Machine Learning based Power Efficient Optimized Communication Ensemble Model with Intelligent Fog Computing for WSNs

doi:10.21203/rs.3.rs-1744293/v1

PDF

Research Article

Machine Learning based Power Efficient Optimized Communication Ensemble Model with Intelligent Fog Computing for WSNs

https://doi.org/10.21203/rs.3.rs-1744293/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Wireless sensor networks have evolved in recent years. Energy consumption is one of the major parameters to access the performance of the network. This study proposes an algorithm which is based on Ant lion optimization along with three different clustering techniques. This study is motivated from fog computing in which Ant lion optimization is applied in the bottom layer and an ensemble model of k-means, SOM and Fuzzy c-means techniques is used for clustering in FOG layer. To access model’s performance, results are compared to the traditional methods. This model has shown significant improvement in terms of energy consumption as compared to traditional model ECCM. Proposed model has also shown improvement in terms of load balancing.

k-means

Fuzzy c-means

SOM

AntLion Optimization

FOG Layer

In recent years, advancements in IOT technology have in turn led to the advancement of wireless sensor networks. Wireless sensor network is nothing but a dedicated network of sensors which are connected wirelessly. Over the years WSNs are introduced in various industries like healthcare [1], [2],, automobile [3], home and offices [4], [5]. WSNs play a vital role while implementing IOT. Performance of WSN can be determined on the basis of two aspects which are energy consumption and load balancing. Energy consumption is dependent upon the routing mechanism [6]–[8]. There are various models that are currently available to optimize route by which information travels from sensor nodes to hub [9]–[13]. Communication is nothing but transfer of data from one end to another and routing mechanism is a phenomenon by which route is selected on which data travels. Routing can be single hop or multiple hops; each hop consumes energy [14]–[17]. Energy management [18]–[21] is important because in most of the cases replacement or recharging of battery is not feasible [22]–[24].Presently available models are highly complex and consume higher amounts of energy. This study proposes a model which is more efficient in terms of load balancing and energy consumption. Proposed model is based on ant lion optimization along with an ensemble model of k-means, SOM and fuzzy c-means clustering techniques.

This study is further divided into following sections:

Related work
Material and Methods
Performance evaluation
Conclusion

In the era of IOT, importance of wireless sensor networks has increased. Energy consumption in communication is dependent on the cluster formation. Selection of optimal routes and nodes should be done in order to reduce energy consumption and latency. At present there are many techniques available for cluster formation.

Over the years numerous algorithms were introduced to optimize cluster routes. Routing algorithms can be classified into uniform and non-uniform clustering algorithms. Authors of [25] proposed non uniform cluster routing algorithm. The objective of this algorithm was to balance energy consumption in the nodes. In study [26] deployment scheme was proposed which reduced long hops in cluster into multiple small hops to reduce transmission time in the wireless sensor networks. Authors of [27] proposed a model based on fog layer computing. This model stores partial information of the nodes in advance in the fog layer, this reduces the interruptions between the communication and was helpful in moving sink node. A low-cost collaborative charging in wireless sensor networks technique was presented in study [28].

In [29] a data dissemination strategy which was based on ant colony optimization techniques was proposed. Authors named this strategy as transmission with multiple load balancing schemes (TMLBS). Load decentralization, load maintenance and load diversion schemes were used in this proposed model. Researchers of [30] proposed a relative low-cost filter-based framework FERA. This model adaptively used push/pull strategies and was able to handle data request in VANET effectively. Authors of [31] proposed DMOA which was motivated by morphological erosion and dilation on binary images. This algorithm was capable of extracting event backbone nodes and estimate the event region on their basis. A novel algorithm named as complex alliance strategy with multi-objective optimization of coverage (CASMOC) which improves the congestion in network was proposed in [32]. Authors of this study also provided a proportional relation between working node and neighboring nodes for energy conversion.

Due to the uncertainty of data, there could be coverage blind area and redundant data. To deal with these authors of [33] proposed a novel model named as Coverage Control Algorithm for Moving Target Nodes based on Sensing Probability (CMTN-SP). Survivability aware connectivity restoration strategy was proposed in study[34]. This strategy involves the design of load equilibrium mechanism and reliability enhancement measure. In study [35] a novel algorithm called energy-based k-coverage control algorithm (EBKCCA) was proposed. Based on the calculations by this algorithm authors concluded that multi transmission consumes less energy than single transmission.

In study [36] Energy efficient method for clustering is proposed. Firstly, a strong node is constructed in fog layer then data aggregation routing is constructed and at last particle swarm optimization is used to elect cluster head for effective routing.

Above discussed algorithms have shown promising results but there is still some room for advancement. Presently available algorithms are highly complex in nature. Apart from that energy efficiency is always a reason of concern.

This section involves the calculation of the energy consumption and assumptions made for simulations. Figure 3.1 shows the architecture of the WSN. Network is divided into three layers as shown in the Fig. 3.1.These layers are cloud layer, fog layer and sensing layer. This study focuses on two of them which are fog layer and sensing layer. Fog layer is a virtual layer that binds sensing layer to the cloud storage layer. This layer is responsible for clustering sensing nodes.

Fog layer helps in modulating the routes of data transfer in sensing layer. This layer is also responsible for selecting cluster heads. In other words, fog layer is majorly responsible for routing and optimal node selection. Sensing layer is the physical layer where sensors are deployed and data is gathered. This data is sent to the sink node using optimized route.

3.1 Parameters and Assumptions for Simulation Environment

In this section all the parameters and assumptions are taken in consideration for simulations. For simulation 200 to 800 sensor nodes were taken into consideration which are deployed in the range of 500 x 500 m². Length of data packet is taken as 1000 bits which (125 bytes), out of the total length of data packet 10 bytes are used as header and rest is the payload. Figure 3.2 shows the structure of data packet.

Cluster range is kept between 2% -5% of the total nodes. Unique ID is allotted to each sensor node. Nodes are uniformly randomly deployed in the simulation range. Initially all the sensors have same energy levels. Each node is capable of data sensing, data computation and data aggregation. Power consumption is dependent upon the transmission distance and all node are static whereas sink node is assumed to be moving.

Figure 3.3. Nodes representation deployed in range with (a) 200 nodes,$\lambda$ 0.08. (b) 400 nodes, $\lambda$ 0.16 (c) 800 nodes, $\lambda$ 0.32 (d) 1500 nodes, $\lambda$ 0.60

Node density is represented by $\lambda$ (number of nodes per 10m²). Figure 3.3 (a-d) shows the variation in representation with different value of $\lambda$ and different number of nodes.

3.2 Energy consumptions at Various Levels in WSN Networks

Proposed algorithm uses clustering mechanism for optimized routing. Each cluster consist of cluster members and cluster head. Data is transferred from cluster members to cluster head in single or multiple hops depending upon the location of the members with respect to the distance from nearest cluster head. Energy is consumed in three phases: one is sensing event, second is during communication and third is when node is in ideal condition. Energy consumption is maximum during the communication process. Figure 3.4 represents the energy consumption during communication process.

Equations (1–2) show the energy consumption during communication. Presented model is based on Heinzelman’s [37] first-order communication radio model.

$${E}_{Tx}\left(k,d\right)=\left\{\begin{array}{c}k*{E}_{elec}+k*{\epsilon }_{fs}*{d}^{2 }d<{d}_{0}\\ k*{E}_{elec}+k*{\epsilon }_{mp}*{d}^{4 }d\ge {d}_{0}\end{array}\right.$$

$${E}_{Rx}=k*{E}_{elec}$$

$${d}_{0}=\sqrt{\frac{{\epsilon }_{fs}}{{\epsilon }_{mp}}}$$

where ${E}_{Tx}$ is the energy consumed for transmitting $k$ bits over a distance ${d}_{j}$; ${E}_{Rx}$ is the energy consumed for receiving $k$ bits which does not depend on distance; ${d}_{0}$ is the threshold distance; and ${\epsilon }_{fs}$ and ${\epsilon }_{mp}$ are the constant values for free space and multi-fading channel path, respectively.

3.3 Proposed Network Model

Proposed model is based on fog computing. In this study optimization technique along with various clustering techniques are used which are explained in the sub-sections below:

3.3.1 Optimization using ALO in Cluster Formation

Ant lion optimization is a nature inspired optimization technique. This optimization is encouraged from the hunting technique of ant lion. Initially, an ant lion walks randomly in the range and build traps for ants. Then ants are entrapped in those traps and calculations are made based on them. These actions are repeated for each iteration. Eq. 4 is used to calculate random walks of ant lion. After every iteration location of ant lion is updated and locations are normalized.

$${X}_{j}^{it}=\frac{\left({X}_{j}^{it}-{a}_{j}\right)\times ({d}_{j}-{C}_{j}^{it})}{({d}_{j}^{it}-{a}_{j})}+{d}_{j} \left(4\right)$$

Where a_j and b_j are the minimum and maximum value of random walk of j^th ant in it^th iteration and ${d}_{j}^{it}$ denotes the maximum value of the ${j}^{th}$ ant in it^th iteration.

Ant lion optimization is used for cluster formation and optimization of cluster members and cluster heads. Single hop communication is used in cluster whereas in cross layer single and multi-hop both types of communications are considered. Predictions made by ALO for cluster numbers are dependent on the fitness function which is given in Eq. 5. Fitness function is dependent on the features like remaining energies, node-neighboring density, and nearest node distance. ALO is applied only in the bottom layer where clusters are formed. ALO is observed to be fast and efficient optimizing technique for this dedicated operation.

$$fitnessfunction=min\{{E}_{ln}+{E}_{for}\}$$

$${ E}_{ln}=\sum _{k=1}^{k=m}{E}_{Tx}(k,{CH}_{k})$$

where ${E}_{ln}$is the transmission energy of${k}^{th}$ node to its respective ${CH}_{k}{k}^{th}$cluster head with a single hopping protocol, m is the number of nodes and M is the number of clusters

$${ E}_{for}=\sum _{k=1}^{k=M}{E}_{Tx}(k,{CFOG}_{k})$$

${E}_{for}$ is the transmission energy of the ${k}^{th}$ fog node considering single hop or multi-hop communication, m is the number of nodes and M is the number of clusters.

3.3.2 Fog Clustering With k-Means

Optimized cluster heads collect data from cluster members and perform data compression respectively. These compressed data packets are further sent to clouds storage. Traditionally, these data packets were sent to clouds with direct communication link from cluster heads but in recent years, intelligent data forwarding scheme is being promoted by researchers which uses FOG layers to collect the data from cluster heads and forward it to clouds using various algorithms. This scheme reduces the energy consumption. One of the ways to cluster the cluster heads by selecting FOG member is using K-means. K-means is a good and simplest technique when it comes to preparing a cluster using various features. In this study locations of cluster heads are used as input features for preparing FOG clusters. Every FOG cluster elected its best possible FOG cluster Head, whose work is to aggregate the data sent by its nearest cluster heads. Figure 3.7(a) shows cluster formed using k-means clustering techniquewhen simulated with 800 nodes, number of FOG clusters is set to 8%, Maximum number of iterations is taken as 100 for converging the solution. For an example if number of optimized FOG clusters are denoted by k then K-means starts with k number of centroids. k-means + + algorithm [52] is used for predicting initial FOG centers. Given below are the steps for implementing k-means + + algorithm.

A random observation is selected uniformly from dataset x. Selected observation is first centroid c.

Distances are calculated from each point to c which is denoted asd(Xi, cj)

Select the next centroid, c₂ at random from X with probability as shown in Eq. 8

$$\frac{{d}^{2}\left({X}_{m},{c}_{1}\right)}{\sum _{j=1}^{n}{d}^{2}\left({X}_{j},{c}_{1}\right)}$$

4. To choose center j:

Calculate the distances from each point to each value of c, and assign each observation to its closest centroid.

For m = 1,...,n and p = 1,...,j − 1, select centroid j at random from X with probability

$$\frac{{d}^{2}\left({X}_{m},{c}_{p}\right)}{\sum _{h;{x}_{h}\in {c}_{p}}^{n}{d}^{2}\left({X}_{h},{c}_{p}\right)}$$

where C_p is the set of all observations closest to centroid c_p and x_m belongs to C_p.

5. Repeat above step 4 until k centroids are selected.

Minimum variance [38] is then calculated using Eq. 10

$$\left(\frac{1}{n}\right)\text{*}\sum \left(\underset{j}{\text{min}}{d}^{2}\left({X}_{i},{m}_{j}\right)\right), fori=1 ton$$

where d(Xi, mj) is the Euclidean distance between Xi and mj. Cluster centroids are referred as {j}k i = 1.

3.3.3 Fog Clustering using Self Organizing Maps (SOM)

Fog Layer clustering can also be done using Self Organizing Maps (SOM). This clustering technique is based on similarity in samples. SOM algorithm is inspired by mammal brain and motor mappings, which has a unique way of problem solving and organizing topologies. This study uses two features as input for SOM model which are x and y coordinates of positions of the cluster heads and 8% of output clusters with 16 layers. Weights and bias updates are generated after every epoch while clustering process and training stops when maximum number of epochs reaches which is set at 1000. Figure 3.5(a) and 3.5(b) shows the SOM neighbor connections and neighbor weights. Figure 3.5(c)and 3.5(d)represents relation of input and output weights when connected with SOM clustering. Figure 3.5(e) and 3.5(f) shows the number of cluster head are clustered with FOG cluster heads and their connected weights. It is very much visible that SOM clustering balances the cluster heads in each FOG cluster very well. Figure 3.7(b) shows final clusters when maximum number of epochs is reached.

3.3.4 Fog Clustering with Fuzzy Networks

Fuzzy c-means which is generally referred as FCM [39] is an apt way to build clusters using predefined features, this approach is based on minimizing the objective function given in Eq. 11. Fuzzy c-means allows samples to become a part of different cluster with different degree of membership. Figure 3.7(c) shows the clusters formation in FOG layer for the first iteration with overlapping value 2 of fuzzy in clustering. Objective Function achieves the value of 3.01e + 5 at 100th iteration.

$${ J}_{m}= \sum _{i=1}^{P}\sum _{j=1}^{N}{\mu }_{ij}^{m}{‖{x}_{i}-{c}_{j}‖}^{2} \left(11\right)$$

Where P and N are the number of data points and number of clusters respectively. m is the exponent of overlap control for fuzzy. x_i and c_j are the ith data point and center of the jth cluster respectively. µ_ij is the membership degree of x_i in the jth cluster. For a given data point, x_i, the sum of the membership values for all clusters is one.

3.3.5 Ensemble Model for Fog Clustering

This work involves clustering process involving FOG layer using three different clustering methods, which further used together to make an ensemble model for predicting final FOG clusters and FOG cluster heads. A weighted Sum criterion is used to predict Ensemble clusters which uses the information passed by three clustering algorithms. Figure 3.6 shows the Ensemble clustered FOG layer, which communicates, between cluster heads and cloud. Figure 3.7(d) shows the cluster made by the ensemble model

Simulation results has been taken for this study using MATLAB9.4 tool. Network performance parameters like residual energy of the sensor nodes, lifetime of network and number of alive/dead nodes are used for comparison with previous studies [31], [40], [41]. Simulation area of $500 \times 500 {m}^{2}$ is used for deployment of randomly uniformly distributed sensor nodes. Sensor nodes with density varying from 0.08 to 0.60 is used for performance evaluation. Initial energy of sensor nodes is assumed to be 5J with 10m of sensing radius.

Table 4.1 shows total residual energy of nodes for simulation time varying from 1000 secs to 2000 secs. It is observed that residual energy improves with percent of 1.35 at 1000 secs when compare with ECCM, whereas the same improvement increases at 4.23% at 2000 secs. This shows a trend of network energy saving over the network run time. Proposed work also shows significant improvement in residual energy of network when compare to algorithms like TBSR and DMOA. Figure 4.1shows the performance variation in terms of network lifetime of proposed method with existing techniques.

Table 4.1

shows total residual energy of nodes for simulation time
Algorithms	Total Networks
Algorithms	1000 (seconds)	1200 (seconds)	1400 (seconds)	1600 (seconds)	1800 (seconds)	2000 (seconds)
DMOA	3450	3250	2900	2500	2000	1500
TBSK	3500	3400	3200	2900	2600	2300
ECCM	3770	3740	3700	3650	3600	3550
Proposed	3820	3810	3800	3750	3730	3700

In this study, Number of sensor nodes are varied from 100 to 400 with resolution of 50 considered. Overall network lifetime shows small improvement when taken for lesser number of nodes and variation increases with increase in number of nodes. Calculations shows a significant improvement in network lifetime by 14.29% when clustering is done using proposed ensemble model as compare to ECCM at 400 nodes.

Another parameter which measures the number of Alive nodes at various time stamps is considered to check the load balancing of nodes as shows in Fig. 4.2. It is observed that in case of proposed technique no nodes are dead till the simulation time of 1500secs where 66 nodes were dead using ECCM approach for FOG layer. At 2000sec of simulation time proposed work has experienced greater number of alive nodes by 7.4% as compare to ECCM.

Energy consumption is always a matter of concern in terms of WSNs. This study proposes an algorithm which shows significant improvement in terms of energy consumptions when compared with the traditional methods. Proposed algorithm encourages fog computing. In this study Ant Lion optimization is used for clustering in bottom layer which finds optimal cluster members and cluster heads. This leads to the next layer which a virtual layer also known as fog layer. All the computations are made in fog layer. Proposed algorithms used an ensemble model of three clustering techniques which are k-means, SOM and Fuzzy c-means (FCM). Results of proposed model is compared to the ECCM model and it was observed that network lifetime was improved by 14.29%. During simulation another observation was that at 2000 sec of simulation time 7.4% more nodes were alive as compared to ECCM. Proposed algorithm has shown better load balancing as compared to the traditional methods.

Funding: Not applicable.

Conflict of Interest: The authors declare that they have no conflict of interest.

Availability of data and material: Not applicable.

Code availability: Not applicable

Pradhan, B., Bhattacharyya, S., & Pal, K. (2021). “IoT-Based Applications in Healthcare Devices,” J. Healthc. Eng., vol. 2021, doi: 10.1155/2021/6632599
Jagadeeswari, V., Subramaniyaswamy, V., Logesh, R., & Vijayakumar, V. (2018). A study on medical Internet of Things and Big Data in personalized healthcare system. Heal Inf Sci Syst, 6(1), 1–20. doi: 10.1007/s13755-018-0049-x
Qin, E., Long, Y., Zhang, C., & Huang, L. (2013). Cloud computing and the internet of things: Technology innovation in automobile service. Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics), 8017, 173–180. doi: 10.1007/978-3-642-39215-3_21. LNCS, no. PART 2
Menon, V. G., Jacob, S., Joseph, S., Sehdev, P., Khosravi, M. R., & Al-Turjman, F. (2020). “An IoT-enabled intelligent automobile system for smart cities,” Internet of Things, no. xxxx, p. 100213, doi: 10.1016/j.iot.2020.100213
Froiz-Míguez, I., Fernández-Caramés, T. M., Fraga-Lamas, P., & Castedo, L. (2018). Design, implementation and practical evaluation of an iot home automation system for fog computing applications based on MQTT and ZigBee-WiFi sensor nodes. Sensors (Switzerland), 18(8), 1–42. doi: 10.3390/s18082660
Wang, T., Luo, H., Zheng, X., & Xie, M. (2019). Crowdsourcing mechanism for trust evaluation in CPCS based on intelligent mobile edge computing. ACM Trans Intell Syst Technol, 10(6), doi: 10.1145/3324926
Liu, X., & Zhang, P. (2018). Data Drainage: A Novel Load Balancing Strategy for Wireless Sensor Networks. Ieee Communications Letters, 22(1), 125–128. doi: 10.1109/LCOMM.2017.2751601
Teng, H., et al. (2019). A novel code data dissemination scheme for Internet of Things through mobile vehicle of smart cities. Futur Gener Comput Syst, 94, 351–367. doi: 10.1016/j.future.2018.11.039
Wang, T., Liang, Y., Jia, W., Arif, M., Liu, A., & Xie, M. (2018). “Coupling resource management based on fog computing in smart city systems,” J. Netw. Comput. Appl., vol. 135, no. September pp. 11–19, 2019, doi: 10.1016/j.jnca.2019.02.021
Wu, J., Dong, M., Ota, K., Li, J., Yang, W., & Wang, M. (2019). Fog-computing-enabled cognitive network function virtualization for an information-centric future internet. Ieee Communications Magazine, 57(7), 48–54. doi: 10.1109/MCOM.2019.1800778
Sun, Z., Xing, X., Wang, T., Lv, Z., & Yan, B. (2019). An optimized clustering communication protocol based on intelligent computing in information-centric internet of things. Ieee Access : Practical Innovations, Open Solutions, 7, 28238–28249. doi: 10.1109/ACCESS.2019.2896250
Liu, X. (2017). “Routing Protocols Based on Ant Colony Optimization in Wireless Sensor Networks: A Survey,” IEEE Access, vol. 5, no. c, pp. 26303–26317, doi: 10.1109/ACCESS.2017.2769663
Yi, S., Li, C., & Li, Q. (2015). “A survey of fog computing: Concepts, applications and issues,” Proc. Int. Symp. Mob. Ad Hoc Netw. Comput., vol. 2015-June, no. August, pp. 37–42, doi: 10.1145/2757384.2757397
Lu, J., Xin, Y., Zhang, Z., Liu, X., & Li, K. (2018). Game-Theoretic Design of Optimal Two-Sided Rating Protocols for Service Exchange Dilemma in Crowdsourcing. Ieee Transactions On Information Forensics And Security, 13(11), 2801–2815. doi: 10.1109/TIFS.2018.2834318
Liu, X., Qiu, T., Zhou, X., Wang, T., Yang, L., & Chang, V. (2020). Latency-aware path planning for disconnected sensor networks with mobile sinks. IEEE Trans Ind Informatics, 16(1), 350–361. doi: 10.1109/TII.2019.2916300
Chen, X., et al. (2020). Android HIV: A Study of Repackaging Malware for Evading Machine-Learning Detection. Ieee Transactions On Information Forensics And Security, 15(8), 987–1001. doi: 10.1109/TIFS.2019.2932228
Wu, Y. K., Huang, H., Wu, Q., Liu, A., & Wang, T. (2019). A risk defense method based on microscopic state prediction with partial information observations in social networks. Journal Of Parallel And Distributed Computing, 131, 189–199. doi: 10.1016/j.jpdc.2019.04.007
Lin, X., Li, J., Wu, J., Liang, H., & Yang, W. (2019). Making Knowledge Tradable in Edge-AI Enabled IoT: A Consortium Blockchain-Based Efficient and Incentive Approach. IEEE Trans Ind Informatics, 15(12), 6367–6378. doi: 10.1109/TII.2019.2917307
Adhianto, L., et al. (2010). HPCTOOLKIT: Tools for performance analysis of optimized parallel programs. Concurr Comput Pract Exp, 22(6), 685–701. doi: 10.1002/cpe
Wang, T., Zhou, J., Liu, A., Bhuiyan, M. Z. A., Wang, G., & Jia, W. (2019). Fog-based computing and storage offloading for data synchronization in IoT. IEEE Internet Things J, 6(3), 4272–4282. doi: 10.1109/JIOT.2018.2875915
Yang, G., Liang, T., He, X., & Xiong, N. (2019). Global and local reliability-based routing protocol for wireless sensor networks. IEEE Internet Things J, 6(2), 3620–3632. doi: 10.1109/JIOT.2018.2889379
Zhang, G., Wang, T., Wang, G., Liu, A., & Jia, W. (2018). “Detection of hidden data attacks combined fog computing and trust evaluation method in sensor-cloud system,” Concurr. Comput., no. November, pp. 1–13, doi: 10.1002/cpe.5109
Sun, Z., Tao, R., Li, L., & Xing, X. (2017). A new energy-efficient multi-target coverage control protocol using event-driven-mechanism in wireless sensor networks. Int J Online Eng, 13(2), 53–67. doi: 10.3991/ijoe.v13i02.6465
Lu, J., Xin, Y., Zhang, Z., Peng, H., & Han, J. (2018). Supporting user authorization queries in RBAC systems by role–permission reassignment. Futur Gener Comput Syst, 88, 707–717. doi: 10.1016/j.future.2018.01.010
Wang, T., Luo, H., Jia, W., Liu, A., & Xie, M. (2020). An Intelligent Trust Evaluation Scheme in Sensor-Cloud-Enabled Industrial Internet of Things. IEEE Trans Ind Informatics, 16(3), 2054–2062. doi: 10.1109/TII.2019.2930286
Liu, X. (2017). Node deployment based on extra path creation for wireless sensor networks on mountain roads. Ieee Communications Letters, 21(11), 2376–2379. doi: 10.1109/LCOMM.2017.2739727
Wang, T., et al. (2017). Trajectory Privacy Preservation Based on a Fog Structure for Cloud Location Services. Ieee Access : Practical Innovations, Open Solutions, 5(8), 7692–7701. doi: 10.1109/ACCESS.2017.2698078
Liu, T., Wu, B., Wu, H., & Peng, J. (2017). Low-Cost Collaborative Mobile Charging for Large-Scale Wireless Sensor Networks. Ieee Transactions On Mobile Computing, 16(8), 2213–2227. doi: 10.1109/TMC.2016.2616309
Liu, X., Qiu, T., & Wang, T. (2019). Load-Balanced Data Dissemination for Wireless Sensor Networks: A Nature-Inspired Approach. IEEE Internet Things J, 6(6), 9256–9265. doi: 10.1109/JIOT.2019.2900763
Lai, Y., Lin, H., Yang, F., & Wang, T. (2019). Efficient data request answering in vehicular Ad-hoc networks based on fog nodes and filters. Futur Gener Comput Syst, 93, 130–142. doi: 10.1016/j.future.2018.09.065
Nie, Y., Wang, H., Qin, Y., & Sun, Z. (2017). Distributed and morphological operation-based data collection algorithm. Int J Distrib Sens Networks, 13(7), doi: 10.1177/1550147717717593
Sun, Z., Zhang, Y., Nie, Y., Wei, W., Lloret, J., & Song, H. (2017). CASMOC: a novel complex alliance strategy with multi-objective optimization of coverage in wireless sensor networks. Wirel Networks, 23(4), 1201–1222. doi: 10.1007/s11276-016-1213-3
Sun, Z., Xing, X., Yan, B., & Lv, Z. (2019). CMTN-SP: A novel coverage-control algorithm for moving-target nodes based on sensing probability model in sensor networks. Sensors (Switzerland), 19(2), doi: 10.3390/s19020257
Liu, X. (2017). Survivability-Aware Connectivity Restoration for Partitioned Wireless Sensor Networks. Ieee Communications Letters, 21(11), 2444–2447. doi: 10.1109/LCOMM.2017.2699174
Sun, Z., Zhang, Y., Xing, X., Song, H., Wang, H., & Cao, Y. (2016). EBKCCA: A novel energy balanced k-coverage control algorithm based on probability model in wireless sensor networks. Ksii Transactions On Internet And Information Systems, 10(8), 3621–3640. doi: 10.3837/tiis.2016.08.011
Zhou, Z., Liao, H., Gu, B., Huq, K. M. S., Mumtaz, S., & Rodriguez, J. (2018). Robust Mobile Crowd Sensing: When Deep Learning Meets Edge Computing. IEEE Netw, 32(4), 54–60. doi: 10.1109/MNET.2018.1700442
Heinzelman, W. B., Chandrakasan, A. P., & Balakrishnan, H. (2002). An application-specific protocol architecture for wireless microsensor networks. IEEE Trans Wirel Commun, 1(4), 660–670. doi: 10.1109/TWC.2002.804190
Sasikumar, P., & Khara, S. (2012). “k-means clustering in wireless sensor networks,” doi: 10.1109/CICN.2012.136
Bezdek, J. C. (1981). Pattern Recognition with Fuzzy Objective Function Algorithms. Pattern Recognit with Fuzzy Object Funct Algorithms. doi: 10.1007/978-1-4757-0450-1
Sun, Z., et al. (2019). An Energy-Efficient Cross-Layer-Sensing Clustering Method Based on Intelligent Fog Computing in WSNs. Ieee Access : Practical Innovations, Open Solutions, 7, 144165–144177. doi: 10.1109/ACCESS.2019.2944858
Tang, J., Liu, A., Zhang, J., Xiong, N. N., Zeng, Z., & Wang, T. (2018). A trust-based secure routing scheme using the traceback approach for energy-harvesting wireless sensor networks. Sensors (Switzerland), 18(3), 4–9. doi: 10.3390/s18030751

PDF

Reviewers agreed at journal
14 Jun, 2022
Reviewers invited by journal
14 Jun, 2022
Editor assigned by journal
12 Jun, 2022
First submitted to journal
09 Jun, 2022

You are reading this latest preprint version

Machine Learning based Power Efficient Optimized Communication Ensemble Model with Intelligent Fog Computing for WSNs

Status:

Version 1

Abstract

Figures

1. Introduction

2. Related Works

3. Material And Methods

3.1 Parameters and Assumptions for Simulation Environment

3.2 Energy consumptions at Various Levels in WSN Networks

3.3 Proposed Network Model

3.3.1 Optimization using ALO in Cluster Formation

3.3.2 Fog Clustering With k-Means

4. To Choose Center :

5. Repeat Above Step 4 Until Centroids Are Selected.

3.3.3 Fog Clustering using Self Organizing Maps (SOM)

3.3.4 Fog Clustering with Fuzzy Networks

3.3.5 Ensemble Model for Fog Clustering

4. Results & Discussion

5. Conclusion

Declarations

References

Status:

Version 1