Optimized Low-powered Wide Area Network within Internet of Things

The Internet of Things is rapidly becoming an integral part of everyday life. Low-powered wide area network are capable of providing reliable connectivity even in low-density areas and device consuming low amount of energy. The exponential increase in the use of Internet of things application across the globe will continue to generate more and more data traffic within the Internet of Things network. Thus, End device within Internet of Things network in the near future will rise up to billions of device operating across Internet of Things network, generating a necessity for more correct and reliable energy conservation Internet of things technology. This prompted the research work on optimized low-powered wide area network. The experiment has been carried out in various stages: firstly running of simulation over wireless sensor network without optimization using MATLAB Simulink and obtained the following result of


Introduction Background to the Study
The Internet of Things (IoT) is a transformative technology designed to establish connections among diverse devices through Radio Frequency Identification (RFID) tags, sensors, actuators, mobile phones, and other means, enabling them to collaborate in achieving specific tasks.This technological paradigm encompasses three key parameters: Things orientation, referring to objects within the IoT context; Internet orientation, facilitating communication across various nodes; and semantic orientation, regulating data traffic among communicating devices within IoT networks (Arslan et al., 2019).
The Internet of Things represents an innovative approach that leverages the power of the internet to facilitate communication between tangible items or entities.These tangible entities encompass a broad spectrum, including commercial equipment and household appliances.Through the integration of sensors and communication networks, these entities generate valuable data, paving the way for a diverse range of services for individuals.For example, strategically managing building energy consumption can lead to significant reductions in energy costs (Motlag et al., 2020).
The majority of IoT-based systems consist of battery-operated devices such as smart sensors, radio frequency identification (RFID) tags, home appliances, surveillance cameras, smartphones, and various other items.However, the rapid proliferation of these devices has given rise to challenges related to energy efficiency, attracting considerable attention in the process (Adul-Qawy et al., 2020).
Researches indicates varying levels of performance in terms of energy power consumption, range, coverage, and latency within the LPWAN sector.This technology is marked by low-power operating devices, cost-effective network components, and broad coverage.Various protocol versions operate under the LPWAN environment to fulfill the domain characteristics across unlicensed free bands.Notably, LoRaWAN stands out as one of the leading protocols in LPWAN technologies, competing for the establishment of large-scale IoT networks.LoRaWAN is anticipated to play a crucial role in connecting billions of devices in the future (Ramli et al., 2021).
The widespread adoption of LoRaWAN technology can be attributed to its strong backing by the LoRa Alliance, a non- profit association comprising over 500 member companies.This work places emphasis on LoRaWAN due to its popularity in the IoT landscape.In the LoRaWAN network's star-of-stars architecture, messages traverse between end devices (EDs) and the central network server (NS) via gateways (GWs).These gateways establish connections with the network server through conventional IP connections, acting as transparent bridges by converting RF transmissions to IP packets and vice versa.LoRaWAN devices utilize the Pure ALOHA media access control (MAC) protocol for communication channel access.Through ALOHA, devices can broadcast packets directly to the communication channel without the need for carrier or collision detection.The simplicity of this protocol contributes to reduced transmission costs, overhead, and, significantly, the power consumption of transmitting devices (Ramli et al., 2021).
When evaluating four wireless technologies, namely: WiFi, BLE, Zigbee, and LoRaWAN, it becomes evident that WiFi consumes the highest amount of power, making it unsuitable for applications reliant on battery power (Sadowski and Spachos, 2018).In contrast, more suitable options for energy-efficient systems, especially in the energy sector, include low-power wide area network communication technologies such as LoRa, Sigfox, and LTE M.These technologies operate in unlicensed bands, offering reliable, cost-effective, low-power, long-range, and last-mile solutions for smart energy management (Motlagh et al., 2020).
IoT applications, spanning areas such as transportation and logistics, utilities, smart cities, smart buildings, consumer electronics, industry, environment, and agriculture.With increasing demand and the vast potential within the IoT ecosystem, LPWANs are anticipated to witness more advanced applications in the future, expanding into domains like remote healthcare, traffic safety and control, smart grid management, complex industrial processes, as well as applications in manufacturing, training, and surgery.Consequently, the increasing demands in these diverse domains underscore the need for LPWANs to exhibit high reliability, availability, and low latency.The batteries of End Devices (EDs), such as sensors, cannot be easily charged or replaced after prolonged use, impacting the network's operational lifespan and hindering its ability to accomplish tasks within the expected time frame, especially given the generated data traffic (Boulogeorgos et al., 2016).
The anticipated rapid increase in EDs within IoT networks in the future, reaching into the billions across public, industrial, and personal networks, emphasizes the need for more accurate and reliable energy conservation technology.Addressing this challenge, the research proposes a technical solution using the particle swarm optimization algorithm to appropriately assign spreading factors to various end devices at the gateway.This optimization aims to facilitate efficient data traffic across the network, ultimately reducing the power consumption of end devices.The remaining parts of the research will cover related work, Frame work, IoT network simulation, results and discussion.The following are the contributions of this study to the body of knowledge: 1. Optimized LoRaWAN protocol by spreading factor allocation using PSO algorithm appropriately 2. Minimized power consumption of end-device over Low Power Wide Area Network.
3. Perform an evaluation of the proposed model using the following metric: throughput, number of nodes, packet lost, data transmission and buffer size.Pau (2015) carried out research on power consumption reduction for wireless sensor networks using a fuzzy approach such as Fuzzy Logic Controller (FLC), which can be used in such a way as to further reduce the energy consumption in a WSN.In fact, the fuzzy-based approach presented in their work dynamically changes the sleeping time in order to increase the battery duration of the sensor devices.Simulation results demonstrate that using the proposed FLC, a substantial reduction in energy consumption is obtained compared to simulations carried out with fixed sleeping times.
The results have been obtained using the Gaussian membership functions because an improvement of 30% has been achieved compared to the approach without FLC.techniques and asymmetric cryptographic techniques for the fog network.For the authentication between gadget and the fog light weight encryption algorithms are employed and asymmetric procedures are implemented for the identification system between the fog and the cloud server.For future work, there is a need for more modification to be made to use this method in various contexts such as BLE or Zigbee.Lavric and Popa (2018), carried out an experiment on Performance Evaluation of LoRaWAN Communication Scalability in Large-Scale Wireless Sensor Networks.The authors specifically find out the number of nodes that can be handled by a given gateway so as to reduce Collison.Lavric and co. (2018), from their obtained result, they conclude that the configuration with the lowest transfer rate ensures the highest level of collisions.Because the airtime of the packet is of a few seconds, it is prone to collisions that can negatively influence the capacity of the communication channel.Based on their evaluation results on collision reduction over Lora network they found out that collision will be minimize using ADR algorithms and limiting the maximum number of transmission packets over the network.Raj and Basar (2019) designed gateway that enables the connection establishment using the simple fuzzy rule basedneural networks clubbed with the clustering to have a proper exploitation of energy evading the wastage caused during the selection of the head and the patterning of the cluster.The learned system enables the cluster patterning and the routing to make optimized choices on the selection of the nodes for conveying and affords a delay less, energy aware and less failure proposed system for transmission.The performance analysis is done with regard to the packet delivery ratio, energy consumption, sensor network life time and delay to evidence it perfect functioning.According to the authors performance evaluation of the proposed system compared to the preceding methods prove that the proposed system is highly adept.Followed by adaption of power parameters from the related literature to be used in the research experiment, formulation of analytical model is a level where power consumption model is derived from the related literature, spreading factor is a level where spreading factor is allocated to end device across simulated wireless sensor network at the gate, MATLAB is platform where wireless sensor network is simulated, power consumption model is evaluated and results were obtained.

IoT Architecture
Based on Design considerations and network architectures for low-power wide-area networks of Chaudhari and Borkar (2020), IoT architecture has four layers which are: LPWAN nodes, LPWAN Gateways, network server and application server as depicted in Figure 2 below.The parameters for the implementation of the proposed model comprise both generic and specific parameters.It was adopted to address the pressing challenge of connecting and networking different computer systems due to the rapid development of information technology (Li et. al., 2015).However, IoT has its own network architecture, which corresponds to each layer of the OSI reference model.Generic parameters were considered part of the design based on the perception of Li et al., (2015) that there is a conventional standard for heterogeneous devices to participate in a network in general called the Open System Interconnection (OSI) reference model.While specific parameters were adopted from the journals of Kalifeh et al. (2021) and Kim et. al. (2020), which are carrier frequency, transmission power, spreading factor (SF), bandwidth (BW), and code rate,

Data
The default spreading factor runs from 7,12 provided the data that we used.LoRa employs multiple orthogonal spreading factors ranging from 7 to 12. SF provides a tradeoff between data rate and range.The choice of a higher spreading factor can increase the range but decrease the data rate, and vice versa.Each symbol is spread by a spreading code of length 2 SF chips.At the transmitter, the spreading code is subdivided into codes of length 2 SF /SF.Then each bit of symbol is spread using the subcode, so it takes 2 SF chips for the spreading of one symbol (SF bits × 2 SF = SF).The substitution of one symbol for multiple chips of information means that the spreading factor has a direct influence on the effective data rate.At the receiver, the spreading code is multiplied by the received bits to regenerate the input data.(Noreen et al., 2017).

Duty cycle 1%
Procedures for Wireless Sensor Network Simulation The processes for wireless sensor network simulation can be described as follows: the network will consist of 50 nodes and Area of interest (AoI) of 20m by 30m is used as justified in the journal of Al-Fuhaidi et al., (2020) using a small number of nodes and small AoI for wireless sensor network simulation gives better performance in terms of coverage, cost and accuracy.MALAB software is used for the simulation of wireless sensor networks using MATLAB Simulink as well as the implementation of spreading factor allocation across the wireless sensor network using the particle swarm optimization algorithm.Spreading factor

Spreading factor (SF)
Proposed Energy Model The first approach consists of active mode, that is, the duration during which the sensor nodes are active, and sleep mode.All the gadgets are powered by 3.3V, with the exception of the sensor unit, which requires a power supply of 2V to operate.The quantity of power the sensor uses could change as a result of the energy used in this mode.Thus, it is needed to take into account the sleep mode consumption.The total energy used by the end node is given by equation ( 1).
Where E Total is total energy used by the end nodes within the wireless sensor network, E Sleep is the energy used when the end node is in sleep mode, and E Active represent energy used when the end node is in on state therefore E Sleep is given by equation ( 2): P Sleep represent the sleep state, and T Sleep represents the time taken in sleep state.E Active is given by equation (3.3): E SU is used energy for the sensor startup, E m represent data measurement, E proc present microcontroller processing, E WUT is the wakeup of the LoRa transceiver, E Tr is energy used by the transmission mode, and E R is the energy used at the reception mode of the packets in the network.The model above depicts the total energy consumption of a wireless sensor network (Bouguera et al., 2018).4. is the experimental result obtained from the phase 1 simulation experiment, where experiment 1 network power consumption indicates the running of simulation across the wireless sensor network without optimization, maintaining the same delay of 1.2574 ms, packet loss of 0.54994 kbps, and throughput of 0.4004 bit/sec throughout the first experiment.
After the simulation without optimization, we obtained a power consumption of 6.399 e-17 joules.
Experiment 2 network power consumption from Table 4. was obtained after running the simulation again without optimization, maintaining the same throughput, packet loss, and delay as the first experiment.The result of power consumption obtained after the simulation is 2.5230e-17 joules.Finally, the percentage was calculated by subtracting the power consumption of experiment 2 from experiment 1, dividing the difference by experiment 1, and multiplying by 100%.(2018) was carried out with respect to the received window (RX 1 ) from the Things Network (TTN) server.Their effort includes testing the power consumption of both RX 1 and RX2, where the result proved that the power consumption of RX 1 is reduced by 25% as compared to RX 2 .Baumker et al., (2019) found that the result of their experiment lowered the power consumption by 50% using the new generation configuration of the transmitter SX1262.The proposed work used the spreading factor allocation technique to lower power consumption by 60%.Based on the figure above, the proposed work outperformed the existing work.The graph above presents the performance between LPWAN protocols; the proposed protocol used LPWAN protocols in relation to the number of nodes and throughput.Figure 7 shows that with a decrease in the number of nodes, there is an increase in the throughput of the proposed protocol.Based on an evaluation of the energy consumption of LPWAN technologies by Rajab et al. (2021), it was found that DASH7 and LoRaWAN protocols are more power efficient than Sigfox.However, the experiment above proved that the LoRaWAN protocol is the most energy-efficient as compared to the DASH7 and Sigfox protocols.The graph above depicts the performance of the LoRaWAN protocol with respect to delivery delay and number of nodes.
The results indicate that with an increase in network density, there is an associated increase in delay in delivery, which corresponds to the delivery delay of other routing protocols, as stated in the journal of Khan et al. (2013).The authors revealed that with an increase in the number of nodes, there is also an increase in delay.Based on Figure 9, the proposed technique and other techniques were tested with respect to the number of nodes and throughput.Figure 9 shows that with a decrease in the number of nodes (density), there is an increase in throughput, and vice versa.The proposed system has maximized throughput to minimize the network density so as to reduce the packet loss of the proposed model.As stated in the journal].

Results and Discussion
The experiment on battery power consumption was carried out in various stages, as follows: The experiment was implemented in a MATLAB environment; a wireless sensor network was simulated randomly using a random function in 2. Using MATLAB simulink for wireless sensor network simulation Lack support for accurate physical modeling which will reduce the level of accuracy of the result.
3. Lack scalability: MATLAB Simulink may not handle large-scale wireless sensor network simulation scenarios efficiently.

Conclusion
The Internet of Things is a connection of other objects that communicate with each other without human intervention via radio frequency identification, actuators, etc. to carry out a task.The tremendous increase in end devices on the Lora network led to more power consumption due to data traffic at the gateway, thus triggering the research work.The authors used the spreading factor allocation technique so as to reduce the data traffic at the gateway, thus minimizing the power consumption of the end device.The researchers carried out the experiment by simulating an IoT network and allocating spreading factors to heterogeneous devices across the wireless sensor network in a MATLAB environment.A particle swarm optimization algorithm was used as an optimization technique that reduced data traffic so as to minimize power consumption across the LoRa network.Thus, the experiment reduced power consumption by 60%.The researchers state that simulators like Arduino and ThingSpeak should be used for online implementation of spreading factor allocation over an IoT cloud platform using PSO and compare the results.
Qeios, CC-BY 4.0 • Article, May 27, 2024 Qeios ID: FZQVZY.2 • https://doi.org/10.32388/FZQVZY.2 4/22 Abdul-Qawey (2020) carried out a study on the classification of energy-saving techniques for optimum utilization across the IoT heterogeneous wireless sensor network.The co-authors work toward energy-saving techniques within the IoT domain for proper network management with respect to efficient power consumption.Abdul-Qawey et al. (2020) mentioned in their work that realizing energy efficiency in heterogeneous wireless sensor devices becomes a key challenge and extends the system's life.The researchers considered the current operating conditions of network devices, such as active, idle, and sleeping states.According to the authors, upcoming developments will focus on the design and implementation parts of the IoT.According to Ismail et al. (2019), LPWAN (Low-Power Wide-Area Network) emerged as an enabling technology for Internet of Things applications.The authors conduct a study on low-power wide-area networks with respect to opportunities, challenges, and future directions within the IoT domain.Ismail and colleagues discussed the characteristics of current low-power wide-area network (LPWAN) technologies, which enable them to provide long-range connectivity, low-power communication, and low deployment costs for a large number of devices.The researchers mentioned in their journal that, due to the increase in demand within the IoT network, several competing technologies are being developed, including LoRa, SigFox, IQRF, RPMA, DASH7, and Weightless.The authors have stated future research in LPWAN as follows: future scalability and coverage, technology coexistence, inter-technology communication, real-time communication, support for control applications, support for mobility, support for data rate, and security.Bouguera et al. (2018) presented an improved energy model for sensor nodes.Various LoRaWAN modes and scenarios based on LoRaWAN Class A were examined for a particular Internet of Things application.The authors also looked into LoRaWAN scenarios to figure out how much power the sensor node would consume.The created model enables analysis of how choices in hardware and software affect the autonomous behavior of the node under study.The authors demonstrated through numerical results that the amount of energy consumed varies depending on the LoRa/LoRaWAN parameters used, which include spreading factor, coding rate, payload size, and bandwidth, in order to reduce the energy consumption of the sensor node.Our research adopted some of these parameters for the formulation of the power consumption model.Zanaj and colleagues (2021) carried out a survey and analysis of short-and wide-area IoT technologies to ascertain the energy efficiency and consumption characteristics of LPSAN and LPWAN technologies.The characteristics of the most important Internet of Things technologies-low-power wide-area networks (LPWAN) and low-power short-area networks (LPSAN)-have been thoroughly studied by the writers.In the end, the most often used methods for increasing energy efficiency could be categorized into multiple broad categories: methods for data reduction, radio optimization, sleep/wake, energy-efficient routing, and method evaluation across multiple technologies were all taken into consideration.Among Low Power Wide Area Network (LPWAN) standards, LoRaWAN has become a prominent protocol, as they have discussed in their journal.Motlagh et al. (2020), conduct an investigation into the existing literature on the application of the Internet of Things in energy systems in general, and smart grids in particular.Many IoT techniques are studied by the researchers.The paper also examines some of the challenges associated with implementing IoT in the energy sector, such as privacy and Qeios, CC-BY 4.0 • Article, May 27, 2024 Qeios ID: FZQVZY.2 • https://doi.org/10.32388/FZQVZY.2 5/22 security, and some of the potential solutions, such as block chain technology.The authors worked on the application of the Internet of Things in the energy sector and consider the Internet of Things to be a platform that promotes energy conservation, specifically efficient energy use, through the use of various LPWAN protocols such as Lora, Sigfox, and others, thus it formed basis of using Lora protocol in our study.Lansky et al. (2022), carried out a study on light weight centralized authentication mechanism for the internet of things driven by fog.The authors developed light weight centralized authentication mechanism for IoT using cryptographic

Figure 1 .
Figure 1.Framework for the Proposed Model

Figure 8 .
Figure 8. Delay Delivery with Respect to No. of Nodes

Figure 10 .
Figure 10.Protocols' Energy Consumption with Respect to Data Transmission & Buffer Size MATLAB across the measured area of 20 m by 30 m.Following the simulation of an IoT wireless sensor network, an optimization technique called particle swamp optimization was implemented by calling the PSO algorithms in the MATLAB library for the assignment of spreading factors to the end device at the LoRa getaway.Some specific key parameters, like data rate, spreading factor, throughput, network density, delay, and packet loss, were used as metrics for the validation of the model.The results of the experiment were obtained as follows: Experiment scenario 1 spreading factor was implemented without optimization across the wireless sensor network, as shown in figure4.Delay, packet loss, and throughput were maintained at the same level throughout the experiment.Experiment scenario 2 was obtained after implementing PSO for SF allocation across the wireless sensor network, as shown in figure5.Finally, the percentage of the reduced power consumption of the network was calculated by subtracting the power consumption of experiment 2 Qeios, CC-BY 4.0 • Article, May 27, 2024 Qeios ID: FZQVZY.2 • https://doi.org/10.32388/FZQVZY.2 19/22 from experiment 1 and dividing the difference by experiment 1 and multiplying by 100.Now, the total power consumption across the LoRa network was minimized by 60% as compared to the existing work of Krome et al. (2019), which uses the receive window (R X ) technique to reduce energy consumption by 25%, and the journal of Baumker et al. (2018), which uses the transmitter configuration method to reduce power consumption by 50%, as depicted in figure 6.Based on Figure 6, the proposed work outperforms the `existing work of Krome et al. (2019) and Baumker et al. (2018).The results of the proposed work indicate that using the Particle Swarm Optimization (PSO) algorithm for appropriate spreading factor assignment across the IoT network leads to lower energy consumption.This demonstrates the effectiveness of the PSO algorithm in finding an optimal spreading factor configuration that minimizes energy consumption.Based on the results of the proposed system, using the PSO in allocating spreading factors in LoRaWAN, the network can achieve a more efficient allocation of resources, resulting in reduced energy consumption.Different graphical models were tested with respect to various network communication parameters such as network density, throughput, buffer size, data transmission, packet loss, and network delay as another form of evaluating the proposed work against the existing work of the recently reviewed literature.Limitations of the study 1.The study considers only device class A while class B and C are not captured for the wireless sensor network simulation.
research investigates a multitude of prior studies concerning low-power wide area networks (LPWANs), focusing on topics such as the energy efficiency and power consumption of Internet of Things (IoT) devices.The examination The

Table 1 .
Summary of techniques mentioned.Formulation of Meta-heuristic AlgorithmFail to account for large size network.
Kim et al. (2020)• Article, May 27, 2024 Qeios ID: FZQVZY.2 • https://doi.org/10.32388/FZQVZY.2 6/22The research framework was developed in alignment with the stated objective of the research project.The proposed methodology outlines the implementation of spreading factor allocation to various end nodes, utilizing the network simulator MATLAB to simulate the IoT network based on the LoRaWAN protocol.The spreading factor allocation is designed to be applied across the IoT network to optimize its performance.The parameters utilized in this framework were adopted from the studies conducted byKim et al. (2020)andKalifeh et al. (2021).The visual representation of the research framework is depicted in the figure below.Qeios, CC-BY 4.0 • Article, May 27, 2024 Qeios ID: FZQVZY.2 • https://doi.org/10.32388/FZQVZY.2 7/22