy information of the vehicles such as location and identities in V2I communication and lack in the enthusiasm to take users to participate are two major issue in the establishing 6G IoV network.
6G IoV uses wireless technologies that arises privacy issues on the information's broadcasting [9][10]. Normally, data are transmitted by a vehicle to the fog node (RSUs) layer that includes the detail of user's personal information locations, travelling route. Recent literatures show that untrusted fog node reveals the user's information for money that leads to serious vulnerabilities to the society [10][11]. Therefore, preserving the privacy of each vehicle and to ensure their integrity and authentication are other issues in the 6G IoV network. In case of security purposes, the identity authentication technique is the most effective approach to protect the integrity, privacy confidentiality, and availability of the data in IoV [12]. Whereas, authentication along with key exchanging technology, public key infrastructure (PKI) systems is also widely used in other fields, including mobile cloud networks, smart grid, IoV network etc. [13][14][15]. Above literatures suffers from workload leads to a high packet loss in the case of locations with heavy traffic and only provide threshold privacy protection and not able to correctly reply to message of users. Which in turn lose to motivate the users to take participate in the V2I communication in 6G IoV network.
The emerging Blockchain technology provide decentralized network for data storage, which provide user's (vehicles) to broadcast the information anonymously in 6G IoV network without worrying about their privacy in non-trustful fog (RSUs) nodes. It also helps to secure the transaction (payment) in terms of smart contract between vehicles to RSUs without using any intermediary. Recent research involves artificial intelligence and machine learning techniques into 6G IoV network to continuously collect the information and optimize the payment in return paid to users through learning fog node [16][17]. In [18], authors have proposed deep learning reinforcement learning algorithms to motivate the users for uploading the information to the fog nodes in 6G IoV network, but lacks in privacy and security of the users. Ensuring the success of blockchain and fog-cloud technology based on artificial intelligence approaches this paper neuro-fuzzy learning algorithm to secure the information during transmission and privacy of vehicles and motivates vehicles by providing incentives to them. The clod-server provides unlimited resources for data storage and it submitted the data in such a way that not even the government agencies or law enforcement can destroy the privacy of the witness by tracking [2]. Besides all the above importance, it is also necessary to attract the attention of the users towards this service.
As aforementioned by the above challenges, privacy and attracting users' attentions towards to get the reward in exchange for uploading the information (images/video, etc.) of the site of interest (labelled as witness service by users) at the fog-cloud 6G IoV network are the main purpose of this paper. In this regard, this paper presents a novel scheme ANFPB using blockchain ensuring the privacy of the vehicles, and a further adaptive neuro-fuzzy technique is used to evaluate reward based on location, timeline, and quality of the information shared by the vehicles. The major contributions of the proposed model as follow: 1) Firstly, a system model, network model, and blockchain technology are presented to define the involved physical entity, fog-cloud network layer for information sharing and to ensure privacy and provides the reward to users respectively in 6G IoV network.
2) Secondly, we briefly define the system initialization process, pseudonyms exchange mechanism for privacy-preserving that include smart contract on blockchain to ensure the authenticity of the vehicles.
3) Thirdly, a novel Adaptive Neuro-Fuzzy Payment based on Blockchain (ANFPB) is presented to evaluate the reward for the vehicles based on the shared traffic information. Further, to make the scheme free from fraudulent users' revocation authority revokes the vehicles based on their pseudonym exchange history table. 4) Finally, performance evaluation is presented to compare the privacy and computation cost of the proposed model concerning state-of-art-models. The paper is organized as follows. Section 2 presents the study of related work. Section 3 discusses the details of the system model, network framework with blockchain structure. Section 4 describes the proposed model. Section 5 explains the details of presented ANFPB algorithm The performance evaluation and the comparison of the proposed model and the state-of-art models have been shown in section 6. Finally, this paper is concluding in the section 7.

Related work
Several issues of protecting the privacy of the data in Internet of Vehicles have emerged in recent years and the research on analyzing such problems have been carried out in both academics and industries to make life secure. The privacy protection of the identities of the vehicle can be done effectively with the help of various authentication approaches [18][19]. These authentication approaches are broadly classified into three types such as cryptography-based authentication technique [18][19][20][21][22][23][24], reputation evaluation-based technique [25][26][27][28][29][30], and hardware-based trust enhancement technique [31][32][33][34]. Cryptography-based authentication technique deals only with the correct evidence holds by the vehicle. This approach neglects the reputation behaviour of the vehicle. Cryptography-based authentication technique includes Identity-Based Encryption technique (IBE), Public Key Infrastructure technique (PKI), etc. On the other hand, reputation evaluation-based techniques can progressively increment or decrement the credibility of the vehicle based on its behaviour to fulfill the trust threshold verification. This is applicable for the profoundly self-organizing IoV. Besides, this technique also has drawbacks of lacking robustness to the insecure vehicle in the IoV network. Lastly, the hardware-based techniques are focused on building a confided computing platform at the terminal layer for the vehicle in IoV. The reliability of the vehicle is controlled by monitoring the hardware unit, software unit, and other units. The hardware unit includes actuators, electronic control unit, interfaces, etc., whereas the software unit includes the operating system environment, onboard applications, etc. Here, the speed of authentication for keeping the information confidential and securely has been improved as it has self-cryptographic system. But the vehicles participating in IoV are co-related. So, security provided only to the terminal layer is not sufficient.
In the case of authentication techniques for privacy protection, several kinds of research have been carried out in recent years. The paper in [35] considers the Zero-Knowledge proof method to verify the identities. In [36], they proved that it is probably going to lessen the total unique certificates of the vehicle by distributing the certificates to their neighboring vehicles. This authentication approach deals only with identity authentication and also fails to meet the privacy of real life. To find the culprit causing a traffic accident, the true identities can be revealed by the traffic control centre. To improve the efficiency of the authentication technique, researchers in academic and industries proposed to embed hardware chips into the vehicles for security purposes. Where both encryption and decryption of the data can be carried out based on the hardware unit. It also helps to keep some private data secure. TPM (Trusted Platform Module) and TPD (Tamper Proof Device) are hardware-based authentication techniques and the comparative analysis for both the technique are demonstrated in [31]. Both the authentication techniques have the data encryption capacity, but TPD has certain drawbacks such as high price, intolerable to extreme temperature. In [33][34] paper, TPM was used to verify the components present in the vehicle whether they are working properly on the command without alleviating the security provides by IoV.
Besides the three types of authentication technique, some other technique includes group signature and ring signature. Earlier Boneh [37] and Lin [38] developed vehicle communication based on group signature. Privacy-preserving protocol along with confidentiality in VANET based on the sign-encryption technique has been introduced by Hu et al [39]. In the case of the group signature framework, vehicles acquired secret keys and public keys to avoid information leakage. With the increase in the number of nodes, time also increases gradually which concludes that group signature consumes an enormous amount of time. To solve this problem, the researchers adopted a solution based on the tamper-proofing of hardware devices. When the enemies attack the hardware device, the security system is compromised automatically [40][41]. In this case, ring signature was developed [42][43], Xiong et. al. proposed the privacy protection protocol and ring signature technology for IoV [44]. But this requires truthful traffic management agencies. Zeng also proposed CARS (conditional anonymous ring authentication solution) for IoV. Liu proposed an authentication technique based on session keys used in complex communication. Wu developed a secret key allocation system. It includes verification codes and establishes security by providing authentication of group keys. Recently, Hu et al. provide an efficient privacy-preserving authentication system in VANETs on the basis of ring signature [45]. They also developed an efficient, trustworthy, privacy-preserving VANETs protocol based on proxy re-signature. Besides this, they developed a protocol based on remote authentication.
Most of the data storage and authentication are constructed based on cloud computing technology. In this technology, data are stored in cloud servers by the authentication centre or RSU (Roadside Unit). Here, the assumption of cloud service providers to be reliable may not be true in real life because certain users and cloud servers may develop conspiracy. Considering the above problems, Chen et al. [46] proposed a light-weight protocol and anonymous aggregation protocol using fog computing for V2I communication schemes. Finally, after considering several security problems occurring in IoV we have designed a model for protecting privacy as well as keeping concerned on the security of the information in IoV.

System Model
The proposed system model is the association of IoV and fog assisted cloud computing technology. Active participants of this model consist of vehicles (sensor nodes) equipped fully with IoV infrastructure such as dedicated short range communication (DSRC)-based on board unit (OBU), cameras, tamper resistant hardware (TRH), Department of Motor Vehicles (DMV), revocation authorities (RAs), law enforcement organization (LEA) and judiciary. The above mention participants are the physical participants in the system model. Further, fog assisted cloud computing technology is employed to store the information collected through road-side cameras and process this information for the forensic investigation purpose. Finally, this processed information is handed over to the judiciary as an evidence of the events. In case of cloud computing, the storage of the information is done in two layers as fog layer and the cloud layer. Fog layer is used to collect the data, aggregate the data, anonymized the data, and dispatch the data. Here, Roadside Unit (RSUs) is considered as fog node. It is mainly used for storing instant data in turn latency in uploading the information reduces by the virtue of fog layer. Addition to that, loss of captured information is less as report (images/text file/audio/video) of an event can be stored in multiple fog. Whereas, cloud infrastructure responsible for storing all the information, processing the related query, data dispatching, and rewarding systems. Further, it also analyse the information for forensic investigation, taking necessary precautionary measures whenever required or generates warning message. Furthermore, the cloud provides details of forensics data to the trustworthy agencies, including government agencies, law enforcement agencies, judiciary, insurance agencies etc. To encourage the user participation on the road, we proposed privacy assure rewarding system (PARS) for providing reward to the vehicle based on the contribution as services. Rewarding systems consist of one physical entity known as reward collection center (RCC). This center can be petrol pump, gas station, post office, zoo, food court etc. This physical entity also consists of other software modules such as receipt collector, receipt receiver, receipt issuer, and reward calculator. The arrangement of the participants in the system model has been depicted in figure 1.

Network Model
The proposed network model has been illustrated in figure 2. This model describes about the process of transferring the captured information through the on-board cameras of vehicles to the nearest RSUs using DSRC. The information can be any events such as accidents, traffic jam, terrorist attack, etc. of the site of interest (SiO). The data are aggregated in the Fog (RSUs) layer and send it to the cloud layer by remotely triggering. In our proposed work, we used passive service which includes several cameras either installed in the vehicles or on the road site in order to capture images of the SiO and transfer it directly to fog layer with low latency and suitable to store instant information. Later it is transferred to the cloud layer with high security using the block chain technique. This data are kept in the cloud layer for future purposed where it can be used for the forensics investigation process. This information also used to avoid terrorist attack, destructive events, deadly accidents, etc.

Blockchain Structure
Blockchain was proposed in 2008 by Satoshi Nakamoto and has become an effective technology with a significant impact of decentralizing the business way. Blockchain is defined as a synchronized and disseminated record keeper in terms of listing blocks. On considering, its immutable and distributed data storage technique, blockchain is applicable to various areas such as banking, health care, supply chain management, trade finance, transaction in IoT network etc. The advantage of blockcahin is depicted in figure 3. In public blockchain structure, there is no central manager involved instead of the participants in the network is responsible for maintaining the public record. Systematically, in public blockchain anyone can add new block by proof-of-work (PoW) mechanism, which is a cryptographic puzzle technique. A node which determines the solution to the puzzle disclosed the solution to all other nodes present in the network. While nodes will accept the solution only when there is a validation of all the transactions occurring in new blocks. Beforehand, it needs to be confirmed that no other solution has been received. Then, the block points correctly to the last block in the block chain. The stored data in any of the block cannot be altered because it will cause invalidation of all the data stored in the previous blocks of the blockchain due to its hash function and this also leads to the breaking of consensus between the vehicles (nodes) associated with block chain. Whereas, in the permissioned blockchain special case of private blockchain, only authorized node have both write and read permission and public node only read the record but cannot able to add new block in the chain. Auditing company and business intelligence company owners set up their own permissioned network, where nodes send the request to join the network after permissioned is granted, that worked in decentralized blockchain manner. An alternative consensus algorithm named as proof of stake (PoS) is used to add the new block in the network instead of POW. In PoS, already exited nodes choose the leader among them that are responsible for creating the new block based on their stake (trust availability). This reduces the computational time and cost with respect to PoW based bolckchain.
Each block has two parts, the header consist of pointer (hash value) to previous block and the latter is body part contains records of all the validated transactions, including users' information, time stamp, receipt etc. as shown in figure 4. This chain of blocks is formed by connecting the current block to previous block by determining the hashed value of current block using the hash value of the previous block. Here, all the vehicles (nodes) hold its own copy of their block chain and further to determine their current state, each transaction needs to be processed in the order their appearance in the block chain. The blocks present in the block chain are partitioned into six sections as a hash function of the previous block in the block chain, nonce, hash function associated with the current block, timestamp, Merkle root, and exchange data.
The blockchain is established on the pre-selected RSUs to share transaction records for audit that remove the authenticity of any intermediate trusted authority. The RSUs (fog node), vehicle, cloud server construct a private blockchain based on consensus (PoS) mechanism. The smart contract (set of digital commitments) is an important factor in blockchain, where all the rules are predefined and executed when there is an event occurs, this rules cannot be modified once it is spread on the blockchain network. It ensures the transparent nature of the network, which encourage the vehicle to upload the pics of the site of interest without worrying about their privacy leakage and finally economic benefit provided to the vehicle by the reception collection centre. In this paper, we proposed security and privacy aware fog-cloud incentive based using private blockchain approach to encourage the vehicles for uploading the pics of site of interest.

Proposed Model
This section describes the privacy and security aware proposed model in IoV. We present a novel model for vehicle's privacy and securing the transferred data using pseudonym exchange mechanism through blockchain technique. The primary concern of the proposed model is to maintain scalability in terms of more number of vehicles participate in the service of upload the captured information to RSUs by providing incentives to participating vehicles. Later on, these rewards redeem on the RSU's as petrol pump for refuelling. On the other hand, providing static infrastructure all-over the roads to capture all the necessary information for every instant of time would cost the administration to an extreme. Instead of static infrastructure, we maintain mobile sensors to timely generate the necessary data to the cloud and save it for future purpose without any interference from the outsiders or attackers. The vehicles (nodes) may be malicious and upload the wrong information to fog nodes so that, revocation authority identified those vehicles and put in blacklist or revoke them. The privacy of vehicles and transaction of deposits (incentives) must not be leaked. To counter the security and privacy problem, permissioned blockchain technique is used with proof-of-stake consensus mechanism. For system setup several parameters are used and the initializations are discussed in the below subsections.

Preliminaries of system initialization
Each vehicle conveys a bunch of pseudonyms imposed at the time of registration by DMV. To revoke a vehicle in IoV, encryption of secret key is performed and this key is stored in RAs. Here, secret keys consist of symmetric key which is used mainly for the generation of pseudonym denoted as Ҡ and individual secret key associated with each vehicle is denoted as Ҡ . For storing such keys in the RAs, an encryption algorithm known as Elgamal encryption technique is used. This technique is far better than ECC (elliptic curve cryptography). Let us consider as a cyclic group with prime order as . A generator is used to generate . At the time of registration of Vehicle the DMV, select any random number as a private key which is denoted as ∈ * and determine the public key using the mathematical expression Ҡ + = . DMV holds the responsibility of distributing the shares of the secret keys to the RAs by using a secret share scheme based on threshold. Here, is divided into ℓ equal parts and ℓ denotes the number of RAs so that each RA can hold a share of secret keys say where ∈ { 1 , 2 , 3 … … . ℓ }. Later, we can conclude that any existing secret sharing mechanism can be employed in such process.

Tamper Resistant Hardware (TRH) initialization
The installation and initialization of tamper resistant hardware devices in a vehicle are done at DMV. So, for this purpose the owner of the vehicles needs to personally visit the DMV and the credentiality of the vehicles is confirmed. Then DMV initialized the TRH in the vehicle by saving several system parameters associated with the TRH. The parameters include { , , , Ҡ + , , } where denotes the secret initial counter of vehicles used in the generation of pseudonym and denotes the factor to increment pseudonym. In addition, DMV preloads vehicles TRH including secret keys such as Ҡ and Ҡ .

Generation of pseudonym
DMV is responsible for generating number of vehicle pseudonym at the time of registration by considering the secret counter of the vehicle and increment the counter by using the pseudonym incrementing factor . It is important to note that pseudonym is allowed to trace secretly to facilitate revocation whenever necessary by the RAs. TRH is a place for storing these pseudonyms and later used it for conditional privacy preserving communication process. The mathematical expression for generating pseudonyms are defined as follows Where ℰ = + and denotes the current count of the pseudonym that has been generated.
denotes the vehicle identification number. DMV records these pseudonyms in a database and index it through the value . All these pseudonyms and anonymous certificates are stored in TRH of the vehicles and distributed all these data to RAs too. These anonymous certificates are used mainly for exchange of pseudonym during the communication process. For the revocation process, the encryption of Ҡ and Ҡ is done by TRH and dispatch to the RAs. Here, in revocation process RAs play the role of trapdoor. The encryption of the previous keys with public master key based on ElGamal encryption technique is defined mathematically as follows Where denotes the random number generated only once (nonce) by TRH. The encrypted information, including { 1, 2}is sent to RAs from TRH. On the other hand, decryption of Ҡ and Ҡ keys can also be carried out based on their warrant and construct from separate by cooperating together. The main purposed for storing the encrypted keys in the database is of two reasons such as in case of privacy any conflict; RAs used the keys for vehicle revocation. For each vehicle, the database is maintained by DMV and the credentials of the vehicles which includes { , , } are save.

Identity exchange using the block chain technique
The identity related to the vehicles is the most important data that needs for privacy preserving. The concept of multiple pseudonym does not guarantee the enhancement of privacy as the pseudonym can be traced and relate to the sender. Hence, new model for privacy preserving using the blockchain in identity (pseudonym) exchange mechanism has been proposed in this paper. Each vehicle has its own block to store the identity exchange record. The data are recorded in the blocks in an automatic and standardized manner so that if there is no trust between people, at least user should have the option to believe that the code and system have been set up effectively and will work respectively. The privacy of vehicles maintain by the pseudonym vehicles, that can be cancel out whenever necessary. Based on the concept of DSRC, every vehicle in IoV is directed to broadcast the information to its nearby vehicle. This information includes current position, current speed, direction, etc. At a point when a vehicle needs to exchange pseudonym with another vehicle by preserving its privacy, a beacon message is raised. An intent flag has also been included in the message to alert the vehicle for pseudonym exchange. In the meantime, all the nearby vehicles have a choice for exchanging the pseudonyms after receiving the beacon message. The beacon message is represented as ℳ = ( || . || ), where denotes the information, including position, acceleration, speed, heading, steering wheel angle, brake status etc. And, .
denotes the parameters such as integrity, authentication, etc. In order to avoid from malicious attack of the exchange pseudonym, the process of exchanging the pseudonym should be anonymous. Since, the knowledge of exchanging pseudonym provides probabilistic and statistical facility to the attackers.
It is also important to note that, the validity of pseudonym is checked before exchanging through the pseudonym revocation list. Meanwhile, the report for exchanging is sent to the any of the available RAs anonymously. And if the vehicle is favourable for exchanging the pseudonym, then RAs exchange the report. Further, it is noted that the benefit for revocation has been shared to all the RAs instead of sharing to a single entity.
The RSUs (fog node) maintain the private blockchain by collecting the real-time information, authentication of vehicles, data integrity and finally upload the data to cloud server. Vehiclevehicle identity exchange could be done on the ledger through smart contracts by employing the blockchain-verified identity. The detailed overview of the smart contract is presented in algorithm 1. The () function initialize the registration process of the vehicle obtain from DMV, vehicle I gets its certificate , which is used to identify itself with identity and licence plate number . Vehicle joins the blockchain network with identity and able to gets its public and private keys ( , ) and its wallet address . The vehicle execute the system initialization and upload its information { , , , , , } to nearest RSUSs, where each infoemation stored into the memory block. To ensure the vehicle authencity and their integrity of data, asymmetric encryption is used by RSUs signature in the blockchain as follow: Where is the digital signature by sender's private key to the transferred information , ( ) is the hash digest of and is the decode function of the information using sender public key. The ( ) function is used to implement new smart contract between the RSUs and vehicle on the agreement of the contract items through signed by their private keys. This smart contract is accessed by all the vehicles and RSUs deployed in the network after successfully verification of consensus mechanism. Each smart contract is responsible for maintain records such as state variables, account address of sender ( ), account address of RSUs ( ) corresponding payment , timestamp and transaction time . In consensus (PoS) mechanism, where number of records in each RSUs treated as their respective stake or trust. For initial stake distribution, the Genesis block of the blockchain has RSUs identifies, public and stakes ({ } =1 , { } =1 , { } =1 ) respectively. Initially, genesis block is empty signed by Cloud server. After, leader is selected to generate new block among the RSUs using their probability of the previous block's stake. The newly created block has block header, block number, time stamp and Merkel hash root from the previous Merkle root hash root tree generated from previous records. Further, RSUs adds this newly block in the network by updating the stake value and records.
The ( ) function excuted after the conseus and performed the smart contract between RSUs and vehicle if the ≥ and perfrom the information (images of SiO ( )) transfer and incentive settlement. Further, system automatically updates the ledger of the blockchains, state variables and transfer the information to the cloud server.
In this way, the proposed model is beneficial for the scalability in the IoV network that can keep up with the large number of vehicles. Unlike, PoW based public blockchains, the proposed consensus (PoS) mechanism based blockchain technology is carried out into small number of preselected RSUs and turnout to be reduction in both transaction latency and cost. Also, total time to create a new block is stable regardless of the network size that ensures the anonymity of the vehicles which is maintained by the authorized RSUs.

Communication through assisted fog and cloud
The sensor camera installed in a vehicle intended to capture the full view of the sites (omnidirectional) known as the full view model. The working of these cameras is based on the sites density where only specific vehicles were allowed to capture images in dense sites and in case of sparse sites most of the vehicles are allowed to take images to cover up maximum area of the sites of interest. So, we assume that in case of congested street there will be enough vehicles on the road to participate in the service. But, all the vehicles available on busy road were not allowed to participate in the service because there will be a wireless traffic due to excess amount of information. Hence, selected and mobile cameras around the street and attached to the vehicles play a vital role in such situation.

Reporting of the events and acknowledgement
In order to encourage the participants involved in the service for capturing images of certain events happening on the road, we developed a model for providing reward to the participants who provide accurate information regarding the events. The confidentiality will be maintained in this rewarding process. The steps involved in this process are described separately in the below section.

Pictorial event reporting
The images captured by the vehicle are collected and the software timestamps together it with GPS data obtained from GPS module which includes pseudonym pick up from the pool and sign and finally send the message to the cloud. The proper format of the message can be represented as Where denotes the DMV issued anonymous certificates of the vehicle. represents the identity number of the event. Location associated with vehicles is denoted by and for location of event is denoted as . is the quality of report of an event. ℘ indicates the vehicle's pseudonym and denotes the images taken by the vehicle. and denotes the private and public key which are responsible for the secure communication process.

Collection of receipt
After confirming the authorization and the contents presented by the contributors, the validity of the is verified by receipt issuer and also examine the pseudonym validity. Once the validity is approved the cloud produces a receipt of the vehicle which contains a receipt ID. This ID act as a coupon to claim the reward. The format of the receipt issued to the vehicle and to Receipt collector is represented as follows denotes the receipt ID. The above receipt format is necessary to claim the reward. Finally, cloud signs the receipt and later sent it to the vehicle.

Acknowledge
After receiving the receipt, the vehicles need to acknowledge to the cloud. Only after a valid acknowledgement, the coupon can be redeemed.Using this acknowledgement, users are allowed to use the same pseudonym in the process of reporting and collecting the reward. The mathematical representation of acknowledgement is Additionally, the hash value has been calculated by using its individual secret key and included in the acknowledgement to avoid from any case of conflict.

Adaptive Neuro-Fuzzy based payment
The reward is given to those participating vehicles, only after analysing their services. This can be done by the use of the Adaptive Neuro-Fuzzy Inference System (ANFIS) [47]. ANFIS works on basically artificial neural network (ANN) and Takagi-Sugeno fuzzy inference system (FIS). The ANN is used for weight adjustment of membership function to lower down the error rate inside the FIS according to feedback from the system. Location of event ( ), timestamp of an event ( ) and Quality of report for an event ( ) are feed as input to the proposed ANFIS for estimating the reward for a vehicle in the return of reporting an event shown in the Fig. 5.   The proposed ANFIS evaluate the output at fifth layer throughout processing the input parameter one by one each total of four layers. The working procedure of each layer as follow:

IF THEN
1. Fuzzy layer-This layer is adaptive in nature due to weight of the membership function node (squared) is automatically tuned based on feedback. This layer uses triangular and trapezoidal where, ={0,1}define the degree of input according to membership function.
2. T-norm layer-This layer non-adaptive in nature (circle with label ), it determines the firing strength of each rule associated with it as output ( 2 ) to the node. This can be done using minimize operator AND, which multiply each rule incident to the concerned node as follow: Where # represent the firing strength of each T-norm node.
3. Normalized layer-Third layer is also non-adaptive type and mainly estimate the firing strength proportion of each rule coming from T-norm layer pertain to each node ( # ) labelled inside the circle. The output of normalized layer 2 is calculated as follow: 4. Defuzzy layer-This layer has essence of adaptively tuned the firing strength of each rule pertain to square labelled node ( # ). The output (consequent parameter) of fourth layer is multiplication of individual firing strength and normalized firing strength of rule is calculated as follow: 5. Aggregated Output Layer-This non-adaptive layer estimated the final output of the ANFIS system treated as performance evaluation layer. This can be done by summation of the signals coming to single node represented with summation ∑ symbol inside single node as follow: The proposed ANFIS is used to calculate the reward in return of reporting an event by the vehicle represented as Adaptive Neuro-Fuzzy Payment based on Blockchain (ANFPB) in the algorithm 2.
The presented ANFPB work in two passes: (1) Forward pass-in this pass, input parameter are propagated from first layer to fourth layer and output of Defuuzy layer is note down. In this pass, all the antecedents' parameters are fixed but consequent parameters are updated. Further, obtained output of the fourth layer is compared with actual output ( ) and Loss is estimated using least square method as follow: (2) Backward pass-In the meantime, Gradient descent method is used to minimize the error and membership function is updated in the adaptive first layer node with the learning rate {0,1} as follow: At this time, consequents parameters are fixed. The complete backward and forward pass is known as one learning episode. The presented algorithm is runs until convergence or maximum number of episode. 5. Aggregated output is estimated using Eq. (11) at output layer. 6. END

Redeem awards
The redeeming of awards is done by showing the coupon which has been collected from the cloud to RCC ((reward collection centre in cloud server as another physical entity) by the users. The contents and the validity of the coupon is verified by RCC and further the validity of the pseudonyms and the contribution is also verified by RCC. Besides this, RAs send the total amount of the reward to RCC and later provides the actual reward to the vehicle. Finally, all the reward transfer process is done through bitcoins.

Revocation system
An authorization letter is must needed to carry out revocation of a node. RAs are responsible for this process. The misconduct will be checked by the departments or expert who includes law enforcement agencies, specialized expert etc. Decision will be taken by these experts either to precede the issue a revocation or not. The first case is proceeding with the revocation; here RAs play the role of retrieving forensics data from the cloud. Then, according to the time interval as mention in the query, the data from the cloud is provided to the RAs. Later, RAs view the values of the message to determine the pseudonym used and also search for the related pseudonym in PEHT (pseudonym exchange history table maintain by RAs) to check whether the original owner used the pseudonym or it has been transacted with other user. Besides, PEHT is subject to make RAs know the next step. Based on recent time the PEHT is searched, where RAs is responsible for constructing from separate by cooperating togetherwhich is related to the pseudonym and the cipher keys is decrypt in session leader. The mathematical expression for decrypt is represented as follows Here, decryption of key Ҡ and Ҡ is performed by RAs and extract the vehicle identity based on the pseudonym.

Performance evaluation
This section describes about the qualitative performance evaluation of our proposed model subject to security and privacy preserving analysis, average transaction confirmation time, communication cost, anonymity and attacking probability of the vehicle with respect to others state-of-art-models. We consider 5000 × 5000 2 city area that includes maximum number of 100 vehicles. We use NS-2 simulator to develop a simulation platform with varying speed of vehicle 20 to 80 km/h and the other simulation parameters is shown in Table 1.

Security preserving analysis
The main purpose of the proposed model is to preserve the security and privacy. An attacker can observe the transaction or transmission of data happening between the vehicles as well as with the cloud. This attacker can analyze the obtained data, but it cannot analyse the pseudonym during the exchanging process because it is anonymous and encrypted before transfer. Hashing is performed using the secret key Ҡ , as sender holds this key the integrity of the data and non-repudiation is provided. This work only under the condition that secret key cannot be compromised. Whereas if Ҡ is compromised, then this alone cannot lead to hazards of the system as in this case only a part of the pseudonym can be obtained. But when we consider both the keys to be compromised, then the condition becomes more disastrous because attackers can easily use the pseudonym. For security purpose, the information about the events happening on the road is reported by the vehicles on the cloud. Where, the vehicle selects the and any ℘ , later the message is sign and perform encryption using the public key of the cloud. It is worth to be noted that, until unless if the vehicle did not compromise, it is not possible for the attacker to interrupt the communication process. On the other hand, if both the keys (public and private) are compromised, then attackers can utilize the information associated with the events and even receive the award. In our proposed work such consequences are avoided and provide a secure environment.

Privacy preserving analysis
In case of privacy preserving, the vehicles report the events anonymously to the cloud so that the attackers are kept away from the original report generators.So, it is hard for the attackers to abuse the privacy of senders' message. Especially, pseudonym exchange and anonymous report make the attackers impossible to detect the original senders. Again, the vehicle ID present in the pseudonym act as a trap door in revocation process. To measure the privacy of the event reports, we used entropy denoted by . The calculation of entropy depends on the anonymity set of the users denoted as (set of users encompasses the site of interest). Let's consider the chances of node to be as a witness among the sets of vehicles as where set of vehicles is denotes as = { 1 , 2 … … , … }. And the entropy of node is defined mathematically as = − ∑ × | | =1

2
. Under normal distribution, the probable outcomes may be | | and the chance of each outcome is 1 | | ⁄ . Since, in this distribution every vehicle has equally likely changes in pseudonym exchange so maximum entropy can be achieved and the mathematical expression is represented as It is also important to know that normal entropy and maximum entropy are equal. We can further say that the entropy value does not depend only on the anonymity set; it also depends on the individual probability. In this situation, more the items in anonymity set, higher the entropy value. However, we consider variable anonymity sets based on its location and density of the traffic. Similarly, in case of rewarding process privacy preserving of the users is also maintained. This is done by providing pseudo identity at the reporting time. The pseudonym involved in such case is selected by the users from the pool of its own pseudonym. Lastly, we can say that without compromising with the security parameters of the users it is hard for the attackers to access the information and remove the reward.

comparison of proposed and existing schemes
The performance of our proposed work is evaluated based on certain parameters by comparing with the existing schemes. Even though the parameters used in existing and our proposed work are not comparable, we can still compare the privacy and security level provided in our posed work and existing work. Besides, the reward winning process can also be compared. Table 2 represents the comparative analysis of existing schemes and the proposed model. In some of the existing schemes, the privacy of real identity is not maintained properly and this needs to be preserved for secure life.
After analyzing these existing methods, we can determine that our proposed model are highly secure and also encourage the users to participate in operating the services successfully by providing privacy preserving of their real identity. Whereas, average total transaction confirmation time refereed to number of consensus mechanism finished for transaction for a vehicles. For this purpose, we set the number of vehicles 40 for 240 minutes. Similar to bitcoin, the traditional blockchain (PoW) model using cloud network, the transaction confirmation time is set to 60 minutes whereas in our proposed ANFPB blockchain (PoS) model using fog node is set to be 15 minutes. The pre-selected RSUs in our proposed model is set to be 10. The frequency of information and incentive transfer in our hour takes from the set value {1, 2, 3, 4, 5}. Fig. 7(a) illustrate that as the frequency of information transferred and corresponding incentive are increases the total transaction confirmation time (average consensus mechanism per hour) increases sharply in blockchain (PoW) model than proposed model ANFPB blockchain (PoS) model. This is due to the fact that our proposed model carries out consensus process done by only preselected RSUs for information and incentive transfer than consensus process is done by all the RSUs in each vehicles of the blockchain (PoW) model. Whereas fig. 7(b) shows that average transaction speed (number of incentive transferred per hour) for one vehicle of the traditional blockchain approach is lower than the proposed blockchain model. This can be attributed to the adding of fog layer to the cloud computing network, which increases the transaction speed and reduces the latency in credit the incentive to the respective vehicle in return of uploaded images of SiO. Thus, it is proved from the results, the proposed model supports fast transaction confirmation time and speed without much delay and it stabilize on increasing frequency of information transfer in fog-cloud based IoV network. In the Fig. 8 illustrate the computational cost (running time) on the RSUs to authenticate the registered vehicle. It is evident from the results as the number of vehicles increases the computational cost (second) in our proposed model is not increases sharply where as traditional blockchain model based on PoW increases sharply and there is sharp increase in without blockchain model and not able to converged regardless of number of vehicles. This can be attribute to the reason that proposed model can efficiently authenticate the vehicles with fewer pairing on preselected RSUs, it does not required to search for the vehicles all over the fog-cloud based IoV network, accordingly running time is less than state-of-art algorithms. It is also worth to note down further increases in vehicle over 60, the communication cost of the proposed model begins to stabilize. This is due to the fact proposed scheme efficiently reduce the latency of computing to generate, receive, transmit and large amount of information in proper way using blockchain with pseudonym mechanism.

Anonymity and being attack probability of the by the attackers over simulation time
For this purpose we choose the number of vehicles in the fog-cloud based IoV network is 50. Fig.  9(a) compares the average probability of the vehicles anonymity over increasing simulation time for the proposed model and state-of-art-models. It can be seen from the result, at the beginning of simulation run, anonymity of the vehicles is increases sharply as the running time of the models increases in all of the models, but after 600 seconds the growth rate in the anonymity of the vehilces tend to stable. This is due to the fact, initially there are more vehicles that need to be anonymous by swapping pseudonym, and after that almost all vehicles have anonymous identity then it tends to stabilize. Whereas fig. 9(b) shows that probability of being attack by outsider attackers for eavesdropping the transferred confidential information or incentives given by RSUs in exchange of the images of SiO for the proposed model and state-of-art-models. From the above results of fig. 9(a) assert that as the anonymity of the vehicles is increases, the probability of being attack by attacker is gradually decreases. It can also be seen from the results, initially the attack probability is 100% after that the simulation time ingresses as the probability of being attack is almost zero within 600 and 800 seconds for the proposed ANFPB model and traditional blockchain with PoW model. Whereas worst performance is shown by without blockchain model, due to the fact our proposed model provide better pseudonym exchange mechanism with blockchain on preselected RSUs.  Fig. 10(a) shows the comparison of the average probability vehicles anonymity over growth of vehicles in simulation area of the fog-cloud based IoV network. It can be seen from the results anonymity of the vehicles shows an upward trend with respect to increase in the vehicles up to 70, thereafter change in the probability of anonymity is negligible. This can be attributing as initially exchange in the pseudonym higher when the vehicle density is low and thereafter model reaches its optimal value and no longer fluctuation in anonymity probability of vehicle. Whereas fig. 10(b) shows the corresponding probability of being attacked by the outsider over increased vehicle density of proposed model and state-of-art-models. It is evident from the fig. 10(b), when the number of vehicle is less than 20 the probability of being attacked is above 80%, as the number of vehicles increases and the attack probability is tends to decreases for all of the state-of-artalgorithms. It is worth to note down that for our proposed model the probability of being attacked by the outsider attacker is quickly stabilize to almost zero with only 70 vehicles with respect to other state-of-art-models. This is due to the fact fog layer assist in changing the pseudonym with lower latency and accordingly RSUs generate report quickly using blockchain approach. Therefore our proposed ANFPB model provides better anonymity and privacy with respect to vehicles density than other state-of-art-algorithms.

Comparison of reward over different incentive algorithms
A comparison of reward over different type of an event with respect to state-of-art-algorithms shown in the figure 11, such as events are ( ) Indifferent refers to broken pavement ( ) Sensitive refers to collision between one or more vehicles, catching fire on some vehicles ( ) Moderate refers to traffic congestion on the road, weather report. It is clearly observed from the result that value of reward of the proposed algorithm ANFPB is higher compared to other state-of-art algorithms. This is due to the fact proposed algorithm uses neuro-fuzzy ANFIS for the evaluation of reward based on event reporting within the timeline, location and quality. Whereas blockchain based algorithm have not any learning approach to evaluating the reward. It is also worthy point to note down that for the sensitive event the reporting must be within 10-15 minutes or live telecast of fire, causality or Collison preferred. In turn, proper action taken by concerned authority to save life of involved persons. As a result, this type of reporting causes more reward as compared to indifferent and moderate type. It is also seen from the result without blockchain approach shows worst performance in the evaluation of reward because of it randomly put the reward value to any vehicles. Fig.11 Reward over different type of event report. Fig.12 Reward over event duration.

Comparison of reward over event duration
A comparison of reward value over event duration with respect to state-of-art-algorithms as shown in the figure 12, as the time duration increases consequently larger number of reports are generated and uploaded to the fog layer by the vehicles. It is evident from the result the as the time duration increases consequently the reward value is also increases up to 90Rs for the proposed algorithm and remains constant afterwards. This is because of as the time duration increases some of reports (sensitive one) have less impact if they are report delay, some of event (road breakage) last longer than a day and as a result the value of reward is not increases further sharply. This can be attributing to the reason that reward value reaches up to the maximum limit. It can be also seen from the result the reward value of the traditional blockchain and w/o blockchain algorithm also increases but they are lagging behind 14% and 24% respectively from proposed ANFPB algorithms. This is due to the fact blockchain based algorithm evaluate the reward value using simple mathematical model and does not able to adapt the reward value according to changes in the location, timeline and quality of the report sent by the vehicles. Whereas w/o blockchain shows worst performance because of it evaluate the reward value without considering period of event reported.

Conclusion
In this paper, In this paper, a novel adaptive neuro-fuzzy based payment using blockchain scheme to preserve the privacy of vehicles in the process of information sharing and also encourage vehicles to participate in the information sharing with RSU in IoV network. A smart contract is also proposed to register the vehicles themselves using an identity exchange process to provide more secure privacy from any attackers to access the contents in midway. Meanwhile, we also introduced the rewarding of most active users in the IoV network to encourage the users to participate using the neuro-fuzzy technique algorithm ANFPB. The simulation results show that our proposed model ANFPB preserves the security and privacy of the users, and encourages the users to participate in the service that also helps to capture the correct evidence of any events happening on the road as compared to other state-of-art-algorithms. In the future, we would like to extend this work by advancing machine learning techniques [53][54] with blockchain. And, also includes the compression of huge data in the cloud to preserving their security.