A Deep-DrpXML and IAG-GWO based CHST fostered blockchain technology for secured dynamic optimal routing for wireless sensor networks

Abstract

In wireless sensor network (WSN), routing is one of the substantial maneuvers for distributing data packets to the base station. But malevolent node outbreaks will happen during routing process, which exaggerate the wireless sensor network operations. Therefore, a secure routing protocol is required, which safeguards the routing fortification and the wireless sensor network effectiveness. The existing routing protocol is dynamically volatile during real time instances, and it is very hard to recognize the unsecured routing node performances. In this manuscript, a Deep Dropout extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Blockchain Technology is proposed for secure dynamic optimal routing in Wireless Sensor Networks (SDOR-DEML-IAgGWO-CHS-BWSN). In this, Crypto Hash signature (CHS) token are generated for flow accesses with a secret key owned by each routing sensor node and it also offers an optimal path for data transmission. Then the secured dynamic optimal routing information is delivered through the proposed Blockchain based wireless sensor network platform with the help of Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf routing algorithm. Then the proposed method is simulated using the NS-2 (Network Simulator) tool. The simulation performance of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provide 76.26%, 65.57%, 60.85%, 48.99% and 42.9% lower delay during 30% malicious routing environment, 73.06%, 63.82%, 59.25%, 44.79% and 38.84% lower delay during 60% malicious routing environment is compared with the existing methods.

Keywords

Wireless sensor network secured routing protocol malicious node attacks Deep Dropout extreme machine learning Improved Alpha-Guided Grey Wolf Crypto Hash Signature token blockchain technology

1 Introduction

WSN has a self-adapting sensor node set for data transmission. It is an auspicious technology has widespread application [1] such as environmental science, industry, agricultural computerization etc. However, the WSN application has perilous problems in terms of securing data by using cryptographic systems and routing protocols [2, 3]. Even though, these tactics are very important for identifying the optimal path in the network. But these tactics cannot prevent the unauthorized node attacks. When an unauthorized node receives data packets from a neighboring node, it discards the packets [4]. Similarly, it does not transmit data packets to its subsequent hop nearest node. It is called as black hole attack during the routing process of the WSN [5]. Consequently, it is essential for presenting some solutions to overcome these problems, which has prompted research in this area by incorporating the block chain proficiencies in the WSN [6].

Block chain is a data structure, which records all transaction information of the routing process. By the recorded transaction information, the block chain system cannot be alternated, and hacked [7]. Generally, block chain is a digital database of transactions, which is an immutable process due to its cryptography method for securing the block chain database. The block chain is based on the peer-to-peer network with multiple nodes without a central server [8]. It has four concerns namely confidentiality, integrity, non-repudiation and authentication. To overcome these four concerns, cryptography method is used in the block chain technology [9, 10]. The word “crypto” in cryptography means secret or hidden and the” graphy” word of cryptography means writing. The cryptography method generally performs encryption and decryption method based on key [11]. This feature can be useful for the wireless sensor network and its custom for block chain token communications, for recording the associated data of every routing node [12 –17]. For safeguarding the BC transaction, each BC system comprehends the explicit consensus procedure [18 –21].

Therefore, it is difficult for existing routing schemes to identify such malicious nodes, because the real-time change of the routing information between two routing nodes are difficult to be accurately distinguished. When a malicious node receives the packets of data from a neighbor node, it directly discards the packets and does not forward the packets of data to its next-hop neighbor node [22, 23]. This creates a data “black hole” in the network; hence it is named as black hole attack, which makes it difficult to detect routing nodes in WSNs. These malicious nodes can be attacked by external intrusion or internal legitimate nodes captured by external attackers. Hence, some solutions need to be put forward to fix these problems has motivated me to do research in this area.

In this manuscript, Proof-of-Authority (PoA) based consensus procedure for effective transaction in Block chain wireless sensor networks has been selected. By this, every routing node has its own registration bond. Here, the Crypto Hash signature (CHS) token are generated for the flow accesses with a secret key belonging to each routing sensor node, which also provides an optimal path for data transmission. Then the secured dynamic optimal routing information is delivered through the proposed Block chain based wireless sensor network platform with the help of Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf routing algorithm.

The main contribution of this manuscript is summarized as follows,

Block chain technology is used to deliver a distributed routing information management platform based on secure block chain token transactions.

Here, the Crypto Hash signature (CHS) token [24] is generated for flow accesses with a secret key belonging to each routing sensor node and also provides an optimal path for data transmission.

Then the secured dynamic optimal routing information is delivered through the proposed Blockchain based wireless sensor network platform with the help of Deep Dropout Extreme Machine learning [25] optimized Improved Alpha-Guided Grey Wolf [26] routing algorithm.

Then the proposed method is simulated using NS-2 (Network Simulator) tool.

The evaluation metrics like delay performance for 30% malicious node, delay performance for 60% malicious node, average latency with energy consumption, block chain token transactions throughput is measured.

Then the simulation performance of the proposed Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain Technology for secured dynamic optimal routing in Wireless Sensor Networks (SDOR-DEML-IAgGWO-CHS-BWSN) is compared with existing method like SDOR-DCNN-SSOA-BWSN [27], SDOR-BWSN [28], ASOR-BWSN [29], SDOR-LD2FA-PSO-WSN [30] and SDOR-APSO-FOA-WSN [31].

The remaining paper is structured as below. Section 2 delineates the Literature survey. Section 3 illustrates about Proposed Methodology. Section 4 demonstrates the result and discussion. Finally, Section 5 concludes the paper.

2 Literature survey

Several research works were already presented in literature based upon trusted routing protocol in block chain based wireless sensor networks, a certain recent research works are reviewed here,

Revanesh and Sridhar [27] have presented reliable allocating routing method for WSN by means of block chain and deep learning method. In this method, the block chain was used to manage distributed routing protocol with Salp Swarm Optimization algorithm for achieving optimal routing. The routing differences among the nodes were predicted, and then optimal routing results were finalized by DCNN approach. The presented routing scheme was executed in NS2, and then examined based on its latency, energy consumption, and throughput metrics. The efficiency of the presented method was 97%. The presented methods have good performance compared to other methods.

Mubarakali, [28] have presented an effective method for authentication based on block chain method for security purpose. In wireless sensor networks, the nodes were communicated via the base station, cluster heads and normal sensor nodes. The block chain network was formed in hierarchical form, with small including global chain, between certain network nodes. Nodes in the fusion models have dissimilar communication circumstances, individuality protected connection was recognized, and designated cluster node detection confirmation was performed. Experimental outcomes show that the presented method has forceful security and improved outcomes. Moreover, the computational ability attains more than 300 bytes per phase.

Lazrag et al., [29] have suggested a Block chain’s capability on prevailing reorganized networks creates it expressly appropriate to design the self-maintaining system on WSN devices. Nowadays, embedded systems and WSN are found promising areas of applications. The count of linked devices were affected to develop, because of insecure nature of these devices. It can be moderated the design with adoption of WSN security standards for creating secured environment on the technology. Therefore, the suggested method offers a routing protocol that utilizes Block chain technology to suggest the shared memory among the nodes of network.

Prithi and Sumathi [30] have suggested a deterministic Finite Automata (DFA) with Particle Swarm Optimization (PSO) for interruption identification and secure data transmission was executed by the optimized route. Learning Dynamic DFA with Particle Swarm Optimization offers the information about the node, packet with route inspection to detect and eliminate the interlopers, thus the data transmission was made in energy effective way for optimal path selection. The suggested method was simulated in MATLAB. The suggested method provides high throughput, high network lifetime and lesser energy consumes.

Pavani and Rao [31] have presented a secured cluster-based routing protocol based on adaptive particle swarm optimization optimized with firefly algorithm through data communication in WSN. The objective was used to reduce the energy consumption separate node and improving the entire lifespan of the network. The suggested secure cluster-based routing protocol contains energy-effective clustering, safe routing, safety verification. The performance was calculated by NS-3, and various metrics, viz time of encryption and decryption, energy consumption, packet drop rate, network lifespan was analyzed. The outcomes were likened to other methods. The outcomes show that the suggested secure cluster-based routing protocol provides better performance than existing methods.

Sun et al., [32] have presented a Secure Routing Protocol with Multi-objective Ant-colony-optimization for wireless sensor networks. The ant colony algorithm was used to enhance the multi-objective routing algorithm. The node trust valuation method was recognized with the pre-processing to evaluate the confidence degree of the nodes with an improved D-S suggestion system. The multi-objective routing outcomes were attained by Pareto optimal solution device by means of the exterior collection technique with an assembling distance principle. The replication outcomes shown with NS2, which the suggested algorithm can attain preferred presentation against the black hole occurrence in WSN routing.

Prithi and Sumathi [33] have presented a hybrid Particle Swarm Optimization–Grey Wolf Optimizer (PSO–GWO) for securing communication of data on increased path. A Learning Dynamic DFA (LD2FA) was introduced to study the active part of the atmosphere. The replication outcomes are got in MATLAB. The suggested method has increase the lifespan of network than other methods. It also demonstrates that the suggested method has throughput and less energy consumption compared to lightweight IDS. Hence, LD2FA based Hybrid PSO–GWO was deliberated to effectively consumes energy in the optimum path.

3 Proposed methodology

In this section, secured dynamic optimal routing for Wireless Sensor Networks is proposed with the help of Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Blockchain Technology. The overall work flow of proposed Wireless Sensor Networks is proposed with the help of Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Blockchain Technology is given in Fig. 1.

Fig. 1

Overall work flow for secured dynamic optimal routing for block chain based Wireless Sensor Networks.

3.1 Block chain based wireless sensor network architecture

In this section, the Blockchain based wireless sensor network architecture is discussed. The Blockchain Technology is commonly termed as disseminated record with interfere-resilient, restructured, and data perceptible. This feature can be useful for the wireless sensor network and its custom for blockchain token communications, recording the associated data of every routing node. Generally, blockchain based wireless sensor routing networks encompasses Server node, routing nodes and terminal devices. The nodes in the blockchain wireless sensor network are static or dynamic. For eg, server node mostly is in static form and the routing node are in dynamic form. So, the entrance and departure of nodes does not affect the blockchain based wireless sensor routing networks due to its dynamic status updation. Every routing node is responsible for distributing the packets and receiving from source to destination terminal node by means of routing policies attained from local learning approach. This local learning approach continuously enquiries the applicable state information of routing network from BC network and it helps for distributing the packets from source to destination terminal node. For safeguarding the blockchain transaction, every blockchain system comprehends the explicit consensus procedure.

In this research work, Proof-of-Authority (PoA) based consensus procedure is chosen for effective transaction in Blockchain wireless sensor networks. It contains substantiate and gofer identities for Proof-of-Authority (PoA) fostered Blockchain wireless sensor networks. Commonly, substantiate the pre-validated nodes of BC contains innovative endorsement. It is also responsible for Proof-of-Authority (PoA) fostered Blockchain confirmation. In this work, server node act as substantiate, which has sophisticated liberties during Proof-of-Authority (PoA) fostered Blockchain. It also has inimitable blockchain address. The role of substantiate is accomplishing bonds, confirming BC transactions, release blocks on BC. Then newly substantiate is included in the blockchain wireless sensor networks by means of election from the authenticated substantiates with the condition of getting 50% votes. If there is any malicious substantiate, it can only disturb a particular block and it can be removed based on the votes of substantiate. Then the gofer identity is generally termed as less-privileged nodes. It could not accomplish the authentication work on Proof-of-Authority (PoA) fostered Blockchain as like as substantiate identity. In this work, every routing node act as a gofer identity, which has less sophisticated liberties during Proof-of-Authority (PoA) Blockchain. It also has inimitable blockchain address. The role of gofer identity is to recruit the token bonds, activation of some bond functions, and enquires about the transaction information in the blockchain WSNs.

In blockchain based wireless sensor routing networks, it consists of diverse block chain Crypto Hash Signature tokens for epitomizing diverse packets to destination sensor nodes. The Crypto Hash Signature token is commonly termed as digitalized evidence related to specific packets stored in a smart bond. The routing nodes may recruit the token bonds for engendering Crypto Hash Signature tokens and it also records the routing information from specified packets. By this, Crypto Hash Signature token transactions will take place with source and destination nodes utilizing token bond. The wireless sensor network architecture depending on proposed Blockchain, every routing node has its own registration bond. The routing data contains the information about the address of the subsequent neighbor routing node, then information regarding total packets forward to the subsequent node with the timestamp. This routing data is approved by the BC Proof-of-Authority (PoA) consensus system using server nodes, also updates in BC wireless sensor network. The learning mode of the respective routing node may derive this information from the BC wireless sensor network, then advice the consequent routing strategy for routing node. The detailed discussion regarding Crypto Hash Signature Token fostered Block chain wireless sensor networks are given below,

3.1.1 Crypto Hash Signature Token fostered Block chain wireless sensor networks

In this section, Secure Hash algorithm (SHA-1) is used for Crypto Hash Signature Token fostered Block chain wireless sensor networks, which is executed for protecting the decentralized database of wireless sensor networks. Here, Crypto Hash signature (CHS) token are generated for the flow accesses with a secret key belonging to each routing sensor node and it also offer the optimal path for data transmission. Basically, the hash function is a mathematical function that considers the capricious length of numeric data, and then converts into a fixed length numerical data. Here, the input is in any length (smaller length or larger length), but the output will be in the fixed size. If the same input is used in multiple times, then the output of the hash value always is same. Similarly, if the input data (text) are changed, it automatically changes the hash value in a very different way. Then the actual input cannot get from the hash value. That means, the reverse is not possible. By this, hash function is not encryption because the reverse process is not possible. The Crypto Hash Signature acquires the input and produces the hash value is given in the following format, $\begin{matrix} Input : Crypto Hash Signature Token fostered \\ Block chain wireless sensor networks \\ Crypto Hash Signature of SHA - 1 hash value : \\ 5 CB 0965 DF 4 B 2 C 594 F 28 CCA 93187322 C 185265 B 1 F \end{matrix}$

Generally, the crypto hash techniques are done based on the following Equation (1) $SHA - 1 Function (transaction) = Output generation$ (1)

Then the public key is generated is based on the following Equation (2) $Public key = {KG}^{transaction} mod (prime number)$ (2)

The conflict confrontation is utilized through crypto hash techniques for two diverse transactions which produce the same hash value. If the same hash value is obtained for two transactions, which is generally called as brute force attack and it is given in the following (3) $\begin{matrix} Hash value (transaction 1) = \\ Hash value (transaction 2) \end{matrix}$ (3)

The first iterative function of the input transaction key is based on the following Equation (4) $Transaction z_{+}^{1} = Hash value z_{+}^{1}$ (4)

Where the hash value of first iterative function of input transaction is given in the following Equation (5) $H a s h v a l u e_{z + 1} = b l o c k c i p h e r (H a s h v a l u e_{z}, T r a n s a c t i o n_{z + 1}) \oplus H a s h v a l u e_{z}$ (5)

Conversely, the sensor node endorsement authorizes the sensor node register instead of storing passwords. Then it submits in the course of the Crypto Hash Signature function before database storage. Then this result is used for sensor node confirmation. The sensor node confirmation is given in the Fig. 2.

3.2 Block chain based wireless sensor network Procedure

Initially the routing network information is required to transmit the block chain network for activating the block chain-based routing network architecture. The routing network information is documented in the smart bond, which comprises the information regarding registration bond, Crypto Hash Signature Token bond, and Crypto Hash Signature Token fostered block chain transactions. This information is demonstrated with the help of legitimate server nodes, and then the result of execution is transferred to the block chain network. The registration bond contains information regarding the uniqueness characteristics of all routing node and server node, which is responsible for asking query related to the entire block chain network node. Generally, the registration bond contain inbuilt capricious such as mapping plot, which are responsible for plotting the block chain addresses to the uniqueness characteristics of all routing node and server node. Similarly, the registration bond contains inbuilt capricious, such as mapping status that is responsible for accumulating the node status, whether it is recorded or not. If a newly node needs for registering the identity information at the register bond, it supposed to activate the bond as bond caller. By this, registration bond automatically records the new node information in its block chain address. Then the registration bond clarifies the new node mapping status, whether it is survived or not. If the status of the block chain address = 0, then it automatically maps the block chain address, which is equal to the identity information of new node. Then the status of the block chain address is updated to 1 in status array, which means the operation is a success. Similarly, if the status of the block chain address = 1, then it automatically shows the registration operation is a failure. After registration, the registration data are not changed. The procedure for Crypto Hash Signature Token fostered block chain transactions are given below in detailed manner,

Fig. 2

Sensor node confirmation.

In the block chain based wireless sensor networks; it verifies the routing process of all routing sensor nodes. The working procedure for block chain based wireless sensor networks token transaction is given in the Fig. 3. By this, the routing sensor nodes get the relevant routing information from block chain based wireless sensor routing networks. Initially the source node transmits the packets to the destination node and it activates the “transfer” function on the Crypto Hash Signature Token bond for changing the status. The status contains information regarding the token amount which was transmitted to next-hop routing node. Then the token amount is issued on the Crypto Hash Signature Token bond. By this, next-hop routing node activates the “confirm” function on the Crypto Hash Signature Token bond, which authenticates the acknowledged packet quantity to the block chain based wireless sensor networks. Then the entire token transaction is authenticated with the help of server node consensus algorithm. This entire process is completed in one time slot. Then the unofficial token transactions are abandoned in the block chain based wireless sensor networks, similarly the failed token transactions are not evidenced in the block chain based wireless sensor networks without concerning the routing information. By this, block chain based wireless sensor routing networks provide secure routing environment.

Fig. 3

Working procedure for block chain based wireless sensor networks token transaction.

But the routing information of block chain based wireless sensor networks is not successfully exploited and it needed to improve the performance. For that, Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf Block chain based routing algorithm are used to learn every sensor node’s information. The detailed discussion regarding Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf Block chain based routing algorithm is given below in the following section.

3.3 Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf Block chain based routing algorithm

In this section, secured dynamic optimal routing information is delivered through the proposed Block chain based wireless sensor network platform is discussed with the help of Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf routing algorithm. Generally, the secured dynamic optimal routing information is about the Crypto Hash Signature Token transaction’s timestamp, then the allocated quantity of Crypto Hash Signature Tokens for each destination routing node, then the enduring quantity of Crypto Hash Signature Tokens and about every authorized destination node’s routing address. Initially, the input parameters are initialized with real routing platform, initial position, regulation constraint and discriminant data constraint. Generally, Block chain based wireless sensor network platform are used for updating the routing node position. Based on this, verdict is taken for doing particular routing actions. The position P implies current state of the packets should be allocated. The “Procedure” is about the accomplishment taken for the delivering packet from the current spot to the next-hop routing node. The action space is generated for collecting all accomplishment at the routing node. But sometimes, it has different queue length for each routing node. This different queue length affects the probability of transferring packets between two routing nodes. To overcome this proposed Deep Extreme Machine learning generates the action space, whose components are sinked autonomously from a Bernoulli distribution. In the deep convolutional neural network, Softmax layer is removed with the help of dropout procedure. This is called as deep dropout Extreme Machine learning. Then dropout procedure is done in the routing nodes with generated action space, which are obtained from the proposed Deep Extreme Machine learning. It is denoted as L_K_. The L_K means the Lagrange function of Deep Extreme Machine learning and it is given in the following Equation (6) $L_{K} = \frac{1}{2} χ^{T} F χ + \frac{1}{2} R \sum_{x = 1}^{A} ℑ^{2} + \sum_{x = 1}^{A} \sum_{y = 1}^{B} (δ_{x, y} χ_{y} \times g (X_{x}) - t_{xy} + ℑ_{xy})$ (6) where $\frac{1}{2} χ^{T} F χ + \frac{1}{2} R \sum_{x = 1}^{A} ℑ^{2}$ represents the loss function of extreme machine learning, ℑ represents the training error, g (.) represents the activation function, χDenotes the regularization constraint to the loss function of extreme machine learning and it is given in the following Equation (7) $χ = {\begin{matrix} {MP}^{T} {[\frac{1}{R} + {MP}^{T} * MP]}^{- 1} \times T; & if NoS ⩽ NoHLN \\ {[\frac{1}{R} + {MP}^{T} * MP]}^{- 1} \times {MP}^{T} T; & if NoS > NoHLN \end{matrix}$ (7) where MP represents the Moore-Penrose generalized inverse of the matrix, R represents the regularization, shift matrix T is determined by the given Equation (8) $T = MP \times χ$ (8)

Similarly, NoS represents the samples number, NoHLN represents the hidden layer neuron number, F represents the configuration dissimilarity matrix and it is calculated with the help of Equation (9) $F = F_{ICSM} - (1 - ℘) F_{WCSM}$ (9) where F_ICSM represents the inter-class scatter matrix and it is given in the following Equation (10) $F_{ICSM} = \sum_{x = 1}^{C} \sum_{y = 1}^{NoC} (X_{y} - {mean}^{x}) {(X_{y} - {mean}^{x})}^{T}$ (10) where C represents data sample categories number, NoC represents categories number, mean^x represents the mean of xth class, F_WCSM represents the intra-class scatter matrix and it is given in the following Equation (11) $F_{WCSM} = \sum_{x = 1}^{C} ({mean}^{x} - mean) {({mean}^{x} - mean)}^{T}$ (11) where ℘ represents the parameter value used for adjusting the inter-class scatter matrix and intra-class scatter matrix, where the values are in the range of 0 and 1. Then train the Deep Dropout Extreme Machine learning with different values. Then create an n-dimensional binary random vector, whose fundamentals are obscure autonomously hidden from the Bernoulli distribution. Then dropout technique is applied in the deep convolutional neural network hidden nodes with the generated n-dimensional binary random vector, which was obtained from the extreme machine learning. So it is called as deep dropout extreme machine learning. Then the probability distribution is obtained based on the Lagrange function of deep dropout extreme machine learning is given in the following Equation (12) ${P_{1 x} (X), P_{2 x} (X), . . . . . . . ., P_{Kx} (X)}$ (12)

Then test the deep dropout extreme machine learning with Lagrange function. Here, the Soft Max layer in the deep extreme machine learning is removed with the help of dropout procedure. Then the label vector is calculated with the help of lagrange function from deep dropout extreme machine learning. Based on majority voting, the “Procedure” is measured with the help of Equation (13) $Pr {ocedure}_{L} (X) = {MAX}_{1 ⩽ L ⩽ Z} \sum_{x = 1}^{Y} P_{xL} (X)$ (13)

Generally, the output of “Procedure” is obtained, which means accomplishment taken for delivering packet to the next-hop routing node from the current spot. In the deep dropout extreme machine learning, probability distribution based on the Lagrange function of deep dropout extreme machine learning is a significant constraint for encouraging the routing node to make better routing decisions. The probability distribution based on the Lagrange function of deep dropout extreme machine learning is determined by means of efficaciously distributed tokens quantity. For example, if the probability distribution value is small, that means there is a malicious node which does not communicate any packet process of routing. By this, the delinquent of hunking routing connection occurred by malicious black hole nodes has suggestively abridged. By this, the proposed deep dropout extreme machine learning [25] is used for forecasting the actual-time changing of routing information amid the 2 routing nodes, which favor near optimal routing decisions. But the optimal parameters of deep dropout extreme machine learning should be optimal for dynamic optimal routing.

In this work, improved alpha-guided grey wolf optimizer [26] can be utilized to optimize the deep dropout extreme machine learning for finding the optimal parameters. Here, improved alpha-guided grey wolf optimization (GWO) has been utilized to tune the hyper parameters with the help of deep dropout extreme machine learning. Usually, some methods are used for parameter configuration such as grid, manual and random search model. Nevertheless, these search method are taking its peculiar weakness concerning iteration time is no stratagem-built knowledgeable exploration.

Improved alpha-guided GWO is a metaheuristic algorithm that impersonators the pyramid edifice and hunting mechanism of grey wolf. The main steps of the improved alpha-guided grey wolf optimizer are trailing and approaching the target, hurtling and encompassing the target, assaultive the target. By this, it is accumulated a prospect model for discovering the dynamic optimal routing. In this paper, improved alpha-guided grey wolf optimizer is selected; it takes fewer iteration times than various tuning model, via grid search, random search finds the optimum hyper parameters. The flowchart for Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf Block chain based routing algorithm is given in Fig. 4. The step-by-step process of deep dropout extreme machine learning using improved alpha-guided GWOare given below

Fig. 4

Flowchart for Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf Block chain based routing algorithm.

Step 1: Initialization

Initialize the initial population of grey wolf with the process value of the routing node with real routing platform, initial position, regulation constraint and discriminant data constraint. For that encircling strategy is used for finding initial position, it is measured by Equation (14, 15) $E = | F * {position}_{Pr ocedure (X)} (Z) - position (Z) |$ (14) $position (Z + 1) = {position}_{Pr ocedure (X)} (Z) - G * H$ (15)

Where Z is the current iteration, position_Procedure(X) denotes the position of Pr ocedure (X) (optimal secured routing), position (Z) denotes the position of grey wolf, G and H are the coefficient vector, which is computed utilizing Equation (16, 17) $G = 2 g * {random}_{1} - g$ (16) $H = 2 * {random}_{2}$ (17) where, the component of g is diminishes linearly from 2 to 0 with iteration, random₁ and random₂ denotes the random number, which is given in the value of 0 and 1.

Step 2: Random generation

After the process of initialization, the input parameters have been generated randomly. Here, the greatest fitness values are selected depending on explicit hyper-parameter situation. Moreover, randomly generate the population of Procedure value for routing node.

Step 3: Fitness Function

The arbitrary quantity of resolutions is engendered after the initialized values. Then, the fitness function is generally employed for obtaining the objective function, such as secure dynamic optimal routing given in the following dynamic optimal secure routing.

Step 4: Updation of position for secure dynamic optimal routing

In this step, updation of position for alpha guided wolf is one of the tuning parameters for the deep dropout extreme machine learning. Updation of position for alpha guided wolf works based on the following procedure. Initially, the maximum iteration is done in the “Procedure” (from Equation 13). It checks the fitness value of best dynamic optimal secure routing F (position_procedure (ζ) = position_procedure (ζ) ^* = output) = best dynamic optimal secure routing

The second dynamic optimal secure routing is denoted as F (position_procedure (ψ))

The third dynamic optimal secure routing is denoted as F (position_procedure (ς))

The remainder routing is denoted as F (position_procedure (ϖ))

F (position_procedure (ζ) is better than fitness value of best dynamic optimal secure routing F (position_procedure (ζ) ^*, which is helpful for finding optimal hyper-parameter setting and it resolute utilizing the given equations (18–24) $E_{ζ} = | F_{1} * {position}_{Pr ocedure (X)} (ζ) - position |$ (18) $E_{ψ} = | F_{2} * {position}_{Pr ocedure (X)} (ψ) - position |$ (19) $E_{ς} = | F_{3} * {position}_{Pr ocedure (X)} (ς) - position |$ (20) ${position}_{1} = {position}_{Pr ocedure (X)} (ζ) - G_{1} * H_{ζ}$ (21) ${position}_{2} = {position}_{Pr ocedure (X)} (ψ) - G_{2} * H_{ψ}$ (22) ${position}_{3} = {position}_{Pr ocedure (X)} (ς) - G_{3} * H_{ς}$ (23) $\begin{matrix} position (Z + 1) = \\ \frac{[{position}_{1} (Z) + {position}_{2} (Z) + {position}_{3} (Z)]}{3} \end{matrix}$ (24)

Formerly discover the loss value subject to optimal hyper-parameter setting, then the value of loss resolute by Equation (25) $\begin{matrix} position (Z + 1) = [0.5 \times {position}_{1} (Z) + 0.3 \\ \times {position}_{2} (Z) + 0.2 \times {position}_{3} (Z) \end{matrix}$ (25)

At that time, this equivalent loss value including hyper-parameter setting is deposited on equivalent paths. These equivalent paths are utilized to the parameter settings with the purpose of evaluations. The Updation of equivalent paths with the help of the following Equation (26, 27) $\begin{matrix} {position}_{Pr ocedure (X)} (ζ)^{*} = MAX {position}_{Pr ocedure (X)} (ζ) \\ + Min {position}_{Pr ocedure (X)} (ζ) - {position}_{Pr ocedure (X)} (ζ) \end{matrix}$ (26) $\begin{matrix} {position}_{Pr ocedure (X)} (ζ)^{* *} = Guassian (1, 1) \\ * {position}_{Pr ocedure (X)} (ζ) \end{matrix}$ (27)

By this, two worst fitness functions for dynamic optimal secure routing are replaced. Then update the Equation (16, 17). Then calculate the fitness of all dynamic optimal secure routing path.

Step 6: Termination

Here, the optimum hyper-parameter have been chosen at deep dropout extreme machine learning with the help of Improved alpha-guided grey wolf optimizer will iteratively repeat step 3 until the halting criteria is met. Then finally deep dropout extreme machine learning finds the best dynamic optimal secure routing with the help of Improved alpha-guided grey wolf optimizer.

3.4 Analysis of Security in Block chain based wireless sensor network

In this section, secured dynamic optimal routing analysis of proposed Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain Wireless Sensor Networks is discussed. Here the secured dynamic optimal routing information management system is implemented based on the routing transaction of routing node and server node. The security analysis is explained based on six viewpoints such as proof of authority consensus procedure, block chain transaction perceptibility, Routing data source, single point attack avoidance, double-spending problem prevention and self-amendment features, which are discussed below,

a) Proof of authority consensus procedure

The Proof of authority consensus procedure is used in the proposed block chain wireless sensor network, which verifies the validated server node for uploading the transaction (packet transmission) by routing information updation. By this, hackers unable to interfere with the routing information of proposed BC wireless sensor network.

b) Block chain transaction perceptibility

The block chain transaction perceptibility is done with the help of validated server node. This validated server node records the issuing token bond transaction, the routing node transaction functioning in the bond, the allocating tokens transaction on the proposed block chain wireless sensor network. By this, all the transaction data are recorded in every block of the proposed block chain wireless sensor network, which cannot be perceptible transversely throughout the block chain network.

c) Routing data source

In the proposed block chain wireless sensor network, every routing sensor node gets the appropriate routing data source. By this, the routing data source cannot be determined by the hackers.

d) Single point attack avoidance

In the proposed block chain wireless sensor network, it does not entail the secured third-party chief expert for accomplishing the routing information. By this, attack is avoided through multiple server node transaction verification.

e) Double-spending problem prevention

In proposed Crypto Hash Signature Token bond stipulates records only one address on every time slot from every routing node address. By this, the routing node could not recruit any token transactions for different routing nodes at similar time slot. By this, double spending problem is prevented.

f) Self-amendment

In the Proposed Deep Dropout Extreme Machine learning optimized with Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Blockchain wireless sensor routing protocol does not generate any routing link transaction with malicious sensor node. Because the regularization constraint value of the routing link is very low from the proposed Block chain routing algorithm.

4 Results and discussions

This segment describes the effectiveness and simulation performance of the proposed Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain Technology for secured dynamic optimal routing in Wireless Sensor Networks. The proposed method is simulated by NS-2 (Network Simulator) tool through wireless sensor module and the coding is done in C++. The data regarding sensor node are sent to the gateway by utilizing Geth 1.8.19a standard protocol. The evaluation metrics like average latency, average energy consumption, BC token transactions throughput is measured. Then the simulation performance of the proposed Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain Technology for secured dynamic optimal routing in Wireless Sensor Networks (SDOR-DEML-IAgGWO-CHS-BWSN) is compared with existing methods, like trusted distributed routing scheme for WSNs using Deep Convolutional Neural network algorithm optimized with Salp Swarm Optimization algorithm fostered Block chain Technology (SDOR-DCNN-SSOA-BWSN) [27], Efficient with secured routing protocol depending on BC technique for WSNs (SDOR-BWSN) [28], Efficient Authentication Scheme Utilizing Block chain Technology for WSNs (ASOR-BWSN) [29], an innovative learning dynamic DFA with PSO algorithm for secured energy efficient routing in WSN (SDOR-LD²FA-PSO-WSN) [30] and Adaptive particle swarm optimization along optimized firefly algorithms for securing cluster-based routing in wireless sensor networks (SDOR-APSO-FOA-WSN) [31].Table 1 tabulates the simulation parameters.

Table 1
Simulation parameter

Parameters Values

Simulation area 1200 × 1200m

MAC layer used Geth 1.8.19a

Frequency 9.47 ×10⁸

Data rate for transmission in network 50 Mbps

Size of packet 1024 bytes

Network load 1000–3000 Kbps

Range of transmission 250m

Number of frames per packet 5

Arrival rate 2(packets/slot)

Concurrent request rate 4500 (times/second)

Parameters	Values
Simulation area	1200 × 1200m
MAC layer used	Geth 1.8.19a
Frequency	9.47 ×10⁸
Data rate for transmission in network	50 Mbps
Size of packet	1024 bytes
Network load	1000–3000 Kbps
Range of transmission	250m
Number of frames per packet	5
Arrival rate	2(packets/slot)
Concurrent request rate	4500 (times/second)

4.1 Testing settings

The testing setting of the proposed Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Blockchain Technology for secured dynamic optimal routing in WSNs is analyzed. Here, 30 virtual servers are used for testing updation of BC transactions chain. Every routing data need through the Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain learning approach is acquired from the public BC transactions. The conglomerate BC is constructed depending upon Geth 1.8.19a., that offers consistent Ethereum transaction facilities. The real packet arrival rates are measured based on 30 terminals from the simulation network area. From the simulation network area, the routing nodes are simulated for receiving and delivering the actual packets from one packet slot with the help of routing policy developed via Deep Dropout Extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain learning model. The information with the average delay with energy consumption and the performance of BC token transactions is lastly documented. The device specification is given in Table 2.

Table 2
Device specification

Restriction Server Block Terminal Block Routing Block

Name chain sensor chain sensor chain sensor

Node Node Node

CPU 3GHz 1.5GHz 600 MHz

RAM 32GB 2GB 64MB

Storage 1 TB 32 GB 512 MB

Network 1100 Mb 110 Mb 110 Mb

Operating system Windows 7 Windows 7 Windows 7

Restriction	Server Block	Terminal Block	Routing Block
CPU	3GHz	1.5GHz	600 MHz
RAM	32GB	2GB	64MB
Storage	1 TB	32 GB	512 MB
Network	1100 Mb	110 Mb	110 Mb
Operating system	Windows 7	Windows 7	Windows 7

4.2 Performance metrics

The evaluation metrics like average latency, average energy consumption, and throughput of BC token transactions is measured.

Delay

In this experiment, the delay performance is measured with malevolent nodes (30% and 60%) in the simulation area. Here three kinds of malevolent nodes are specified with the same probability. The first kind is malevolent nodes issues counterfeit short queue length data, which communicates the data packets to other routing nodes. The second kind is malevolent nodes issues exact queue length data, which does not communicate the data packets to other routing nodes. The third kind is malevolent nodes issues counterfeit short queue length data, which does not communicate the data packets to other routing nodes.

Average Token Transaction Latency

The average Token Transaction Latency is measured based on the transaction packaging time. It also archives the elapsed time while miner’s place the Ethereum token transaction at BC with increase arrival rate.

Average Token Transaction energy consumption

Average Token Transaction energy consumption is a distinct unit used in Ethereum networks for computing works are done by the attacker. This resolute through the count of system instructions functioned utilizing Ethereum transaction. For Ethereum platform bond implementation, it required firm amount of energy consumption. For extra computational resources, it required more Token Transaction energy consumption. Then, the Token Transaction energy consumption converts as related ether currency for paying the BC miner

Token transaction throughput

It measures the BC system’s capability to grip simultaneous token transactions.

Effectiveness (Vulnerability Detection)

Overall vulnerability periods is defined as the time required to detect attacks versus the percentage of newly introduced attacks in the network

Network Overhead

The time takes to transmit data on a packet-switched network. Each packet requires extra bytes of format information that is stored in the packet header, which, combined with the assembly and disassembly of packets, reduces the overall transmission speed of the raw data.

Efficiency (Block Generation Latency)

Blockchain network latency is the time between submitting a transaction to a network and the first confirmation of acceptance by the network. Block latency is based on the ratio of attack transactions in the network. Attack ratio is calculated based on below Equation (28) $Attack ratio = \frac{no . of attac ker s in thent work}{no of nodes in the network}$ (28)

4.3 Simulation results

Figures 5 to 6 shows the simulation result of average packet delivery delay performance with 30% and 60% malevolent nodes. Figures 7 to 9 shows the simulation result of average latency, average energy consumption, and network throughput of BC token transactions.

Fig. 5

Average packet delivery delay performance with 30% malevolent nodes.

Fig. 6

Average packet delivery delay performance with 60% malevolent nodes.

Fig. 7

Average transaction latency performance of BC scheme.

Fig. 8

Average transaction energy consumption performance of block chain system.

Fig. 9

Average transaction throughput performance of block chain system.

4.4 Discussion

Figure 5 shows the Average packet delivery delay performance with 30% malevolent nodes. At arrival rate 0, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 71.428%, 63.636%, 55.555%, 46.667% and 33.333% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 0.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 75%, 66.629%, 57.143%, 53.846% and 45.454% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1.0, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 75%, 64.285%, 58.333%, 44.444% and 50% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 79.959%, 66.667%, 66.611%, 50% and 42.857% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 2, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 79.959%, 66.667%, 66.611%, 50% and 42.857% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Figure 6 shows the Average packet delivery delay performance with 60% malevolent nodes. At arrival rate 0, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 67.741%, 58.33%, 50%, 37.5% and 28.57% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 0.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 69.465%, 60%, 55.556%, 38.461% and 32.203% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1.0, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 75%, 66.67%, 62.5%, 50% and 45.454% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 76.627%, 67.058%, 64.102%, 49.09% and 44% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 2, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 76.47%, 67.058%, 64.102%, 48.905% and 44% lower delay compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Figure 7 shows the Average transaction latency performance of block chain system. At arrival rate 0, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 62.711%, 60%, 58.49%, 55.102% and 52.173% lower average transaction latency performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 0.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 64.912%, 61.53%, 60%, 57.44% and 54.545% lower average transaction latency performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 67.272%, 64%, 62.5%, 59.09% and 55% lower average transaction latency performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 68.518%, 64.583%, 62.222%, 58.536% and 55.263% lower average transaction latency performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 2, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide % lower average transaction latency performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Figure 8 shows the Average transaction energy consumption performance of block chain system. At arrival rate 0, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 55.556%, 53.488%, 50%, 47.368% and 41.176% lower average transaction energy consumption performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 0.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 59.09%, 56.097%, 52.631%, 52.631%, and 43.75% lower average transaction energy consumption performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 64.285%, 61.538%, 59.459%, 55.882% and 50% lower average transaction energy consumption performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 1.5, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 70%, 67.56%, 64.705%, 62.5% and 57.142% lower average transaction energy consumption performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At arrival rate 2, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 73.684%, 71.428%, 69.697%, 66.667% and 61.538% lower average transaction energy consumption performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Figure 9 shows the Average transaction throughput performance of block chain system. At concurrent request rate 500, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 78.66%, 42.011%, 41.732%, 41.732% and 41.453% higher average transaction throughput performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At concurrent request rate 1500, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 77.647%, 51%, 25.833%, 20.8% and 0.667% higher average transaction throughput performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At concurrent request rate 2500, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 90%, 85.185%, 63.398%, 51.515%, and 42.857% higher average transaction throughput performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At concurrent request rate 3500, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 86.667%, 84.603%, 48.571%, 30.46%, and 18.97% higher average transaction throughput performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At concurrent request rate 4500, the proposed SDOR-DEML-IAgGWO-CHS-BWSN provide 88.889%, 85.181%, 63.398%, 50.178%, and 31.788% higher average transaction throughput performance of block chain system compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Figure 10 shows the Comparison of Overall Vulnerability Periods in Networks. In this simulation, there were seven different potential attacker nodes in the network. The newly introduced attack ratios varied from 0% to 100%, as shown in Figure 10. For attack ratio 100, the detection time of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 87.5%, 45.4%, 55.53%, 49.66% and 42.6% lower detection time compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. For attack ratio 60, the detection time of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 76.6%, 61.0%, 55.8% 32.7% and 34.56% lower detection time compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. For attack ratio 20, the detection time of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 76.6%, 61.0%, 55.8%, 32.7% and 34.56% lower detection time compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. For attack ratio 0, the detection time of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 24.12%, 57.28%, 83.13%, 12.76% and 44.36% lower detection time compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Fig. 10

Comparison of Overall Vulnerability Periods in Networks.

Figure 11 represents an overhead cost comparison of proposed method between the existing methods like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. Overhead is measured based on the average size of packets. For time interval 50 s, the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 36.12%, 45.28%, 63.13%, 34.5% and 49.20% lower average size of packets compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. For time interval 100 s, the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 65.45%, 55.34%, 46.3%, 47.45% and 54.53% lower average size of packets compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. For time interval 150 s, the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 43.5%, 56.67%, 45.64%, 29.66% and 39.67% lower average size of packets compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Fig. 11

Comparison of Average Overheads.

Figure 12 shows the comparison of Block Generation Latency Based on Attack Transactions. At the attack ratio 0.1, transaction block latency of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 27.192%, 45.53%, 38.77%, 58.4% and 23.455% lower transaction latency than the existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At the attack ratio 0.2, transaction block latency of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 24.192% and 22.455% lower transaction latency than the existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively. At the attack ratio 0.3, transaction block latency of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides 25.192%, 56.75%, 49.66%, 28.56% and 20.455% lower transaction latency than the existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN respectively.

Fig. 12

Block Generation Latency Based on Attack Transactions.

4.5 Justification of result

A trusted routing scheme is very important to ensure the routing security and efficiency of wireless sensor networks (WSNs). The paper proposes a Deep Dropout extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Blockchain Technology for secured dynamic optimal routing in Wireless Sensor Networks (SDOR-DEML-IAgGWO-CHS-BWSN). As a decentralized system, the blockchain network provides a feasible scheme for routing information management and a platform for Deep Dropout extreme Machine learning of routing scheduling. Then use the blockchain token to represent the routing packets, and each routing transaction is released to the blockchain network through the confirmation of the validator nodes. By making every routing transaction recorder traceable and tamper-proof, routing nodes can obtain dynamic and trusted routing information on the blockchain network. In this Simulation results illustrate that the proposed SDOR-DEML-IAgGWO-CHS-BWSN method attains better results compared with the SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN method in terms of delay performance for 30% malicious node, delay performance for 60% malicious node, average latency with energy consumption, block chain token transactions throughput, overall Vulnerability Periods in Networks, Average Overheads and Block Generation Latency Based on Attack Transactions. The proposed SDOR-DEML-IAgGWO-CHS-BWSN method provides lower delay performance for 30% malicious node 15.068%, 27.059%, 20.513%, 38.58% and 43.87%, lower delay performance for 60% malicious node 9.23%, 21.33%, 18.56%, 23.64% and 31.91%, lower latency with energy consumption 18.64%, 30.43%, 28.35%, 32.67% and 45.39%, higher token transactions throughput 48.27%, 40.98%, 21.12%, 58.31% and 56.34%, lower Vulnerability Periods in Networks 16.21%, 8.86%, 7.50%, 45.56% and 23.31%, lower Overheads 21.75%, 7.55%, 5.56%, 47.69% and 58.47% and lower Block Generation Latency Based on Attack Transactions 39.86%, 38.98%, 33.38%, 52.41% and 36.69% compared with existing methods like trusted distributed routing method for WSNs utilizing Deep Convolutional Neural network algorithm optimized with Salp Swarm Optimization algorithm fostered Block chain Technology (SDOR-DCNN-SSOA-BWSN), Efficient including secured routing protocol depending on Block chain model for WSNs (SDOR-BWSN), Efficient Authentication Scheme Utilizing the Technology of Block chain for WSNs (ASOR-BWSN), An innovative learning dynamic deterministic finite automata along particle swarm optimization approach for secured energy efficient routing at WSN (SDOR-LD2FA-PSO-WSN) and Adaptive particle swarm optimization with optimized firefly algorithms for secured cluster-based routing in WSNs (SDOR-APSO-FOA-WSN).

5 Conclusion

In this manuscript, a Deep Dropout extreme Machine learning optimized Improved Alpha-Guided Grey Wolf based Crypto Hash Signature Token fostered Block chain Technology is successfully implemented for secured dynamic optimal routing in Wireless Sensor Networks (SDOR-DEML-IAgGWO-CHS-BWSN). Here security analysis is analyzed in the proposed block chain wireless network based on six viewpoints such as proof of authority consensus procedure, block chain transaction perceptibility, Routing data source, single point attack avoidance, double-spending problem prevention and self-amendment. Then the proposed method is simulated using the NS-2 (Network Simulator) tool. Then the simulation performance of the proposed SDOR-DEML-IAgGWO-CHS-BWSN method provide 66.64%, 63.06%, 61.2%, 57.83% and 53.98% lower Average transaction latency performance of block chain system, 64.52%, 62.02%, 59.29%, 57% and 50.72% lower Average transaction energy consumption performance of block chain system, 88.37%, 86.79%, 85.94%, 75.25% and 56.78% higher Average transaction throughput performance of block chain system, which are compared with existing method like SDOR-DCNN-SSOA-BWSN, SDOR-BWSN, ASOR-BWSN, SDOR-LD²FA-PSO-WSN and SDOR-APSO-FOA-WSN.

References

Rida

J.F.A.

Development of a remote health care wireless sensor network based on wireless spread spectrum communication networks, Materials Today: Proceedings, (2021).

Mousavi

S.K.

, Ghaffari

, Besharat

and Afshari

, Security of internet of things based on cryptographic algorithms: a survey, Wireless Networks 27(2) (2021), 1515–1555.

Yadav

D.K.

, Jayanthu

, Das

S.K.

, Chinara

and Mishra

, Critical review on slope monitoring systems with a vision of unifying WSN and IoT, IET Wireless Sensor Systems 9(4) (2019), 167–180.

Narayanan

K.L.

, Krishnan

R.S.

, Julie

E.G.

, Robinson

Y.H.

, Shanmuganathan

Machine Learning Based Detection and a Novel EC-BRTT Algorithm Based Prevention of DoS Attacks in Wireless Sensor Networks, Wireless Personal Communications (2021), 1–25.

Kaushik

, Sharma

Black hole attack and its security measure in wireless sensors networks, In Handbook of Wireless Sensor Networks: Issues and Challenges in Current Scenario’s, (2020), 401–416. Springer, Cham.

Goyat

, Kumar

, Alazab

, Saha

, Thomas

and Rai

M.K.

, A secure localization scheme based on trust assessment for WSNs using blockchain technology, Future Generation Computer Systems 125 (2021), 221–231.

Min

, Blockchain technology for enhancing supply chain resilience, Business Horizons 62(1) (2019), 35–45.

Wang

, Guo

, Ai

and Cao

, RBT: A distributed reputation system for blockchain-based peer-to-peer energy trading with fairness consideration, Applied Energy 295 (2021), 117056.

Saberi

, Kouhizadeh

, Sarkis

and Shen

, Blockchain technology and its relationships to sustainable supply chain management, International Journal of Production Research 57(7) (2019), 2117–2135.

10.

Gupta

, Tanwar

, Kumar

and Tyagi

, Blockchain-based security attack resilience schemes for autonomous vehicles in industry 4.0: A systematic review, Computers&Electrical Engineering 86 (2020), 106717.

11.

Luo

, Ouyang

, Liu

and Cao

, An image encryption method based on elliptic curve elgamal encryption and chaotic systems, IEEE Access 7 (2019), 38507–38522.

12.

Liu

, Yu

F.R.

, Li

, Ji

and Leung

V.C.

, Blockchain and machine learning for communications and networking systems, IEEE Communications Surveys& Tutorials 22(2) (2020), 1392–1431.

13.

Aggarwal

, Chaudhary

, Aujla

G.S.

, Kumar

, Choo

K.K.R.

and Zomaya

A.Y.

, Blockchain for smart communities: Applications, challenges and opportunities, Journal of Network and Computer Applications 144 (2019), 13–48.

14.

Mythili

, Thiyagarajah

, Rajesh

and Shajin

F.H.

, Ideal position and size selection of unified power flow controllers (UPFCs) to upgrade the dynamic stability of systems: an antlion optimiser and invasive weed optimisation algorithm, HKIE Trans 27(1) (2020), 25–37.

15.

Rajesh

and Shajin

F.A.

, Multi-Objective Hybrid Algorithm for Planning Electrical Distribution System, European Journal of Electrical Engineering 22(1) (2020), 224–509.

16.

Shajin

F.H.

, Rajesh

Trusted secure geographic routing protocol: outsider attack detection in mobile ad hoc networks by adopting trusted secure geographic routing protocol, International Journal of Pervasive Computing and Communications, (2020).

17.

Thota

M.K.

, Shajin

F.H.

and Rajesh

, Survey on software defect prediction techniques, International Journal of Applied Science and Engineering 17(4) (2020), 331–44.

18.

Ghosh

, Gupta

, Dua

and Kumar

, Security of Cryptocurrencies in blockchain technology: State-of-art, challenges and future prospects, Journal of Network and Computer Applications 163 (2020), 102635.

19.

Teeluck

, Durjan

, Bassoo

Blockchain Technology and Emerging Communications Applications, In Security and Privacy Applications for Smart City Development, (2021), 207–256.

20.

Patil

, Sangeetha

, Bhaskar

Blockchain for IoT access control, security and privacy: a review, Wireless Personal Communications, (2020), 1–20.

21.

Makhdoom

, Zhou

, Abolhasan

, Lipman

and Ni

, PrivySharing: A blockchain-based framework for privacy-preserving and secure data sharing in smart cities, Computers& Security 88 (2020), 101653.

22.

Al-Andoli

, Tan

S.C.

, Cheah

W.P.

Parallel stacked auto encoder with particle swarm optimization for community detection in complex networks, Applied Intelligence, (2021), 1–21.

23.

Al-Andoli

M.N.

, Tan

S.C.

, Cheah

W.P.

and Tan

S.Y.

, A Review on Community Detection in Large Complex Networks from Conventional to Deep Learning Methods: A Call for the Use of Parallel Meta-Heuristic Algorithms, IEEE Access 9 (2021), 96501–96527.

24.

Qazi

, Qureshi

K.N.

, Bashir

, Islam

N.U.

, Iqbal

and Arshad

, Security protocol using elliptic curve cryptography algorithm for wireless sensor networks, Journal of Ambient Intelligence and Humanized Computing 12(1) (2021), 547–566.

25.

Vinayakumar

, Alazab

, Soman

K.P.

, Poornachandran

, Al-Nemrat

and Venkatraman

, Deep learning approach for intelligent intrusion detection system, IEEE Access 7 (2019), 41525–41550.

26.

Hu , Pin , Chen

, Huang

, Zhang

and Liu

, Improved alpha-guided Grey wolf optimizer, IEEE Access 7 (2018), 5421–5437.

27.

Revanesh

, Sridhar

A trusted distributed routing scheme for wireless sensor networks using blockchain and meta-heuristics-based deep learning technique, Transactions on Emerging Telecommunications Technologies, (2021), e4259.

28.

Mubarakali

, “An Efficient Authentication Scheme Using Blockchain Technology for Wireless Sensor Networks, Wireless Personal Communications, (2021), 1–15.

29.

Lazrag

, Chehri

, Saadane

, Rahmani

M.D.

Efficient and secure routing protocol based on Blockchain approach for wireless sensor networks, Concurrency and Computation: Practice and Experience, (2020), e6144.

30.

Prithi

and Sumathi

, LD2FA-PSO: A novel learning dynamic deterministic finite automata with PSO algorithm for secured energy efficient routing in wireless sensor network, Ad Hoc Networks 97 (2020), 102024.

31.

Pavani

and Rao

P.T.

, Adaptive PSO with optimised firefly algorithms for secure cluster-based routing in wireless sensor networks, IET Wireless Sensor Systems 9(5) (2019), 274–283.

32.

Sun

, Wei

, Zhang

and Qu

, Secure routing protocol based on multi-objective ant-colony-optimization for wireless sensor networks, Applied Soft Computing 77 (2019), 366–375.

33.

Prithi

and Sumathi

, Automata Based Hybrid PSO–GWO Algorithm for Secured Energy Efficient Optimal Routing in Wireless Sensor Network, {Wireless Personal Communications 117(2) (2021), 545–559.

A Deep-DrpXML and IAG-GWO based CHST fostered blockchain technology for secured dynamic optimal routing for wireless sensor networks

Abstract

Keywords

1 Introduction

2 Literature survey

3 Proposed methodology

3.1.1 Crypto Hash Signature Token fostered Block chain wireless sensor networks

4 Results and discussions

Table 2 Device specification Restriction Server Block Terminal Block Routing Block Name chain sensor chain sensor chain sensor Node Node Node CPU 3GHz 1.5GHz 600 MHz RAM 32GB 2GB 64MB Storage 1 TB 32 GB 512 MB Network 1100 Mb 110 Mb 110 Mb Operating system Windows 7 Windows 7 Windows 7

5 Conclusion

References

Table 2
Device specification

Restriction Server Block Terminal Block Routing Block

Name chain sensor chain sensor chain sensor

Node Node Node

CPU 3GHz 1.5GHz 600 MHz

RAM 32GB 2GB 64MB

Storage 1 TB 32 GB 512 MB

Network 1100 Mb 110 Mb 110 Mb

Operating system Windows 7 Windows 7 Windows 7