Available via license: CC BY 4.0
Content may be subject to copyright.
ML-Based Adaptive MAC Protocol for Real Time
Data Transmission in Wireless Sensor Networks
Archana R Raut Archana Kakade
Rashtrasant Tukadoji Maharaj Nagpur University
Research Article
Keywords: Wireless Sensor Networks, Machine Learning, MAC, real-time, Priority, adaptive learning
Posted Date: May 21st, 2024
DOI: https://doi.org/10.21203/rs.3.rs-4398735/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Additional Declarations: No competing interests reported.
ML-Based Adaptive MAC Protocol for Real Time Data
Transmission in Wireless Sensor Networks
Dr. Archana R Raut*1, Dr. S. P. Khandait2
1Assistant Professor, CSE Department, G H Raisoni College of Engineering, Nagpur, India
2Professor, IT Department, K D K College of Engineering, Nagpur, India
Abstract
Wireless Sensor Networks (WSNs) are expressively utilized in various real-time control and monitoring
applications. WSNs have been expanded considering the necessities in industrial time-bounded applications to
support the dependable, time-bound delivery of data. Recently, Machine Learning (ML) algorithms have been used
to address various WSN-related issues. The use of machine learning techniques supports dynamically modifying
MAC settings based on traffic patterns and network conditions. In WSNs to control the communication between a
large number of tiny, low-power sensor nodes while preserving energy and reducing latency, effective MAC
protocols are essential. This paper addresses the ML-based Adaptive MAC (ML-MAC) protocol to provide a
priority-based transmission system. In this research, depending upon the predictions of the machine learning model,
the MAC parameters are dynamically adjusted to find priority-based channel access and the optimal routing path to
meet the deadline of critical data packets. From the result analysis, the average throughput and delay of the
proposed ML-MAC algorithm is improved as compared to the existing I-MAC protocol.
© 2017 Elsevier Inc. All rights reserved.
Keywords: Wireless Sensor Networks; Machine Learning; MAC; real-time; Priority; adaptive learning.
1. Introduction
A wireless sensor network (WSN) accommodatingly observes the physical or environmental
conditions. Initially, WSNs were mostly planned for applications like military area observation and
environmental and industrial control and monitoring. In these applications increasing the energy
efficiency and network lifetime were the primary design objectives. Providing the best-effort data
transmission with improved network lifetime is worth it in such situations. Currently, WSNs have been
used in various real-time control and monitoring applications, like object tracking, environmental
monitoring, home monitoring, localization, personal health monitoring, industrial control and
monitoring, etc. [1-4]. Most of the applications stated above require minimum end-to-end delay with
reduced packet losses while data transmissions. These rising applications are revealed by requirements
concerning Quality of Service (QoS) fulfillment such as time-bound and reliable data transmission based
on the criticality of data packets [5-7]. In time-critical applications, whenever critical data packets
arrive, they must be transmitted within specific time bounds. Excessive delay in such cases may result in
system uncertainty, financial harm, or social life risk in the employed zone.
A lot of studies and improvements have been passed out in designing new architectures and protocols
for minimizing the energy depletion of the wireless sensor network. However, very few researches have
been carried out to increase the WSN performance in terms of provisioning QoS. In real-time
monitoring applications, a wireless sensor network is supposed to be an operational, useful and highly
probable tool. Some predominant real-time (RT) application areas in WSNs where provisioning the QoS
is an important concern are mentioned in Figure 1.
2
Fig. 1. WSN Real-Time Applications
2. Literature Survey
To support the time-bounded performance of the network when developing time-bounded
applications for WSNs, we also need to analyze the resource constraints and node-to-node transmission
reliability. Numerous MAC and routing protocols with a focus on quality of service have recently been
made available in wireless sensor networks [8-10]. Since most present protocols consider energy to be a
valuable resource in sensors, they focus on increasing energy effectiveness. Additionally, a few other
challenges arise as a result of the RT time-constrained data delivery applications that demand new
routing protocols to effectively utilize the sensor nodes and gather data in a short amount of time. The
network layer is accountable for routing data packets to the destination node within bounded time limits,
while the MAC layer is accountable for ensuring the delay when accessing the channel.
While designing traditional network layer protocol parameters such as end-to-end route detection,
resource reservation is mainly considered. But even if having similar characteristics, it may not be
appropriate for WSN for various causes. Spending time in optimal route detection is not tolerable for
critical non-periodic packets. Also, reserving the resources for such irregular non-periodic packets is
again inconvenient. Even if the traffic is periodic and continuous, still these methods are not concrete in
dynamically changing WSN environment due to disturbance in service while path recovery increases the
delay in data transfer which is unacceptable in time bounded emergency applications. Guaranteeing the
QoS in varied traffic types is a challenging problem at the network layer in WSNs because of the
following features of the WSN:
• Dynamic topology changes: Topology changes in the network are caused by the failure of
nodes or due to node mobility.
• Large-scale sensor nodes.
• Generation of periodical plus un-periodical traffic in the network having diverse urgencies and
different real-time application requirements.
• Data redundancy is formed by simultaneous sensor nodes.
2.1. Machine Learning Algorithms in WSNs
3
In the ML approach, the system obtains the information from the analysis of the existing data which
is updated timely. Using ML-based systems real-time decisions can be taken dynamically without being
explicitly programmed [11-16]. Using ML techniques also makes the computations reliable, more
dominant, and efficient even in the case of complex data. Figure 2 summarizes various machine learning
approaches with methods associated with each approach. The appeal of ML algorithms lies in their use
of architecture to learn and improve performance to offer generic solutions. Due to its interdisciplinary
character, it is essential in several disciplines, including engineering, medicine, and computers. The vast
volume of sensor data can be simply gathered and analyzed using an ML-based approach to excerpt
relevant evidence from the sensor data.
Fig. 2. Machine Learning Approaches and Methods
The three primary types of machine learning algorithms are supervised, unsupervised, and
reinforcement learning. Currently, machine learning (ML) techniques are proposed to resolve issues and
find optimized solutions for real-time data communication applications in WSNs. Different ML-based
techniques and algorithms like classification, regression, Decision tree, clustering, Association rules,
Neural Networks etc. are used in WSNs to fulfill the QoS necessities of various real-time applications.
Several of the applications for ML in WSNs are listed in [17–19] as follows.
• Tracking throughout time any rapid dynamic changes in the ecosystems.
• Calibrating a WSN to get new environmental information.
• Modeling complicated systems, which are challenging for mathematicians to model.
• Extracting important information from sensor data and suggesting system improvements in
advance.
• Developing wise judgment and self-governing control systems.
There have been numerous ML algorithms used in WSNs that have been emphasized [11-25], such
as localization, Target tracking, Congestion control, Anomaly detection, Routing, Data aggregation,
Energy harvesting, Quality of Service, MAC, Event detection etc. Recent expansions have used ML to
address a variety of WSN-related issues [17–25]. The use of ML will increase the effectiveness of
WSNs and reduce the need for human intervention or reprogramming. To tackle the problems with QoS
provisioning in WSNs, we have reviewed earlier work. Additionally, we analyzed machine learning-
based methods used to solve various problems in WSNs in more recent times. Figure 1 gives different
aspects of current methodologies used in various application domains. Many researchers have
developed energy and latency-efficient communication protocols using ML-based techniques for WSNs
[26], to improve the wireless sensor network performance of WSNs a few deep learning techniques is
conducted in [27-28]. According to the current situation, routing decisions can be made similarly to
those made by a humanoid mind using machine learning algorithms. This extra intelligence will aid the
4
network in identifying the network topology and link settings variations so that it can update its
situational awareness and adjust the protocol parameters appropriately.
Time Division Multiple Access, or TDMA, allots a specific time slot for transmission to each node
by dividing the available time into fixed-size slots. By ensuring that transmissions are planned ahead of
time, this lowers contention and collisions. For time-sensitive applications where exact timing is crucial,
TDMA may be appropriate.
A MAC layer protocol called TSCH (Time-Slotted Channel Hopping) combines channel hopping
and TDMA [29]. Nodes synchronize their schedules to send and receive within the time windows that
are allocated for this purpose. Channel hopping increases dependability and reduces interference. Time -
sensitive applications that need dependability and energy economy can benefit from TSCH.
A routing protocol called RPL was created specifically for lossy and low-power WSNs [30]. It
supports two different routing data modes: storing mode and non-storing mode. RPL can assist in
ensuring that time-sensitive data is delivered on schedule by effectively routing data across the network.
In queuing based on the priorities method, the criticality or temporal sensitivity of each packet
determines its priority [31]. High-priority packets are transmitted by nodes ahead of low-priority ones.
This guarantees that, even in the event of congestion or contention, time-sensitive data gets sent
promptly. Using this method, the criticality or temporal sensitivity of each packet determines its
priority. High-priority packets are transmitted by nodes ahead of low-priority ones. This guarantees that,
even in the event of congestion or contention, time-sensitive data gets sent promptly. Accordingly, the
prioritizing for accessing the medium will be accomplished by the medium access control layer [32-36].
Based on the above survey, several MAC protocols are created expressly to reduce data transfer
latency. These protocols frequently use techniques like shortened contention windows and quick
acknowledgment schemes, giving priority to time-sensitive traffic over non-essential traffic.
Machine learning-based I-MAC (Intelligent MAC protocol) is suggested in [37]. Based on the
characteristics of the data packets, an intelligent choice system selects either a TDMA-based or
CSMA/CA-based MAC. Here, both the impact of inherent as well as external parameters is taken into
consideration to address the application restriction problem experienced in one MAC protocol. Because
competitive and non-competitive MAC protocols function differently, the author describes a machine-
learning strategy for identifying the most appropriate protocol model based on a scenario.
It is the most appropriate model using machine learning to select the most suitable MAC protocol
specifically based on the existing situation using fuzzy set theory. Thus, it helps to progress the decision
ability of protocol, by using a strong and operational protocol selection scheme. It also uses multi-
granule expandable knowledge-gaining procedures to construct a training data set that develops
progressively to get the accurate classification outcome. As stated above, dynamic TDMA as well as
CSMA/CA, are the two main MAC protocols discussed in the research [37] to support RT applications.
Traditional machine learning is built upon adaptive machine learning to produce a more sophisticated
response to dynamic environments with fluctuating input. Adaptive machine learning, as its name
suggests, can adjust to quickly changing data sets, which increases its applicability to real-world
scenarios. Compared to classic machine learning, adaptive ML is more reliable and effective and
integrates agility, improved accuracy, and higher sustainability. Large amounts of data may be
processed by adaptive ML, and its operational parameters can be more easily changed as the
requirements of the firm using it change. Adaptive ML can quickly adjust to new information and offer
in-the-moment insight into the potential applications of that data.
In the proposed system adaptive-based machine-learning model is used. Some of the benefits of such
a protocol are:
• Better Performance: The protocol may optimize network performance in real-time, resulting in
higher throughput, lower latency, and increased energy efficiency by dynamically adjusting
MAC parameters.
5
• Robustness: The protocol is more resistant to oscillations and interference since it can adjust to
shifting network circumstances and traffic patterns.
• Energy Efficiency: The protocol can assist extend the life of the network by preserving energy
by intelligently controlling transmission power and communication schedules.
3. Proposed Methodology
In the suggested approach, we have employed a machine learning-based methodology to achieve the
requirements for service quality in WSNs. Figure 3 gives the system model of the proposed approach.
Firstly, the data collected by sensor nodes are classified based on the criticality of data packets after
comparing them with the benchmark value called a threshold value (Th). The threshold value is decided
based on application requirements. When an emergency data packet is identified, the communication is
carried out using a proposed machine learning MAC (ML-MAC) algorithm. The proposed algorithm
finds the optimal routing path for transmission of critical data packets which ensures that the closest and
least energy-intensive route will be chosen at that moment based on the current network situation. The
normal data packets in this scheme must wait until all the time-critical packets have successfully
reached their destinations. When multiple packets with the same priority emerge at once, a queue is
maintained to broadcast packets according to the principle of FIFO (first come, first served).
Fig. 3. Proposed System Model
The ideal routing path for the delivery of crucial data packets is modified in the suggested algorithm
to account for changes in wireless network settings. This gives the network an edge in recognizing the
changes in the network and deviations in link settings so that required parameters can be adjusted
dynamically.
Sensor Network Environment
Sensor Node Data Collection
Data Prioritization based on
Threshold value (Th)
Choose Optimal
Route using
ML-MAC
Data
Communication
If data>Th
Move to Priority
Queue
If Pr.Queue
empty?
Y
Y
N
N
6
Figure 4 shows the system flow chart of the proposed model. In the proposed methodology, an
intermediate node is selected based on the K-nearest neighbors (KNNs) classification algorithm. The
distance between each node pair is determined using Euclidean distance from the source node to the
destination node. After the selection of intermediate nodes, the total cost of the route is calculated
considering the distance, residual energy of the nodes, and link quality in the selected route as
Where, dSD is the Euclidean distance between a source node and the destination node, dn,n+1 is the
direct distance between nodes n and n+1, Ei, En are the initial and residual energy of a node n
respectively, and Ln is the link quality of the selected channel c. Based on the value of tc for each
routing path, the route_cost value is calculated.
Fig. 4. Proposed System Flowchart
7
Where, Nr is the number of routing nodes in the transmission path. In the end, the number of
solutions are generated as shown in Table 1. The path having the highest route_cost value is considered
to be an optimal routing path and is selected for packet transmission. Table 1 shows the path identified
based on source node ‘5’ and destination node ‘41’. As shown in Table 1 total of 19 paths were
identified after the last iteration. Also, route_cost value is calculated at every iteration and the final path
for transmission is selected based on the highest route_cost value.
In the end, each solution has a route_cost value that is calculated. To choose the best path, the path
with the highest route_cost values is taken into account. The ultimate best path for data transmission is
selected depending on the results of all iterations. In the previously discussed example, the optimal data
transmission route from source node '5' to destination '41' was chosen as 5-> 8-> 39-> 41.
Normal data packets will now have a longer end-to-end latency thanks to this method. Lower energy
usage and a higher packet delivery ratio increase the system's overall performance. As a result, it is
quite beneficial for applications utilizing wireless networks that are sophisticated and smart.
Table. 1. ML-MAC-based Routing Table
Packet scheduling for different types of data is shown in figure 5. As shown in figure 5, if 4 packets
are arrived P1, P2, P3 and P4 with P1, P2, P3 as priority data packets and P4 as normal data packets.
Based on the arrival time of priority data packets a Priority queue (Pr. QUEUE) is maintained and
packets are transmitted one after another till the queue becomes empty. The normal packets will be
transmitted after the successful transmission of all critical data packets. Here priority ‘1’ is considered
as the high priority and ‘0’ is considered as the low priority or normal data.
Source
Node (S)
Destination
Node (D)
Solution
No. (Ns)
Relay
Nodes
total_cost
route_cost
5
41
0
15, 37
10.98
5.49
5
41
1
8, 39
37.316
18.658
5
41
2
2
5.91
5.91
5
41
3
15, 38
8.81
4.405
5
41
4
16
5.94
5.94
5
41
6
40, 35
7.29
3.645
5
41
7
13, 20
7.76
3.88
5
41
8
8
5.811
5.811
5
41
9
2, 47, 32
8.76
2.92
5
41
10
13
5.88
5.88
5
41
11
38, 39
7.95
3.975
5
41
12
38
5.9
5.9
5
41
14
18, 20
7.31
3.655
5
41
15
1, 13, 25
9.49
3.16333
5
41
16
15, 19
9.19
4.595
5
41
17
19
5.64
5.64
5
41
19
16, 31
7.54
3.77
8
Fig. 4. Packet scheduling
4. Results and Discussion
The proposed protocol (ML-MAC) is implemented in NS-2 network simulator. The simulation
environment settings used for confirming and calculating the performance of the proposed scheme are
as follows:
Routing Protocols: AODV
Number of nodes: 60
MAC: TDMA
PHY: 802.11
Packet size: 1000 bytes
Simulation time: 500s
Packet interval: .01 s.
Map Size: 1000 X 1000 m
9
Fig. 5. Communication amongst sensor nodes
Figure 5 shows the implementation of the proposed protocol on network simulator NS2. Grounded
on the existing literature, we compared the performance of the proposed algorithm with the existing
protocols taking into account some performance metrics for quality of service assurance. The proposed
system is equated with I-MAC considering TDMA MAC taken into account parameters such as
throughput and delay in transmission.
Figure 6 shows the comparison between protocols I-MAC and ML-MAC considering the metric
transmission delay. Delay is evaluated as the time gap amongst the times when the packet was sent by
the source node till the time it reached by the target node. Figure 5 shows, the average delay versus
intervened time plotted for I-MAC and ML-MAC algorithms.
Fig. 6. Average Delay over elapsed time I-MAC vs. ML-MAC
10
Figure 7 shows the comparison between protocols I-MAC and ML-MAC considering average
throughput metric. It is calculated as the total information the network is able to manage and
administered by the system within specific time period. It also helps to evaluate the overall network
performance based on the number of successful communications in the network. A plot of the average
throughput verses intervened time is shown in figure 7, for I-MAC and ML-MAC algorithms.
Fig. 7. Average Throughput over elapsed time I-MAC vs. ML-MAC
From the result analysis, it is clear that the average throughput and delay of the proposed algorithm is
improved as compared to the existing I-MAC protocol. It shows that under varying network
environments, our algorithm performs superior in terms of transmission delay and throughput over
existing schemes. Considering mission-critical applications, it is found that according to changing
network conditions, the suggested scheme provides reliability and delay.
5. Conclusion
Many MAC and network layer procedures were projected in WSNs as per the existing literature to
offer time-restricted data transportation. The proposed MC-MAC algorithm uses an adaptive learning
approach for channel allocation and optimal routing path selection to support mission-critical data
transfer in WSNs. A machine learning approach is employed to select the most efficient and optimal
routing path for the transmission of emergency data packets specifically based on the existing situation.
Based on the experimental findings the average throughput of ML-MAC is 22% more than that of I-
MAC. Similarly, the average transmission delay proposed ML-MAC algorithm is 75% less as compared
to the existing I-MAC protocol. It shows that under varying network environments, the proposed
algorithm achieves improvement in terms of transmission delay and throughput over existing schemes.
The suggested ML-MAC techniques are hence very advantageous for smart wireless sensor network
management. Recently ML-MAC has been designed considering the static sensor nodes in the network
and numerical data collected from sensor nodes. In the future, it can be extended to work in dynamic
sensor node networks with image data classification and transmission to the destination grounded on a
priority of data packets.
11
References
[1] Yanjun Li, Chung Shue Chen, Ye-Qiong Song, Zhi Wang. Real-time QoS support in wireless sensor networks: a survey. 7th
IFAC International Conference on Fieldbuses & Networks in Industrial & Embedded Systems - FeT'2007, Nov 2007,
Toulouse, France. 2007.
[2] T. Arampatzis, J. Lygeros,, S. Manesis, “A survey of applications of wireless sensors and wireless sensor networks,” in
Proceedings of the 20th IEEE International Symposium on Intelligent Control (ISIC 05), pp. 719–724, June 2005.
[3] Jiang Wu, Xuefeng Ding, "Using Wireless Sensor Network to Remote Real-Time Monitoring and Tracking of Logistics Status
Based on Difference Transmission Algorithm", Journal of Sensors, vol. 2021, Article ID 4084288, 10 pages, 2021.
https://doi.org/10.1155/2021/4084288.
[4] B. Dalal, S. Kukarni, S. Yellampalli, “Wireless Sensor Networks: Applications”, Published in the book “Wireless Sensor
Networks - Design, Deployment and Applications”, 2021, Ch. 1 https://doi.org/10.5772/intechopen.97079, 978-1-83880-910-
2
[5] D. Chen and P. K. Varshney, “QoS Support in Wireless Sensor Network: A Survey,” Proceedings of International Conference
on Wireless Networks (ICWN2004), Las Vegas, Nevada, USA, June 2004.
[6] Josip Balen, D Zagar, G Martinovic, “Quality of Service in Wireless Sensor Networks: A Survey and Related Patents”.
[7] Meena Pundir, Jasminder Kaur Sandhu, “A Systematic Review of Quality of Service in Wireless Sensor Networks using
Machine Learning: Recent Trend and Future Vision”, Journal of Network and Computer Applications, Volume 188, 2021,
103084, ISSN 1084-8045, https://doi.org/10.1016/j.jnca.2021.103084.
[8] Petcharat Suriyachai, Utz Roedig, and Andrew Scott, “A Survey of MAC Protocols for Mission-Critical Applications in
Wireless Sensor Networks”, IEEE Communications Surveys & Tutorials, Vol. 14, No. 2, Second Quarter 2012.
[9]W. Ye, J. Heidemann, and D. Estrin,”An energy-efficient MAC protocol for wireless sensor networks,” in Proc. 21st Annu.
Joint Conf. IEEE Computer and Communications Societies, New York, NY, USA, 2002, vol. 3, pp. 1567-1576.
[10]T. V. Dam, and K. Langendoen, “An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks,” in Proc. 1st
ACM Conf.Embedded Networked Sensor Systems, Los Angeles, CA, USA, 2003, pp. 171-180.
[11] T. O. Ayodele, "Types of machine learning algorithms" in New Advances in Machine Learning, Rijeka, Croatia:InTech,
2010.
[12] M. A. Alsheikh, S. Lin, D. Niyato and H. Tan, "Machine Learning in Wireless Sensor Networks: Algorithms, Strategies, and
Applications," in IEEE Communications Surveys & Tutorials, vol. 16, no. 4, pp. 1996-2018, Fourthquarter 2014, doi:
10.1109/COMST.2014.2320099.
[13] J. Prajapati and S. C. Jain, "Machine Learning Techniques and Challenges in Wireless Sensor Networks," 2018 Second
International Conference on Inventive Communication and Computational Technologies (ICICCT), Coimbatore, India, 2018,
pp. 233-238, doi: 10.1109/ICICCT.2018.8473187.
[14] Praveen Kumar, TarachandAmgoth, Chandra Sekhara, Rao Annavarapu, “Machine learning algorithms for wireless sensor
networks: A survey”, Information Fusion, Volume 49, September 2019, Pages 1-25.
[15] Raut, A.R., Khandait, S.P., Chavhan, N. (2021). QoS Aware Machine Learning Algorithms for Real-Time Applications in
Wireless Sensor Networks. In: Advances in Automation, Signal Processing, instrumentation, and Control. Lecture Notes in
Electrical Engineering, vol 700, pp. 2665-2673, Springer, Singapore. https://doi.org/10.1007/978-981-15-8221-9_2494.
[16] Kai Lin, Joel J. P. C. Rodrigues, “Energy Efficiency QoS Assurance Routing in Wireless Multimedia Sensor Networks”,
IEEE SYSTEMS JOURNAL, VOL.5, NO.4, DECEMBER 2011.
[17] S. Kamalesh & P. Ganesh Kumar (2017) Data aggregation in wireless sensor network using SVM-based failure detection and
loss recovery, Journal of Experimental & Theoretical Artificial Intelligence, 29:1, 133-147, DOI:
10.1080/0952813X.2015.1132262.
[18] Chaoyun Zhang, Paul Patras, and Hamed Haddadi, “Deep Learning in Mobile and Wireless Networking: A Survey”, IEEE
COMMUNICATIONS SURVEYS & TUTORIALS, arXiv: 1803.04311v2, 17 Sep 2018, pp. 1-53.
[19] Meena Pundir, Jasminder Kaur Sandhu, “A systematic review of Quality of Service in Wireless Sensor Networks using
Machine Learning: Recent trend and future vision, 2021, Journal of Network and Computer Applications – Article.
[20] J. Barbancho, C. León, F. Molina, and A. Barbancho, “A new QoS routing algorithm based on self-organizing maps for
wireless sensor networks,” Telecommunication Systems, vol. 36, pp. 73–83, 2007.
[21] Shrawankar, U. N.; Khandait, S. P.; Raut, Archana R., “E-MAC: Efficient MAC Protocol for Time-Critical Transmission in
Wireless Sensor Networks”, HELIX, 2018, DOI: 10.29042/2018-3853-3857
[22] S.K. Gharghan, R. Nordin, M. Ismail, J.A. Ali, Accurate wireless sensor localization technique based on hybrid PSO-ANN
algorithm for indoor and outdoor track cycling, IEEE Sens. J. 16 (2) (2016) 529–541.
[23] S. Phoemphon, C. So-In, D.T. Niyato, A hybrid model using fuzzy logic and an extreme learning machine with vector
particle swarm optimization for wireless sensor network localization, Appl. Soft Comput. 65 (2018) 101–120.
[24] ] S. Zidi, T. Moulahi, B. Alaya, Fault detection in wireless sensor networks through SVM classifier, IEEE Sens. J. 18 (1)
(2018) 340–347.
[25] M. Jiang, J. Luo, D. Jiang, J. Xiong, H. Song, J. Shen, A cuckoo search-support vector machine model for predicting
dynamic measurement errors of sensors, IEEE Access 4 (2016) 5030–5037.
12
[26] Chaoyun Zhang, Paul Patras, and Hamed Haddadi, “Deep Learning in Mobile and Wireless Networking: A Survey”, IEEE
COMMUNICATIONS SURVEYS & TUTORIALS, arXiv: 1803.04311v2, 17 Sep 2018, pp. 1-53.
[27] Qian Mao, Fei Hu, and Qi Hao, “Deep Learning for Intelligent Wireless Networks: A Comprehensive Survey”, IEEE
COMMUNICATIONS SURVEYS & TUTORIALS, VOL. 20, NO. 4, 4th QUARTER 2018, pp. 2595-2621.
[28] Raut, AR, Khandait, SP. An intelligent MAC protocol for mission critical applications in wireless sensor networks.
Concurrency Computat Pract Exper. 2023; 35(23):e7813. doi: 10.1002/cpe.7813
[29] S. Hammoudi, A. Bentaleb, S. Harous and Z. Aliouat, "Scheduling in IEEE 802.15.4e Time Slotted Channel Hopping: A
Survey," 2020 11th IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON),
New York, NY, USA, 2020, pp. 0331-0336, doi: 10.1109/UEMCON51285.2020.9298043.
[30] Musaddiq, A., Zikria, Y.B., Zulqarnain et al. Routing protocol for Low-Power and Lossy Networks for heterogeneous traffic
network. J Wireless Com Network 2020, 21 (2020). https://doi.org/10.1186/s13638-020-1645-4.
[31] S. K. Swain and P. K. Nanda, "Adaptive Queue Management and Traffic Class Priority Based Fairness Rate Control in
Wireless Sensor Networks," in IEEE Access, vol. 9, pp. 112607-112623, 2021, doi: 10.1109/ACCESS.2021.3102033.
[32]WOLF-BASTIAN P¨OTTNER and HANS SEIDEL, “Constructing Schedules for Time-Critical Data Delivery in Wireless
Sensor Networks”, ACM Transactions on Sensor Networks, Vol. V, No. N, Article A, January 2016.
[33]BM.Mohammad, S.Abd Kader,H A.Konber, “Designing a Channel Access Mechanism for Wireless Sensor Network
Moustafa”, Hindawi Wireless Communications and Mobile Computing Volume 2017, Article ID 7493269, 31 pages
https://doi.org/10.1155/2017/7493269.
[34]ViktorRichert,B,NaumanIsrar, “Implementation of a Modified Wireless Sensor Network MAC Protocol for Critical
Environments”, Hindawi Wireless Communications and Mobile Computing Volume 2017, Article ID 2801204, 23 pages
https://doi.org/10.1155/2017/2801204.
[35]Ananda Kumar Subramanian, Ilango Paramasivam, “PRIN: A Priority-Based Energy Efficient MAC Protocol for Wireless
Sensor Networks Varying the Sample Inter-Arrival Time”, DOI 10.1007/s11277-016-3581-5, Springer Science Business
Media, New York 2016.
[36] Vijayakumar S., Rizwan P., Mohammad S. Khan, Suresh Kallam, “Reliable and energy‐efficient emergency transmission
in wireless sensor networks ”, Internet Technology Letters, Wiley Online Library, 25 January 2019
https://doi.org/10.1002/itl2.91.
[37] Mu Qiao, H Zhao, S Huang, Li Zhou and Shan Wang, “An Intelligent MAC Protocol Selection Method based on Machine
Learning in Wireless Sensor Networks”, KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL.
12, NO. 11, Nov. 2018
