Secure Low-Energy Routing Protocol Based on Dynamic Trust Awareness and Load Balancing in Wireless Sensor Networks

Xiaogang, Yuan

doi:https://doi.org/10.1155/2023/6772435

Security and Communication Networks

On this page

Abstract Introduction Conclusion Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Research Article Erratum

!

An Erratum for this article has been published. To view the article details, please click the ‘Erratum’ tab above.

Research Article | Open Access

Volume 2023 | Article ID 6772435 | https://doi.org/10.1155/2023/6772435

Secure Low-Energy Routing Protocol Based on Dynamic Trust Awareness and Load Balancing in Wireless Sensor Networks

Yuan Xiaogang¹

Academic Editor: Vincenzo Conti

Received23 Feb 2023

Revised30 Jul 2023

Accepted13 Aug 2023

Published09 Sept 2023

Abstract

Since wireless sensor networks (WSNs) have the requirements of high security and energy conservation, a distributed secure low-energy routing protocol (SLERP) based on dynamic trust awareness and load balancing is proposed. In order to reduce the adverse influence of malicious nodes in the network, the Chebyshev neural network is used to predict the dynamic trust degree of network nodes to accelerate the speed and accuracy of malicious node detection. Based on the comprehensive consideration of the average dynamic trust degree of cluster-head nodes, the load of the cluster-head node, network energy consumption, network lifetime, and accurate route evaluation model are established by the analytic hierarchy process (AHP). The search methods of secure low-energy routing and chromosome crossing and the mutation method are designed based on the genetic algorithm (GA), so as to quickly establish the optimal cluster-head node-set and the optimal routing path of each node. Simulation results show that SLERP can significantly improve the detection speed and detection success rate of malicious nodes, reduce the network energy consumption and load, and effectively extend the network lifetime.

1. Introduction

Wireless sensor networks (WSNs) are composed of a large number of sensor nodes deployed in the detection area. Their main role is to collect and analyze data information in the monitoring area and transmit it to the base station. Wireless sensor nodes are vulnerable to various factors such as power, bandwidth, and energy, especially when WSNs are randomly deployed in a complex environment, and they are vulnerable to routing attacks from malicious nodes [1]. Therefore, the research of WSNs routing protocol needs to consider reliability, node energy consumption, and load balancing.

In order to improve the security of the network and prolong the lifetime of the network, a large number of research works have been carried out from various angles to reduce the impact of malicious nodes in the network and improve the energy efficiency of the network. Studies have shown that trust-based schemes have been proven to have good resistance to internal node attacks, and dividing nodes into different clusters can effectively prolong the lifetime of WSNs [2, 3]. The comprehensive node trust calculation model constructed from the node behavior dimension and energy dimension can simultaneously realize the security and efficiency of wireless sensor networks by quantifying multiple QoS indicators [4]. Based on reducing the calculation amount and traffic volume of trust and taking trust and energy into comprehensive consideration when making routing decisions, the energy optimization security protocol can optimize the load balancing of the network while ensuring data transmission [5]. By taking node trust, residual energy, and data transmission distance as factors for routing selection, the secure clustering routing protocol can reduce the probability of malicious nodes participating in data transmission and can improve the security and reliability of WSNs [6]. The trust-aware secure clustering routing protocol using an enhanced genetic algorithm (SCRGA) can simultaneously select the optimal cluster-head node and search the secure route of the best path, and build a fitness function based on the maximum comprehensive trust of the cluster-head node and the minimum energy consumption of the whole network and load balancing, which can effectively extend its lifetime on the basis of guaranteeing network security [7]. In dynamic underwater WSNs, the trust evaluation and update mechanism based on the C4.5 decision tree algorithm (TEUC) has a good performance in malicious node detection and energy consumption [8]. By updating the threshold of node trust and selecting more secure and higher energy nodes as cluster-head nodes, the energy and trust models based on low-energy adaptive clustering hierarchy (ETM-LEACH) can ensure the security and rationality of cluster structure [9]. Using an intelligent trust classification scheme to judge whether devices in the improvement of operations and maintenance techniques (IoMTs) are malicious devices can effectively solve the problem of trust uncertainty and improve the detection accuracy of malicious devices [10]. The reliable transmission routing protocol (RTRPT) adopts the gradient clustering model and calculates the comprehensive trust of neighbor nodes with multiple index periods, so the protocol can prolong the network lifetime under the condition of reducing node packet loss rate and packet transmission delay [11]. Trust management based on artificial intelligence is a promising example of providing trusted and reliable services, and the research result shows that the trust management method based on generative adversarial learning has an excellent performance in ensuring network security and service quality in 6G wireless networks [12].

The secure routing protocols proposed in the abovementioned literature have many advantages but also face the following shortcomings:(1)The network structure, node trust, load, energy consumption, and lifetime involved in secure low-energy routing in WSNs are not fully considered, which leads to some defects in the overall performance of the routing.(2)The multiobjective function of secure routing in WSNs adopts an artificial weight coefficient, which makes the evaluation model of routing performance to be less accurate.(3)The comprehensive trust degree of nodes is obtained by adding the current trust degree and the historical trust degree of nodes, which makes the timeliness and dynamic of the trust degree evaluation of routing nodes lack, and leads to the delay of each step decision on routing security.

Considering the security and energy saving of WSNs routing, this paper proposes a distributed secure low-energy routing protocol based on dynamic trust awareness and load balancing, which can effectively guarantee the network data transmission and prolong the network lifetime. The main contributions of this paper are as follows:(1)A Chebyshev neural network is used to predict the comprehensive trust degree of nodes, thus obtaining the dynamic trust value of nodes, so as to improve the accuracy and real-time evaluation of node trust degree, and improve the detection speed and detection success rate of malicious nodes.(2)Taking the highest average dynamic trust of cluster-head nodes, the minimum load of cluster-head nodes, the minimum network energy consumption, and the longest network lifetime as evaluation indexes, the analytic hierarchy process (AHP) was used to construct an accurate multiobjective optimization evaluation model for routing. This multiobjective weight coefficient determination method is more objective and accurate, which is suitable for the actual situation of WSNs and can improve the network performance better.(3)The genetic algorithm (GA) is adopted to search for the optimal path, and the initial population is established based on the criteria of high average cluster-head load and short packet transmission distance. In addition, the fitness function with constraints, crossover operator of random crossover of normal nodes, mutation operator of cluster-head node reselection, and elite retention strategy are designed in detail, so as to accelerate the search speed of the optimal path while maintaining the diversity of the population. Thus, the overall optimal route for security and energy saving can be found quickly.

2. Secure Routing Model

The secure low-energy routing in WSNs is sensitive to the parameters such as trust, load, energy consumption, and survival time of nodes. In order to solve the optimal routing of multiple parameters in WSNs at the same time, it is crucial to construct a multiobjective optimization model and an evaluation method for WSN routing [13, 14].

2.1. Network Model

It is assumed that WSNs consist of a base station and K sensor nodes, and it is distributed in a square monitoring area with a side length L. is the cluster-head nodes set and is the normal node-set. So, K is equal to . is the distance between the node and node . The network has the following characteristics [15, 16]:(1)The locations of sensor nodes and base stations will not change after they are determined, and all sensor nodes have an identification (ID)(2)The relative position of nodes in the network is directly estimated by the location method(3)The initial energy of all sensor nodes is the same, and the energy of the base station h₀ is unlimited(4)Sensor nodes have an energy control mechanism, which can adjust the amount of transmitted energy according to the transmission distance(5)The base station is responsible for the calculation and storage of parameters related to routing, such as the average trust degree of cluster-head nodes, minimum running time of cluster-head nodes, maximum load of cluster-head nodes, and network energy consumption

2.1.1. Dynamic Optimal Number of Cluster-Head Nodes

The network adopts the idea of dynamic clustering, and the optimal number of cluster-head nodes is calculated by referring to the determination method of the optimal number of cluster-head nodes in the classical WSNs routing protocol LEACH [17]. It is assumed that K_live active nodes are randomly distributed in the square monitoring area with a side length L, and the optimal cluster-head node percentage is [18].where is the magnification of the free space model, is the magnification of the multipath transmission model, and is the distance from the node to the base station.

The number of optimal cluster-head nodes M_opt is the number of active nodes K_live multiplied by the percentage of the optimal cluster-head nodes p_opt and rounded to an integer and is represented as follows:

2.1.2. Energy Consumption Model

The energy E_Tx (u, d) consumed by a node sending u bits of data to a location with distance d is calculated as follows [19]:where E_elsc is the energy consumption of sending or receiving 1 bit of data.

The energy consumption E_Rx (u) of a node to receive u bits of data is

The energy E_DA (u) consumed by a node to fuse u bits of data iswhere E_fu is the energy consumption of 1-bit data fusion.

2.2. Dynamic Trust Degree of Nodes

WSNs secure routing based on a trust mechanism has obvious advantages in solving internal attacks, identifying malicious nodes, and improving system security, reliability, and fairness [20, 21]. In view of the complexity and variability of the network environment and attack behavior, in order to reduce the trust of malicious nodes quickly, this paper uses the Chebyshev neural network to perform vertical analysis and dynamic prediction of node trust, so as to improve the accuracy of trust evaluation and the detection speed of malicious nodes.

2.2.1. Node Trust Model

The direct trust can be obtained through the statistical expectation of beta distribution. At the same time, considering the influence of external factors such as node communication behavior, the original model is improved by introducing an abnormal attenuation μ factor [5]. The formula for calculating the direct trust degree of node j to node i is as follows:where α_ji represents the number of packets successfully forwarded by node j from node i and β_ji represents the number of packets unsuccessfully forwarded by node j from node i. μ represents the probability that node abnormal behavior is malicious attack behavior. Num_instausion represents the number of noncooperations caused by the node attack behavior, and Num_detection represents the total number of noncooperation of nodes.

By using the abnormal attenuation factor μ to attenuate the number of noncooperation detected by node j on node i, the influence of external factors on trust degree can be reduced and the accuracy of trust evaluation can be improved.

Assuming that node i has J neighbor nodes, the comprehensive trust degree of node i is calculated as follows:

2.2.2. Dynamic Trust Model Based on the Chebyshev Neural Network Prediction

The trust degree of nodes in WSNs is predictable [22, 23]. The Chebyshev neural network is a neural network based on Chebyshev orthogonal polynomials, which has a superior approximation performance. The orthogonality of the basis function ensures that the neural network can quickly solve the weights at the nodes of the polynomial and can better meet the needs of WSNs node trust prediction. In this paper, the Chebyshev neural network is used to dynamically predict the comprehensive trust degree of sensor network nodes.

The single-input Chebyshev neural network consists of an input layer, a hidden layer, and an output layer, as shown in Figure 1. The input layer, hidden layer, and the output layer of the Chebyshev neural network are , , and and . is the Chebyshev polynomial of the first kind.

The prediction of the comprehensive trust degree of nodes in WSNs takes the comprehensive trust degree of the current moment as the input of the Chebyshev neural network, and the output of the network is the dynamic predicted value of the comprehensive trust degree of nodes.

The neural network is first preprocessed when the prediction model has been established, and the main work includes the normalization of the sample set and the construction of the training set and the test set. 30% of the sample is randomly selected as the training set, and the remaining 70% of the sample is used as the test set. The system uses the training set to train the neural network, and uses the gradient descent method to modify the weight until the prediction success rate reaches the preset standard [24, 25].

2.2.3. Average Trust Degree of Cluster-Head Nodes

In order to find a routing path with a high credibility, all cluster-head nodes are required to have a high trust. This paper uses the average dynamic trust of all cluster-head nodes to measure. The number of cluster-head nodes in the network is M_opt, and the average dynamic trust of cluster-head nodes is as follows:

2.3. Load of Cluster-Head Nodes

The cluster-header node transmits data both from its cluster member node and possibly from the cluster-head node at the next level. Assuming that the transmitted data comes from a total of R nodes, then the load of the cluster-head node h_m is calculated as follows:where Load is the data traffic generated by a single sensor node.

The maximum load among the M_opt cluster-head nodes is denoted by Load_CH-max which is calculated as follows:

2.4. Network Energy Consumption

In sensor networks, the energy consumption per second of each cluster-head node and the energy consumption per second of normal nodes are respectively, expressed as follows:where is the upper cluster-head node (or base station) of the cluster-head node h_m, is the distance between the cluster-head node h_m and its upper cluster-head node or base station, is the cluster-head node of the normal node , and is the distance between the node and its cluster-head node.

The energy consumption of the whole network E_total is obtained by using the following equation:where N_live is the number of surviving normal nodes in the network.

2.5. Minimum Remaining Running Time of All Cluster-Head Nodes

The minimum running time of all cluster-head nodes can directly reflect the lifetime of the whole network, because the death of the cluster-head nodes may cause network coverage holes and network forwarding path discontinuity. Therefore, the running time of the network is directly related to the death time of the first cluster-head node. In this paper, the remaining running time of the cluster-head node h_m is estimated by using the current residual energy and the energy consumption level of the cluster-head node. The calculation formula is as follows:where represents the residual energy of the cluster-head node h_m.

The minimum remaining running time of all cluster-head nodes is expressed as

2.6. Routing Performance Evaluation Model Based on the Analytic Hierarchy Process

The optimization of secure routing in WSNs is a multiobjective optimal problem which needs to maximize the average trust degree of cluster-head nodes and the minimum running time of cluster-head nodes , and minimize the maximum load of cluster-head nodes and network energy consumption E_total.

The evaluation model Eva of routing performance is represented as follows:where E_0max is the maximum energy consumption of a single sensor.

There is a contradictory relationship between objectives in multiobjective optimization problems, which makes the multiobjective optimization problem more complex [26]. In this paper, the AHP is used to accurately estimate the weight of each factor, as shown in Figure 2.

When the AHP is used to determine the four weights, the relative weights of indicators at each layer should be compared in pairs according to the nine-level scaling method shown in Table 1, and then the comparison table shown in Table 2 and the basic judgment matrix A shown in equation (17) can be obtained [27, 28].

The largest characteristic root of the matrix is calculated as .

The consistency index CI is calculated as follows:

In order to measure the consistency of the judgment matrix, the average random consistency index RI is introduced. The RI values of the 1–8 order matrix are shown in Table 3.

When the order of the judgment matrix is less than three, the judgment matrix is always consistent. When the order of the judgment matrix is greater than or equal to 3, the consistency ratio CR of the judgment matrix is expressed by the following equation:

The consistency index of the judgment matrix is calculated as . It can be seen from Table 3 that the consistency index RI of the fourth-order matrix is 0.89, and the consistency ratio indicates that the judgment matrix A meets the consistency test.

After calculating the weight vector according to , the normalized weight vector ω can be obtained as .

The routing performance evaluation model Eva defined in this paper is represented as follows:

3. Secure and Low-Energy Consumption Routing Method Based on GA

In this paper, GA is used to find the optimal routing path in WSNs. According to the requirements of routing security, energy consumption, and load, the appropriate coding scheme, selection, crossover, and mutation operators are designed in detail.

3.1. Chromosome Coding and Selection of Initial Population

Using real number coding, the network consists of K nodes, and each node is assigned a unique positive integer between 1 and K. The network routing adopts the structure of multihop clusters and single-hop within clusters.

3.1.1. Selection of the Cluster-Head Node and Search of the Route of the Cluster-Head Node

The routing of the cluster-head node is selected as follows:(1)In order to improve the security of routing, we prolong the survival time of nodes and accelerate the convergence of the algorithm, and only when the residual energy and comprehensive trust of the node are greater than the average residual energy and the average comprehensive trust of the network, the node will participate in the election of the cluster-head node. According to equation (2), M_optcluster-head nodes are randomly selected. Therefore, the length of the chromosome is M_opt.(2)The cluster-head nodes are sorted according to the distance from the base station, and the cluster-head nodes are marked as CH₁, CH₂, , and , and the distance from the cluster-head node to the base station is . The results are shown in the section on cluster-head node ID in Table 4.

The routing path from the cluster-head node to the base station is determined in the order of CH₁, CH₂, , and . If the base station h₀ is within the direct communication distance of the cluster-head node, then the cluster-head node communicates directly with the base station h₀. If the cluster-head node CH_i is not within the direct communication distance of the base station (i.e., ), then the cluster-head node CH_i selects its upper forwarding cluster-head node according to equation (21). To be specific, CH_i selects the cluster-head node within the range of direct communication distance and the closest routing distance to the base station in CH₁, CH₂, , CH_i-1 is its upper-layer forwarding cluster-head node.where is the signal transmission distance of data from the cluster-head node h_i to the base station h₀ after forwarding through other cluster-head nodes and is the signal transmission distance of data from any cluster-head node to the base station h₀.

The routing path from all cluster-head nodes to the base station can be obtained as shown in Figure 3.

3.1.2. Cluster Selection of Normal Nodes

In order to speed up the convergence of GA and obtain an effective solution, each normal node in WSNs selects the cluster-head node which is within its communication range and has the nearest routing to the base station as its own cluster-head node according to equation (22), as shown in Table 4.where is the distance between the node and the node h_i, is the signal transmission distance of data from the node to the base station h₀ after forwarding through the cluster-head node h_i, and is the signal transmission distance from the node to the base station h₀ after data are forwarded by any cluster-head node.

3.2. Fitness Function

The secure low-energy routing protocol (SLERP) constructs the fitness function f based on the route evaluation model Eva. In order to avoid the emergence of nodes that cannot enter the route, and considering the difference of importance between cluster-head nodes and normal nodes in cluster-based WSNs, the fitness function needs to be multiplied by a penalty coefficient if a node cannot enter the route.where s is the number of cluster-head nodes that cannot be connected to the network and t is the number of normal nodes that cannot be connected to the network.

The chromosome with the highest fitness value at the end of the iteration is the optimal path, and the optimal route with safe and reliable low-energy consumption can be constructed by this function.

3.3. Genetic Operations

Genetic operations mainly include selection operation, crossover operation, mutation operation, and elite retention operation.

3.3.1. Selection Operator

Using the traditional roulette wheel method, the fitness value of each individual is calculated, and then the proportion of this fitness value in the total fitness value of the population is calculated, which indicates the probability of the individual being selected in the selection process.

3.3.2. Crossover Operation

In order to enhance the diversity of the population, random crossover is performed within a chromosome. All cluster-head nodes of each chromosome are unchanged. Each normal node randomly selects its own new cluster-head node from the cluster-head nodes within its communication distance according to equation (24), thus forming a new gene.

3.3.3. Mutation Operators

When the chromosome is mutated, it randomly selects its own cluster-head nodes again, and forms a cluster-head node path according to equation (21) and the principle of the nearest distance. Normal nodes select their own cluster-head nodes according to equation (22), and each cluster-head node forms its own gene. A new chromosome is formed by reselection of cluster-head nodes and normal nodes.

3.3.4. Keep Elite Strategy

In order to ensure that the result when the algorithm terminates is the best solution that has ever been reached in the whole search, a memory device is introduced to store the best solution in the whole iteration process. At the end of each new iteration, the best solution of the current population is compared with the solution of the memory device, that is, if the best solution of the current population is better than the solution in the memory device, then the solution of the memory device is replaced by the best solution of the current population, otherwise the solution of the memory device is kept unchanged. After the whole optimization process, the solution in the memory device is compared with the best solution in the optimization results, so as to select the best solution.

3.4. Termination Condition of the Algorithm

The iterative termination conditions of GA are as follows:(1)The number of iterations reaches the maximum genetic evolution algebra(2)The optimal value searched by the algorithm remains unchanged for several successive generations

3.5. Routing Maintenance

When the trust value of any cluster-head node is lower than the average trust value of the surviving nodes in the network, or the residual energy of any cluster-head node is lower than 70% of the average residual energy of the surviving nodes in the network, then steps 3.1 to 3.3 are executed to rebuild the network routing.

When a normal node is judged as a malicious node or runs out of energy, it is isolated from the network.

According to the abovementioned analysis, the route selection process based on GA is obtained, as shown in Figure 4.

4. Simulation Results and Performance Evaluation

In this section, the performance of SLERP is analyzed by using MATLAB and compared with SCRGA proposed in [7] and ETM-LEACH proposed in [9]. 100 nodes are randomly deployed in a 100 m × 100 m area, the base station is set at the middle position of the edge, and 1 to 10 malicious nodes are randomly set in the network to launch selective attacks at random. The initial trust value of a node is set to 0.5 and the threshold of the trust value is set to 0.35, so any node will be isolated from the network when the trust value of the node is lower than the threshold [6, 17]. The population number of GA is 200, the maximum evolution generation is 500, the crossover rate is 0.8, and the mutation rate is 0.1. The average value of 10 simulation results is taken. The simulation parameters are shown in Table 5 [6–12].

4.1. Success Rate of Malicious Node Detection

The success rate of malicious node detection refers to the probability that the malicious node is identified, and the network security can be evaluated according to the detection rate. It can be seen from Figure 5 that with the increase of the proportion of malicious nodes in the network, the success rate of malicious node detection of the three protocols decreases, and the final recognition rate is between 75% and 90%. In general, SLERP is slightly better than SCRGA, and it is significantly better than ETM-LEACH because the SLERP uses the Chebyshev neural network to predict the dynamic trust of nodes, the node evaluation has a better real-time performance and accuracy, and the final malicious node recognition rate is slightly higher than the other two routes. It can be seen that SLERP can isolate malicious nodes from the network faster and reduce the impact of malicious nodes on the security performance of other nodes in the network.

4.2. Packet Loss Rate

Packet loss is caused by collision, congestion, or internal attack while the packet is in transit, and the loss ratio is called the packet loss rate. The comparison of packet loss rate can reflect the situation of network security and load balance. We attack the network, when the number of malicious nodes is from 1 to 10, and observe the percentage of packet loss rate of the three routing protocols. As can be seen from Figure 6, when there are no malicious nodes, the packet loss rate of the three routing protocols is within 1% because of the delay and communication quality issues. With the increase in the number of malicious nodes, the packet loss rate of the three protocols increases, but SLERP has the best performance. When the number of malicious nodes is 10, the packet loss rate of SLERP is 3% and 6% lower than that of SCRGA and ETM-LEACH, respectively. Since the SLERP makes a dynamic prediction of node trust and gives priority to selecting nodes with a high trust as cluster-head nodes in routing selection, it reduces the trust of malicious nodes more quickly. Malicious nodes are isolated from all nodes and isolated from the network to avoid communication with them, which can effectively reduce the packet loss rate in the process of data transmission and ensure the safety of the network.

4.3. Transmission Delay

The end-to-end transmission delay of the network is the time difference between the source node sending a packet and the base station receiving the packet. It includes the transmission delay and the data processing delay. The smaller the delay, the more unobstructed the network. The simulation result in Figure 7 compares the changing trend of the average end-to-end delay of the three routing algorithms when the number of malicious nodes increases. As the number of malicious nodes increases in WSNs, the delay of the three routing algorithms also increases due to the existence of potential malicious nodes in the network. When the number of malicious nodes is 10, the transmission delay of SLERP is about 4 ms and 6.5 ms, which is lower than that of SCRGA and ETM-LEACH. This is due to the frequent packet loss of the network, and the upper network protocol needs to wait for the establishment of links between the communication nodes and the retransmission of data packets, so the delay increases. The long-term dynamic prediction method of the trust model of SLERP can better evaluate the security of nodes, and the shortest path is given priority in the route establishment process, which can quickly obtain the optimal path. The higher the trust degree of the node, the lower the packet loss rate, and the higher the probability of being selected as the cluster-head node, the lower the delay. A node with a higher trust is more likely to be selected as a cluster-head node, so the routing has a lower delay.

4.4. Network Lifetime

The survival time of the network is the change from the beginning of the sensor nodes in the network to when all nodes die due to energy depletion, and it is an important indicator to evaluate the effect of the routing algorithm of WSNs. The long survival time of nodes indicates that the routing algorithm has a good performance, and the network topology is reasonable. The death time of the first node and the death time of half of the nodes are both selected as the evaluation parameters for comparison in this paper, and the comparison results are shown in Figures 8 and 9.

The simulation results show that the network lifetime of SLERP is longer than that of the other two protocols even when there is no malicious node. In terms of the occurrence of the first dead node, the performance of SLERP is slightly lower than that of SCRGA when the number of malicious nodes is small, but when the number of malicious nodes increases to more than 7, the performance of SLERP is comparable to SCRGA. It can also be seen that the performance of the SLERP protocol is significantly better than that of the ETM-LEACH protocol in terms of the occurrence of the first dead node. The performance of SLERP is better than SCRGA and ETM-LEACH in terms of the time when half of the nodes die. It can be seen from the results in Figures 8 and 9 that SLERP significantly improves the network lifetime compared with SCRGA and ETM-LEACH. In general, the SLERP protocol considers the trust, path length, and residual energy at the same time in the routing discovery stage, which not only guarantees route reliability but also effectively improves the load balancing of the entire network, thus greatly improving the survival time of the network.

5. Conclusion

Aiming at the influence of malicious nodes on the security of WSNs and the limitation of node energy on the network life cycle, a routing performance evaluation model is established, which considers the dynamic trust of nodes, the load of cluster-head nodes, network energy consumption, and node lifetime. Based on AHP and GA, a secure and low-energy consumption routing protocol for WSNs is proposed. The simulation results show that SLERP has a better performance in terms of malicious node detection success rate, packet loss rate, data transmission delay, and network lifetime, and has a good defense effect against malicious node attacks. SLERP needs to calculate the prediction value of node trust, crossover, and mutation operations of GA, so the algorithm is a little computationally heavy. The abovementioned calculation work is mainly completed in the base station, so it has no effect on the energy loss of the wireless sensor network. However, it is necessary to simplify the calculation process of the algorithm in the future research, so as to speed up the search and establishment of the optimal route.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The author declares that there are no conflicts of interest.

Acknowledgments

This research was supported by the Young Doctor Foundation of the Education Department of Gansu Province under Grant no. 2022QB-132, the Soft Science Special Project of Gansu Basic Research Plan under Grant no. 22JR11RA106, and the Key Research Project of Gansu University of Political Science and Law under Grant no. GZF2022XZD08.

References

Y. Han, H. Hu, and Y. Guo, “Energy-aware and trust-based secure routing protocol for wireless sensor networks using adaptive genetic algorithm,” IEEE Access, vol. 10, pp. 11538–11550, 2022.
View at: Publisher Site | Google Scholar
H. Hu, M. Yao, L. Wang, and Y. Han, “Energy aware clustering routing protocol based on improved AP and genetic algorithm,” Journal of Jilin University (Science Edition), vol. 59, no. 6, pp. 1525–1531, 2021.
View at: Google Scholar
Z. Sun and P. Wu, “Research on IWSN secure routing based on trust evaluation model,” Chinese Journal of Sensors and Actuators, vol. 32, no. 6, pp. 858–865, 2019.
View at: Google Scholar
D. Qin, S. Jia, S. Yang, J. Ma, Y. Zhang, and Q. Ding, “Research on trust sensing based secure routing mechanism for wireless sensor network,” Journal on Communications, vol. 38, no. 10, pp. 60–70, 2017.
View at: Google Scholar
Y. Tao, X. Xu, P. Li, and L. Pan, “Energy-optimized secure routing algorithm in WSN,” Computer Engineering and Design, vol. 39, no. 12, pp. 3638–3642, 2018.
View at: Google Scholar
B. Pan, Y. Wu, Q. Yang, and L. Ge, “The impact of major dietary patterns on glycemic control, cardiovascular risk factors, and weight loss in patients with type 2 diabetes: a network meta-analysis,” Journal of Evidence-based Medicine, vol. 12, no. 1, pp. 29–39, 2019.
View at: Publisher Site | Google Scholar
C. Wang, X. Wang, H. Hu, H. Zhao, and Y. Han, “Secure clustering routing protocol based on improved GA and trust-aware for wireless sensor networks,” Journal of Jilin University (Science Edition), vol. 59, no. 9, pp. 1237–1244, 2021.
View at: Google Scholar
J. Jiang, X. Zhu, and G. Han, “A dynamic trust evaluation and update mechanism based on C4.5 decision tree in underwater wireless sensor networks,” IEEE Transactions on Vehicular Technology, vol. 69, no. 8, pp. 9031–9040, 2020.
View at: Publisher Site | Google Scholar
J. Wang, J. Liu, and F. Zhang, “Heterogeneous clustering routing protocol based on energy and trust,” Computer Engineering and Design, vol. 42, no. 4, pp. 927–933, 2021.
View at: Google Scholar
L. Yang, Y. Li, S. Yang, Y. Lu, T. Guo, and K. Yu, “Generative adversarial learning for intelligent trust management in 6G wireless networks,” IEEE Network, vol. 36, no. 4, pp. 134–140, 2022.
View at: Publisher Site | Google Scholar
Q. Zhang and X. Ren, “Trust-based reliable transmission routing protocol for wireless sensor network,” Application Research of Computers, vol. 39, no. 5, pp. 1514–1518, 2022.
View at: Google Scholar
L. Yang, K. Yu, S. Yang, C. Chakraborty, Y. Lu, and T. Guo, “An intelligent trust cloud management method for secure clustering in 5G enabled internet of medical things,” IEEE Transactions on Industrial Informatics, vol. 18, no. 12, pp. 8864–8875, 2022.
View at: Publisher Site | Google Scholar
F. Su and K. Du, “Trust based energy efficient opportunistic routing algorithm in wireless sensor networks,” Computer Science, vol. 47, no. 2, pp. 300–305, 2020.
View at: Google Scholar
G. Zhang, Y. Yang, D. Zhang, and J. Li, “Secure routing mechanism based on trust against packet dropping attack in internet of things,” Computer Science, vol. 46, no. 6, pp. 153–161, 2019.
View at: Google Scholar
H. Hu, Y. Han, and M. Yao, “Trust based secure and energy efficient routing protocol for wireless sensor networks,” IEEE Access, vol. 10, pp. 10585–10596, 2022.
View at: Publisher Site | Google Scholar
A. Beheshtiasl and A. Ghaffari, “Secure and trust-aware routing scheme in wireless sensor networks,” Wireless Personal Communications, vol. 107, no. 4, pp. 1799–1814, 2019.
View at: Publisher Site | Google Scholar
Y. Han, H. Hu, and M. Yao, “A trust-aware secure routing protocol for wireless sensor networks,” Computer Engineering, vol. 47, no. 9, pp. 145–152, 2021.
View at: Google Scholar
H. Qi, L. Lyu, and H. Chen, “Optimized LEACH algorithm based on power control and clustering quantity,” Transducer and Microsystem Technologies, vol. 37, no. 10, pp. 137–139+146, 2018.
View at: Google Scholar
C. Wang, X. Liu, and H. Hu, “Energy-efficient and load-balanced clustering routing protocol for wireless sensor networks using a chaotic genetic algorithm,” IEEE Access, vol. 8, pp. 158082–158096, 2020.
View at: Publisher Site | Google Scholar
D. Chen, H. Qiu, K. Zhu, Q. Wang, and J. Zhu, “An inter-domain routing reputation model based on autonomous domain collaboration,” Sctentta Stntca Informationis, vol. 51, no. 9, pp. 1540–1558, 2021.
View at: Google Scholar
X. Wang, “Low-energy secure routing protocol for WSNs based on multiobjective ant colony optimization algorithm,” Journal of Sensors, vol. 2021, Article ID 7633054, 9 pages, 2021.
View at: Publisher Site | Google Scholar
M. Zeng, H. Jiang, X. Wang, and W. Liu, “Reputation evaluating model and security routing protocol of wireless sensor networks based on grey Markov model,” Application Research of Computers, vol. 30, no. 12, pp. 3758–3761, 2013.
View at: Google Scholar
Y. Liu, N. Liu, X. Li, and Q. Ji, “Load balancing routing protocol based on traffic prediction for wireless mesh networks,” Computer Science, vol. 44, no. 1, pp. 109–112, 2017.
View at: Google Scholar
X. Yuan and Y. Chen, “Secure routing protocol based on dynamic reputation and load balancing in wireless mesh networks,” Journal of Cloud Computing, vol. 11, pp. 77–79, 2022.
View at: Publisher Site | Google Scholar
J. Liang, W. Sun, N. Xiao, W. Chen, and Z. Guo, “Intelligent routing strategy for software-defined satellite network based on Chebyshev neural network,” Journal of National University of Defense Technology, vol. 42, no. 5, pp. 23–30, 2020.
View at: Google Scholar
F. Wang and Z. Xu, “Research on threat level assessment method for infrared multi-target based on AHP,” Ship Electronic Engineering, vol. 41, no. 61, pp. 150–153, 2021.
View at: Google Scholar
M. Bi, Y. Ju, and R. Hou, “Environmentally-adaptive secure routing protocol for satellite networks,” Journal of Xidian University, vol. 47, no. 1, pp. 66–72, 2019.
View at: Google Scholar
H. Yuan and Y. Cao, “An improved algorithm of RPL based on fuzzy analytic hierarchy process,” Chinese Journal of Sensors and Actuators, vol. 31, no. 12, pp. 1900–1905, 2018.
View at: Google Scholar

Copyright

Copyright © 2023 Yuan Xiaogang. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies