Clustering in Wireless Sensor Networks Using Adaptive Neuro-Fuzzy Inference Logic

Authored by: Seema Gaba , Radhika Gupta , Sahil Verma , Kavita , Imran Taj

Information Security Handbook

Print publication date:  February  2022
Online publication date:  February  2022

Print ISBN: 9780367365721
eBook ISBN: 9780367808228
Adobe ISBN:

10.1201/9780367808228-3

 

Abstract

Wireless sensor networks are a powerful category of mobile ad-hoc networks used to provide easy and efficient communication using technology. Wireless sensor networks are widely used for the results they provide, leading to a reduction in human work. Sensors are deployed in groups in required areas where nodes collect data from the surroundings and send it to the sink node through multiple-node communication. This process involves lots of energy dissipation at the individual node level, leading to early failure of the network. To solve this problem, clustering is used in hierarchical routing protocols. The clustering process also lacks efficiency since cluster heads are selected randomly. The appropriate selection of cluster heads may prove to be an effective and logical way to regulate energy consumption and increase network life. This paper proposes an efficient neuro-fuzzy logic-based technique to improve energy consumption and network performance. The wise selection of a cluster head will aid in data-transmission efficiency, improving functioning to ensure network life in emergencies. Adaptive neuro-fuzzy logic helps in training the parameters to meet the requirements of becoming cluster heads. The candidate cluster head parameters are tested against the training data, and the appropriate one is selected as head. The proposed technique is tested for different network cases and has shown good results in the case of the packet delivery ratio.

 Add to shortlist  Cite

Clustering in Wireless Sensor Networks Using Adaptive Neuro-Fuzzy Inference Logic

3.1  Introduction

Numerous applications of self-organized networks that operate without any centralized support have tempted many researchers to study this area. Wireless sensor networks are the most well-known types of self-organized networks being used in many applications. These networks are called ad-hoc networks. The main provision of creating these networks is to readily provide services [1] to end users. The absence of infrastructure or central coordination works to the advantage of mobile ad-hoc networks, or MANETs, due to reduced cost, complexity, and, most importantly, time required for network deployment. Apart from these reasons, MANETs are also suitable for emergency situations like natural or human-induced disasters, military conflicts, emergency medical situations, and business solutions, etc. MANET introduces its branches based on the network and application requirements. The main reason these networks succeed is on-the-go setup. Apart from this capability, a sensor network's success depends on the availability of information, type of information, quality of information, duration of applicability [2], and a couple of other details. It is very important to send the collected information to the base station to fulfil the network's deployment purpose. These type of networks have their use in many fields, including the medical and military fields. So, it is important that data transfer is robust and fault tolerant due to constrained resources, limited bandwidth, and a few other limitations; it becomes important that for efficient data transfer, the network should be fault tolerant and have a longer life. Network lifetime ultimately relies upon the energy utilization of individual nodes in the network. In one of the studies, authors have discussed challenges that will likely present when industrialization reaches its peak. Security and availability are mentioned among those challenges, which comprise future research of WSNs. Because the sensors depend on battery power, that power must be used [3] wisely to prevent unnecessary use. In the previous study, clustering is given as a reliable solution to conserve the node's energy, hence improving network lifetime. The clustering concept reduces [4] the amount of information received at base station from each and every node and provides managed load balancing by organizing the network in a connected hierarchy. Clustering may follow one hop [5] or multi-hop communication. Clustered networks often suffer from the hot spot problem, where cluster heads near the base station die much earlier due to the load of data forwarding. Such deaths of data-forwarding nodes result in network coverage loop holes. However, clusters can be made energy aware in different environments by different methods. The point of concern is the selection of cluster heads. Low energy adaptive clustering hierarchy (LEACH) is one of the prominent routing protocols for sensor networks; it uses the concept of clustering. In this protocol, [6] heads are chosen on the basis of probability value. The chances are that a node with low energy [7] or at a far distance from the sink node may get selected as head. The heads are required to be more capable to collect information [8] from their members, filter it, and send it to sink. Therefore, a lot of new work has been applied to overcome the shortcomings of LEACH. One of the studies combines LEACH, mobile sink, and other points to utilize LEACH and improve CH selection. This paper proposes a fuzzy logic-based technique to provide more refined selection of cluster heads. The fuzzy-inference system is a decision-making system that draws outputs from fuzzy inputs, i.e., uncertain parameters. The logic works on IF-THEN rules and deals with reasoning on the high level using information acquired from users. Fuzzy logic provides an adjustable support to address the aspects of dynamic networks [9] with major uncertainties, which leads to failures and overloads. Adaptive neuro fuzzy is a type of artificial neural network in which neuro-adaptive learning techniques provide a fuzzy modeling method to learn about particular data sets. In basic terms, neuro fuzzy is the tuning of one of the fuzzy inference systems, named sugeno, by using training data. It combines the learning power of an artificial neural network and the explicit knowledge representation of a fuzzy inference system. Key features of this system include use of input-output patterns to adjust the inference rule system. The information is collected and then the inference system is trained to recognize the patterns and behave accordingly. To summarize, we need a training data set for the appliance of neuro-fuzzy logic. In the field of wireless sensor networks [10], adaptive neuro-fuzzy logic has been used for various purposes. The tuned sugeno fuzzy logic can be used in processes like data aggregation, sensing, and remote environmental monitoring in sensor network. For example, [11] proposed an ANFIS based algorithm to monitor temperature gathering in a sensor network. The main advantage is that using ANFIS instead of Mamdani fuzzy type can give better results in terms of accuracy. ANFIS technique is also used to optimize network parameters like calculation of inner packet gap, measurement of latency period, and others. A few studies say that ANFIS gives good outcomes when input parameters do not exceed five. More inputs can cause computational complexes, hence causing delay in calculations. Considering this outcome, in this paper, fewer than five parameters are taken.

The successful outcomes of ANFIS can be based on the robustness of results it provides. ANFIS has highly generalized capabilities of machine-learning techniques or neural networks. ANFIS can take crisp input values, represented in the form of membership functions and fuzzy rules, and also generate crisp output of fuzzy rules for reasoning purposes.

The remaining paper is organized as a review of the existing work done for wireless sensor networks using fuzzy logic, followed by the proposed technique and simulation results of different networks. The last section of the paper covers the conclusion and future work in this context.

3.2  Related Work

Dhananjay Bisen et al. (2018) worked on improving the performance of the AODV routing protocol by reducing the broadcast of unnecessary hello messages. They used a Mamdani fuzzy inference system and adaptive neuro-fuzzy inference system (ANFIS) to calculate the resultant optimal interval of the hello transfer. Energy level and mobility speed of the nodes are used as input for the inference system. The proposed technique is helpful, but it cannot be applied for sensor network because AODV is not a routing protocol for sensors. It can be checked with routing protocols like LEACH and others to be used for sensor networks.

K. P. Vijayakumar et al. (2018) authors have used the fuzzy approach to protect the network against a jamming attack. The fuzzy approach uses two network metrics as input, namely, the packet delivery ratio and received signal strength to detect jamming. To detect jamming, authors say the fuzzy approach gives better results compared to true detection ratio.

Thanga Aruna Muthupandian et al. (2017) presented a survey for techniques for selection of forwarding node (cluster head) in a sensor network. Along with the cluster head paper, the authors also tell about the selection of the next node to be selected for data transmission to send it across nodes to the base station. Authors have presented a comparative study of the fuzzy logic and adaptive neuro-fuzzy logic based on network parameters like the packet delivery ratio, distance, traffic, energy, and others.

Mohammad Shokouhifar, Ali Jalali (2017) has proposed a new energy-optimized routing algorithm for wireless sensor networks called LEACH-FS [12]. To provide uniformity in making an efficient cluster, the algorithm uses a fuzzy c-means method along with bee colony to adjust the fuzzy rules. The fuzzy system selects the appropriate cluster head for the network. The inputs given to fuzzy inference are distance from sink, residual energy, and distance from the cluster centroid. The artificial bee colony algorithm is used to automate the tuning of rules. This automatic tuning eliminates the manual loading of rules.

S. A. Sahaaya et al. (2017) proposed an enhanced zone-stable, fuzzy-logic based cluster head election algorithm (ZSEP-E) [13] for a wireless sensor network. Fuzzy logic inputs are remaining energy, density, and the distance from sink node of the network. A zone partition algorithm forms three zones, with two zones having homogeneous sensor nodes with same capabilities. Nodes are not location aware and not made to be mobile. Also, the initial energy of nodes is categorized as low, intermediate, and advanced, indicating the static nature of selecting heads as nodes; nodes with high energy are the most probable candidates for selection as the cluster head. Fuzzy rule-based optimization is given as a future scope of the work.

B. Baranidharan et al. (2016) have given a distributed load-balancing technique for clustered networks. Authors say load balancing plays an important role so that the network can serve for longer. The fuzzy approach is used for selecting the cluster head. The input parameters for a Mamdani inference system are residual energy, highest number of neighbours, and distance from base station. The cluster size is kept unequal so that the cluster head near the base station can serve for longer, keeping the cluster small. Taking this work forward, the neuro-fuzzy concept is experimented with in this paper to check real-time parameters after every data transmission.

Julie (2016) has proposed an artificial intelligence technique: adaptive neuro fuzzy inference system is used, which forms neuro-fuzzy energy aware clustering scheme (NFEACS) [14]. Using the neural network property of ANFIS, signal strength and energy are trained. Therefore, based on errors found in testing data against training data of these two factors, fuzzy-logic cluster heads are selected. The sink location is centralized here. More refined results are possible if one or more parameters, like node degree, are taken for training set data.

M. Selvi, R. Logambigai et al. (2016) Paper presents a fuzzy temporal logic for clustering. The process differs in that two relay nodes, called a cluster head [15] and super cluster head, are formed. The super cluster head performs routing across clusters, and the cluster head transfers data within clusters. The fuzzy-rule combinations have proven ineffective in decision making. However, the metrics like flow control and congestion control are not considered in this work, which leads to improved quality of service.

Zahra Beiranvand et al. (2013) presents an energy-efficient algorithm, improved LEACH (I-LEACH) [16], which considers location and a number of neighbour nodes, along with energy and distance from sink node, to select appropriate cluster head. However, once deployed, the network is kept static, i.e., the mobility factor of the ad-hoc sensor network is not taken for this work.

Hakan Bagci et al. (2013) has proposed a fuzzy approach to solve the complications of hot spot in network. The algorithm [17] forms clusters of different size, and the clusters near the sink node are smaller in size, reducing intra-cluster work. The mobility factor of sensors is not considered in this work, which may produce more efficient results.

Krasimira Kapitanova et al. (2012) used fuzzy rule-based inference system for robust event detection in sensor networks. However, the authors say that it is difficult to tune the exponentially growing size of rule base. Instead of using crisp values, it is preferred to use fuzzy values for handling uncertainty. The issue of managing a larger rule base can be managed if we have enough data sets that can be trained with existing rules. This can be done applying artificial intelligence with fuzzy logic. This is what paper has tried to implement using adaptive neuro-fuzzy logic.

Safdar AbbasKhan et al. (2012) authors, for detection of faults in the sensor networks, implement another application of the fuzzy inference system. Each node in the network has its own fuzzy model, which is based on a rule set. The rule set gets input as the sensor measurements of neighbouring nodes. Based on this, the node's actual measurement is outputted. Based on the difference in actual measurement and fuzzy output measurement, a node is declared faulty. Participation of faulty node in data transmission can lead to security breach and network shutdown. Therefore, using fuzzy logic, authors have provided a better approach.

Hasung-Pin Chang et al. (2012) presents a hybrid communication protocol named cross-layer energy efficient protocol (CEEP) [18] to minimize information exchange back and forth in the network by the cross-layer optimization concept. The hybrid routing is applied by combining both reactive and proactive schemes, thereby implementing proactive routing within zones and reactive between zones. Authors have only two regions in the network, which can create additional overhead if the network is bigger. In that case, even the intra-zone routing can cause delay to reach the sink, which can be in another zone.

Sudip Misra et al. (2010) has given a simple least-time energy-efficient protocol with one-level data collection (LEO) [19]. The protocol is of a proactive nature, with a modification that every node only contains information about its neighbours. One-level aggregation is done at the node, which is in the closest vicinity of the sink node. Protocol provides a route that takes less time to transfer data from nodes with higher energy and less distance to sink. However, factors like mobility and security are not considered for this work.

Hoda Taheri et al. (2012) paper proposes an energy-aware distributed dynamic clustering protocol for the clustering in sensor networks. Authors have chosen tentative cluster heads based on their remaining energy, and then for a set of nodes, fuzzy logic is applied to check the node fitness for electing as head. The paper presents use of if-then sugeno fuzzy rules to have final selection of cluster head. If considering other network parameters, this work can provide more reliable results.

Toleen Jaradat et al. (2013) have proposed a fuzzy-based cross-layer routing for sensor networks. The fuzzy inference system is inputted three parameters, remaining battery life, link quality, and transmission power for nodes. The combination of if-then rules of these parameters gives the output as a node, which will be next relay node. Every time, fuzzy logic will select the next node, which will communicate the data.

Nimisha Ghosh et al. (2017) an energy-efficient routing is proposed [2032] for wireless homogeneous sensor network, with the use of a mobile collector. The genetic algorithms, which include particle swarm optimization (PSO) and ant colony optimization (ACO), are applied for the results optimization. The main theme follows the selection of cluster head from the clusters in network using fuzzy logic. The mobile collector gathers information from these heads and delivers to sink node. The LEACH routing protocol is improved in a manner where, rather than selecting the cluster head in each round, it is selected on demand. This on-demand selection will reduce the overhead caused during each round in the earlier practices. The linguistic variables for fuzzy logic used here are node degree, packet drop probability, and node centrality. On the basis of these variables, a value's chance for a node to become a cluster head is calculated. PSO is used to optimize the membership function for the better range outputs (chance).

3.3  Proposed Work

The proposed technique aims to optimize the network efficiency by improving network performance parameters and packet-delivery ratio, hence optimizing network energy consumption. The neuro-fuzzy logic is used to help in selecting appropriate cluster heads. The algorithm of the proposed work is as follow:

  • Read the total number of nodes and initial energy.

  • Initialize cluster and cluster head to 1.
  • Check for changed/updated energy values of nodes.
  • For all nodes in network form the clusters. If energy constraint satisfies, then check this condition c(i)<=(P/(1-P*mod(r,(1/P)))) and increase candidate head count. Here c(i) is random value for each node between 0 and 1 and P and r are the head probability and iteration, respectively.
  • Store cluster head id and energy and calculate distance from base station.
  • Calculate packet drop for each candidate node by using 1-(10*log10(distance.^error_rate)) here, distance is distance between candidate head and other members of cluster.
  • For all selected candidate nodes for head input energy, distance from sink and packet drop value to fuzzy model.
  • Check for the testing data (cluster head parameters fed to neuro-fuzzy model) with training data. If obligatory output, train the parameters by adjusting energy and distance.
  • Calculate the minimum distance route to perform routing.
  • Check the required results.

3.4  Simulation Results

The algorithm was implemented in Matrix Laboratory (MATLAB®) using MATLAB scripting language. Table 3.1 represents the simulation parameters taken. The network was set up randomly deployed in an area of M*M dimensions with n number of nodes and a sink node. Initially, each node has 2J of energy. A minimum distance of 30 m is set for communication between nodes.

Table 3.1   Simulation Metrics

Simulation Metrics

Parameters

Values

Area

200 m * 200 m

Energy Level

2J

Mobility Model

Random way point

Number of Nodes

Varied

Threshold distance

30 m

Number of iterations

Varied

The experiment was run for varying nodes at different iterations. Figure 3.1 represents the energy-consumption graph for the case 1, 50 nodes run for 25 iterations. It can be seen from the graph that at the 25th iteration, only 32% of energy is consumed and 68% energy is still preserved. The packet-delivery ratio is one of the important network-performance parameters also noted for 50 nodes on 25 rounds. Table 3.2 represents the packet-delivery ratio values obtained for case 1.

The graph shows the energy consumption for 25 rounds of 50 nodes. Rounds show on X-axis, and the nodes shows on Y-axis. The negative upward curve shows that the rounds are taken in 25 times and no of nodes are 50.

Figure 3.1   Energy consumption for 25 rounds of 50 nodes.

Table 3.2   Packet Delivery Ratio for Case 1

Rounds

Values of packet delivery ratio

1

1

2

1

3

1

4

1

5

1

6

1

7

.996

8

.991

9

.985

10

.98

11

.976

12

.97

13

.965

14

.96

15

.957

16

.953

17

.95

18

.948

19

.945

20

.94

Figure 3.2 represents the graph for packet-delivery ratio of 50 nodes for case 1. The average packet-delivery ratio obtained in this case is 97.58%.

Figure represents the graph for packet-delivery ratio of 50 nodes for case 1. Nodes are shown in X-axis, and no. of rounds is shown in Y-axis of graph. The average packet-delivery ratio obtained in this case is 97.58%.

Figure 3.2   Packet delivery ratio for case 1.

The simulation was run again for case 2 having 50 nodes for 50 iterations. Figure 3.3 represents the energy consumption graph for this case. It is seen from the graph that on 50th iteration 75% of energy is consumed and 25% is still left. Packet delivery ratio values obtained for this case are shown in Table 3.3.

For Figure 3.2 with 50 nodes, the simulation was conducted again for 50 iterations. This case's energy-consumption curve defines as first increase, then constant curve, and then again increases. The graph shows that by the 50th iteration, 75% of the energy has been spent and 25% is still available.

Figure 3.3   Energy consumption for 50 rounds of 50 nodes.

Table 3.3   Packet Delivery Ratio for Case

Rounds

Values of packet delivery ratio

1

1

2

1

3

1

4

1

5

1

6

1

7

.998

8

.998

9

.997

10

.997

11

.996

12

.995

13

.995

14

.9945

15

.994

16

.994

17

.993

18

.991

19

.991

20

.99

Figure 3.4 represents the graph for packet-delivery ratio of 50 nodes for case 2. The average packet-delivery ratio obtained on the 75% utilization of energy is 99.59%.

The bar graph represents the packet-delivery ratio of 50 nodes for case 2. The average packet-delivery ratio obtained on the 75% utilization of energy is 99.59%. Case 3 had the 100 nodes, and simulation was run for 25 rounds.

Figure 3.4   Packet delivery ratio for case 2.

Case 3 had the 100 nodes, and simulation was run for 25 rounds. Figure 3.5 represents the energy-consumption graph for this case. It is seen from the graph that on the 25th iteration, 45% of energy is consumed and 55% is still preserved.

The graph represents the energy consumption for this case. Nodes are shown in X-axis, and no. of rounds shown in Y-axis of graph. The curve is a negative curve; it is seen from the graph that on the 25th iteration, 45% of energy is consumed and 55% is still preserved.

Figure 3.5   Energy consumption for 25 rounds of 100 nodes.

Table 3.4 represents the values of packet-delivery ratio obtained for the case 3.

Table 3.4   Packet Delivery Ratio for Case 3

Rounds

Values of packet delivery ratio

1

1

2

1

3

1

4

1

5

1

6

1

7

.998

8

.998

9

.9965

10

.996

11

.995

12

.995

13

.995

14

.9945

15

.994

16

.993

17

.993

18

.991

19

.99

20

.99

Figure 3.6 represents the graph for packet-delivery ratio of 100 nodes for case 3. The average packet-delivery ratio obtained on the 45% utilization of energy is 99.61%. Next was testing 100 nodes for 50 iterations.

Figure represents the bar graph for packet-delivery ratio of 100 nodes for case 3. Nodes are shown in X-axis, and no. of rounds shown in Y-axis of graph. The average packet-delivery ratio obtained on the 45% utilization of energy is 99.61%. Next was testing 100 nodes for 50 iterations.

Figure 3.6   Packet delivery ratio for case 3.

Figure 3.7 represents the energy consumption graph for case 4. The graph depicts almost full consumption of energy at 50th iteration for 100 nodes.

Figure represents the energy-consumption graph for case 4. Nodes are shown in X-axis, and no. of rounds shown in Y-axis of graph. The graph depicts almost full consumption of energy at 50th iteration for 100 nodes.

Figure 3.7   Energy consumption for 50 rounds of 100 nodes.

In Figure 3.8, graph for packet delivery ratio of 100 nodes for case 4. The average packet-delivery ratio obtained on the 99.8% utilization of energy is 99.54%.

Figure shows the graph for packet-delivery ratio of 100 nodes for case 4. Nodes are shown in X-axis, and no. of rounds shown in Y-axis of graph the average packet-delivery ratio obtained on the 99.8% utilization of energy is 99.54%.

Figure 3.8   Packet delivery ratio for case 4.

Table 3.5   Packet Delivery Ratio for Case 4

Rounds

Values of packet delivery ratio

1

1

2

1

3

1

4

1

5

.998

6

.997

7

.997

8

.9965

9

.9965

10

.996

11

.996

12

.995

13

.995

14

.9945

15

.994

16

.992

17

.992

18

.991

19

.99

20

.99

Table 3.5: The graphs shows that at full consumption of energy, the obtained packet-delivery ratio is 99.54% and when only 32% energy is consumed, then the results for packet-delivery ratio obtained are also good, 97.58%. The performance metrics values, energy utilization, and packet-delivery ratio for different cases is given in Tables 3.6 and 3.7.

Table 3.6   Performance Parameters for 50 Nodes

Number of nodes = 50

Number of Iterations

Energy Consumption (%)

Packet Delivery Ratio (%)

25

32

97.58

50

75

99.59

Table 3.7   Performance Parameters for 100 Nodes

Number of nodes = 100

Number of Iterations

Energy Consumption (%)

Packet Delivery Ratio (%)

25

45

99.61

50

99.8

99.54

Maximum value obtained for the packet-delivery ratio is 99.61%, which is obtained at 45% of energy consumption. On 45% utilization of energy, good results are obtained; therefore, more than 50% of energy is still preserved.

3.5  Conclusion and Future Scope

Wireless sensor networks are deployed in thousands of numbers; hence, clustering is considered as one of the best ways for preserving energy of the nodes during communication. The proposed work aims to improve the selection of cluster heads in the network. Use of adaptive neuro-fuzzy logic includes the artificial behaviour in network in a way that data is tested against an already trained set to fit into the requirement. The fuzzy inference system works on rule sets, and combining it with artificial intelligence (ANFIS) can provide better-optimized results, which are received in our experiment as per considered metrics. The simulation results have shown good results for both the parameters considered here. However, the network's efficiency is not based on only these two parameters. As we have mentioned, other authors have used this technique with other parameters; like delay, throughput can also contribute in testing the network performance. Considering more metrics may result in more optimized and improved results. However, with multiple network metrics, it may become difficult to have a larger rule base if the fuzzy approach is used, but techniques for load optimization of rule base can be tested. Also, mobility is an important factor, which can become the part of neuro-fuzzy logic for better results.

References

1
H. Hasani and S. Babaie , Selfish node detection in ad hoc networks based on fuzzy logic, Neural Computing and Applications 3, 1–12, 2018.
2
A. Abuarqoub , M. Hammoudeh , B. Adebisi , S. Jabbar , A. Bounceur , and H. Al-Bashar , Dynamic clustering and management of mobile wireless sensor networks, Computer Networks, 117, 62–75, 2017.
3
Mian Ahmad Jan , Bin Dong , Syed Rooh Ullah Jan , Zhiyuan Tan , Sahil Verma , and Kavita , A Comprehensive Survey on Machine Learning-Based Big Data Analytics for IoT-Enabled Smart Healthcare System MONET, Springer, 2021.
4
X. Liu and Xuxun , A Survey on clustering routing protocols in wireless sensor networks, Sensors, 12(8), 11113–11153, 2012.
5
K. Akkaya and M. Younis , A survey on routing protocols for wireless sensor networks, Ad Hoc Networks, 3(3), 325–349, 2005.
6
E. G. Julie , S. Tamilselvi , and Y. H. Robinson , Performance analysis of energy efficient virtual back bone path based cluster routing protocol for WSN, Wireless Personal Communications 91(3), 1171–1189, 2016.
7
H. Yaser , S. Ali , N. Z. Jhanjhi , M. Humayun , A. Nayyar , and M. Masud , Role of fuzzy approach towards fault detection for distributed components, CMC-Computers Materials & Continua 67, (2), 1979–1996, 2021.
8
Y. G. Ha , H. Kim , and Y. C. Byun , Energy-efficient fire monitoring over cluster-based wireless sensor networks, International Journal of Distributed Sensor Networks, 2012, 2012.
9
M. Humayun , N. Z. Jhanjhi , and M. Z. Alamri , Smart secure and energy efficient scheme for E-Health applications using IoT: A review, International Journal of Computer Science and Network Security 20(4), 55–74.
10
I. Batra , S. Verma , Kavita , U. Ghosh , J. J. P. C. Rodrigues , G. N. Nguyen , A.S.M. Sanwar Hosen , and V. Mariappan , Hybrid logical security framework for privacy preservation in the green internet of things, MDPI-Sustainability 12(14), 5542, 2020.
11
I. J. Su , C. C. Tsai , and W. T. Sung , Area temperature system monitoring and computing based on adaptive fuzzy logic in wireless sensor networks, Applied Soft Computing 12(5), 1532–1541, 2012.
12
S. Hafsa , H. Ashraf , H. Javed , M. Humayun , N. Z. Jhanjhi , and M. A. AlZain , "Energy optimised security against wormhole attack in IoT-based wireless sensor networks,” CMC-Computers Materials & Continua 68(2), 1966–1980, 2021.
13
S. A. Sahaaya , A. Mary , and J. B. Gnanadurai , Enhanced zone stable election protocol based on fuzzy logic for cluster head election in wireless sensor networks, International Journal of Fuzzy Systems, 19(3), 799–812, 2017.
14
S. Soobia , et al., Analyzing the performance and efficiency of it-compliant audit module using clustering methods, in Proceedings of the Industrial Internet of Things and Cyber-Physical Systems: Transforming the Conventional to Digital. IGI Global, 351–376, 2020.
15
M. Selvi , R. Logambigai , S. Ganapathy , L. S. Ramesh , H. K. Nehemiah , and K. Arputharaj , Fuzzy temporal approach for energy efficient routing in WSN, in Proceedings of the International Conference on Informatics and Analytics, 1–5, 2016.
16
Z. Noor , L. T. Jung , F. Alsaade , and T. Alghamdi , Wireless sensor network (WSN): Routing security, reliability and energy efficiency, Journal of Applied Sciences 12(6), 593–597, 2012.
17
H. Mamoona , N. Z. Jhanjhi , M. Alruwaili , S. S. Amalathas , V. Balasubramanian , and B. Selvaraj , Privacy protection and energy optimization for 5G-aided industrial internet of things, IEEE Access 8, 183665–183677, 2020.
18
H. Chang , A hybrid intelligent protocol in sink-oriented wireless sensor networks, in Proceedings of the 2012 International Conference on Information Security and Intelligent Control, 57–60, 2012.
19
S. Maryam , H. Ashraf , A. Ullah , M. Masud , M. Azeem , N. Z. Jhanjhi , and M. Humayun , Robust cluster-based routing protocol for IoT-assisted smart devices in WSN, CMC-Computers Materials & Continua 67(3), 3505–3521, 2021.
20
N. Ghosh , I. Banerjee , and R. S. Sherratt , On-demand fuzzy clustering and ant-colony optimisation based mobile data collection in wireless sensor network, Wireless Networks 1–17, 2017.
21
H. S. Jawad , M. Irfan , N. Z. Jhanjhi , K. Hussain , and H. Mamoona , Performance enhancement in wireless body area networks with secure communication, Wireless Personal Communications 116(1), 1–22, 2021.
22
P. Rani , Kavita , S. Verma , and N. G. Nguyena , Mitigation of black hole and gray hole attack using swarm inspired algorithm with artificial neural network, IEEE Access 30, 2020.
23
B. Vijayalakshmi , K. Ramar , N. Z. Jhanjhi , S. Verma , M. Kaliappan , K. Vijayalakshmi , Kavita , U. Ghosh , and S. Vimal , An Attention Based Deep Learning Model For Traffic Flow Prediction Using Spatio Temporal Features Towards Sustainable Smart City, IJCS,Wiley, 2021.
24
S. Kar , S. Das , and P. K. Ghosh , Applications of neuro fuzzy systems: A brief review and future outline, Applied Soft Computing, 15, 243–259, 2014.
25
A. Tatar , A. Barati-Harooni , A. Naja_-Marghmaleki , B. Norouzi-Farimani , and A. H. Mohammadi , Predictive model based on an_s for estimation of thermal conductivity of carbon dioxide, Journal of Molecular Liquids 224, 1266–1274, 2016.
26
S. Verma , S. Kaur , D. B. Rawat , C. Xi , L. T. Alex , and N. Z. Jhanjhi . Intelligent framework using IoT-based WSNs for wildfire detection, IEEE Access 9, 48185–48196, 2021.
27
M. M. Afsar and M.-H. Tayarani . Clustering in sensor networks: A literature survey, Journal of Network and Computer Applications 46,198–226, 2014.
28
R. Madan , et al., Cross-layer design for lifetime maximization in interference-limited wireless sensor networks, Wireless Communications 5(11), 3142–3152, 2006.
29
J.-H. Chang and L. Tassiulas , Maximum lifetime routing in wireless sensor networks, IEEE/ACM Transactions on Networking (TON) 12(4), 609–619, 2004.
30
R. Madan and S. Lall , Distributed algorithms for maximum lifetime routing in wireless sensor networks, Wireless Communications 5(8), 2185–2193, 2006.
31
A. Giridhar and P. R. Kumar , “Maximizing the functional lifetime of sensor networks,” in Proceedings of the 4th International Symposium on Information Processing in Sensor Networks. IEEE Press, 2005.
32
S. Mottaghi and M. R. Zahabi , Optimizing LEACH clustering algorithm with mobile sink and rendezvous nodes, AEU-International Journal of Electronics and Communications 69(2), 507–514, 2015.
Search for more...
Back to top

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.