1. Introduction
Diverse unforeseeable disasters such as tsunamis, floods, earthquakes, and torrential rainfall strike many people worldwide. Although disasters are different, disaster management problems are almost identical and repeated, mainly linked to the field complexity, the interoperability problems, and the socio-cultural components. Therefore, communications between functional and dysfunctional areas are significant issues to consider to save people’s lives. Practical communication techniques between first responders and victims rely on the efficient mission-critical transfer of voices and information between victims and first responders [
1]. In such circumstances, there are limited resources and services, low reliability and network availability, energy/power loss, and no available communication infrastructures, restricting the implementation of information communications [
2].
One of the most commonly shared features of all disasters is the failure of critical communications [
3]. The failure of telecommunications infrastructure, whether partial or complete, leads to an inevitable loss of life by causing the delay of disaster relief emergency response. In many scenarios, the breakdown of power networks causes communication networks to disconnect. However, batteries power the communication devices in the disaster region, and their power may eventually run out. Therefore, energy consumption is a significant concern in public safety networks (PSNs). Thus, a PSN needs to implement fewer energy-consuming networks and have the ability to harvest energy.
Energy harvesting (EH) can supply communication devices and wireless networks with the energy harvested. Recently, the EH has become an appealing solution to prolong the wireless networks’ lifetimes. Renewable energy sources such as solar and wind power may provide EH with unlimited environmental energy. Therefore, research in renewable energy has received much attention, notably in cellular communication [
4]. Since the robustness and availability of energy are far more critical in a disaster area than in typical situations, using solar panels and wind turbines is not guaranteed to generate the required energy inside the affected area to support the established wireless network. Therefore, ambient radio signals may be potentially viable wireless energy harvesting (WEH) resources, with RF devices converting received signals into energy sources [
5,
6]. This method is a typical choice for energy-constrained wireless networks. Since emergency energy is mainly restricted, RF energy can be a good solution for power and information transfer.
Researchers have been motivated to examine the simultaneous wireless information and power transfer (SWIPT) and its usefulness in disaster communication methods as a promising technology. SWIPT was developed and allowed energy and data to receivers equipped with RF energy harvesting circuitry. SWIPT can provide continuous and suitable energy needs to wireless networks. In a communications system with SWIPT capability, power transfer and information are simultaneous across wireless media. Consequently, power splitting (PS) and time switching (TS) were proposed as two practical SWIPT strategies in reference [
7]. Moreover, the transmission rate optimization problem for a dual-hop multi-relay IoT system with a decode-and-forward (DF) relay supporting the SWIPT technique was investigated by Lu et al. [
8].
The essential aspect of our proposed technique is to provide wireless coverage to the disaster area when wireless networks based on the D2D communication technique or unmanned aerial vehicles (UAVs) technique are ineffective due to the considerable distance between the formed clusters or the disaster area size beyond the UAV coverage area limitations. Thus, this paper’s contribution expands the previous work, where the RS-assisted wireless communication network is aided with an energy harvesting strategy to ensure the stability of the network. Besides, it utilized the TS protocol at the main cluster heads (MCHs) that act as a relay node of the cluster to transmit all UE information to BS. Moreover, we enabled reliable connectivity for the RS to MCH and D2D among the clusters communication ranges and calculated the MCH and MCM/SCH power consumptions to ensure that the energy harvested is greater than or equal to at least the power consumption. The system model is expected to perform better using an energy harvesting strategy and become more efficient for disaster-resilient operations.
The rest of this paper is organized as follows:
Section 2 summarizes the related work.
Section 3 provides a system model for analyzing the energy harvesting technique and D2D performance in the disaster scenario regarding outage probability.
Section 4 discusses a disaster recovery framework based on D2D, clustering, and EH. The simulation results and performance of EH and D2D with clustering are detailed in
Section 5, while
Section 6 concludes the paper.
2. Related Work
In critical events such as disasters, the primary intention is to search for and rescue victims. Thus, the importance of communication networks arises to serve in such cases successfully. According to PSN standards, wireless networks are an appropriate alternative for post-disaster relief operations since they are simple to implement in emergencies and do not require pre-existing infrastructure. In the literature, several strategies tailored to various scenarios have been suggested.
In reference [
9], the authors proposed using a drone–femtocell technology and constructing an algorithm capable of locating any mobile terminal in a particular monitoring region to search for and identify missing individuals in natural disaster circumstances. This technique uses a series of power measurements based on the reference signal received power (RSRP) to classify the terminal inside or outside the monitoring region. Consequently, even in the presence of obstacles that cause the radio signal’s propagation to be non-isotropic, it roughly determines the position with 1 m accuracy.
In reference [
10], the concept of a movable and deployable resource unit (MDRU) was developed by Nippon Telegraph and Telephone (NTT) Corp. The MDRU’s concept is to deploy an entire resource unit to establish a recovery network to the disaster site.
Altay et al. [
11] suggested a stand-alone eNode-B architecture that secures service without a backhaul connection by leveraging its own integrated virtual evolving packet core (EPC). The stand-alone eNode-Bs are devised to build backhaul connections, expanding the coverage without a central EPC structure. The stand-alone eNode-B architecture provides improved interoperability and enhances data transmission functionality, particularly in emergencies and disaster events. The work in reference [
11] nevertheless did not handle the power consumption problem in the event of a disaster.
Castellanos et al. in reference [
12] suggested a capacity-deployment tool for designing and evaluating the backhaul network for UAV-assisted networks in disasters. This tool assigns resources to the ground users and the backhaul network simultaneously, taking into account power limitations and backhaul capacity. They investigated three different backhaul scenarios using 3.5 GHz with carrier aggregation (CA), a 3.5 GHz link, and the 60 GHz band with three different types of drones.
In reference [
13], the impact of the relay mobility was handled. The authors evaluated the mobile relay’s capacity and coverage extensions and the impact of mobility on the expected availability duration and route probability establishment. However, the scope of this work is limited to a point-to-point connection with single cells in an assumed idealized circular region. In contrast, we deal with the communication links throughout the multi-hop network in our work, where coverage extends from RS-assisted wireless communication links with clustering techniques to MCHs/SCHs and MCMs/SCMs through D2D links.
In reference [
14], a multi-path routing system for PSNs supported by reinforcement learning (RL) and UAV was proposed. The goal is to improve the PSN’s energy efficiency (EE) and thereby increase network lifetime. To begin with, different clustering algorithms are used to generate network configurations. The RL is then used to design a routing topology that considers both the transmission path’s immediate energy cost and its total distance cost.
Since D2D communications allow close UEs to communicate without a base station, this approach may ensure high-speed data transmission and stable, continuous real-time communications. Therefore, In reference [
15], the authors suggested a D2D multicast emergency communications technique to make PSN more flexible. This technique is divided into three steps. Firstly, the distance between UEs is used to divide the alternate cluster head. Secondly, there are two types of cluster head selection schemes. One is based on the number of extended UEs, while the other is based on terminal power. Finally, the Hungarian algorithm based on throughput awareness is used for channel multiplexing.
D2D communications and an unmanned aerial vehicles (UAV) approach assisted by D2D links and clustering techniques as an underlay to recover cellular networks in disasters recently attracted much attention. These technologies can help improve energy management, which is a significant concern. Therefore, some recent studies have investigated the possibility of harvesting energy via the RF signals in a cooperative wireless network.
To develop a highly efficient UAV-based wirelessly powered communication network (U-WPCN), the service area for the UAV platform should be selected so that the UAV platform hovers at the appropriate location depending on the positions of the group’s ground terminals, which will be enhanced if the U-WPCN is efficient. Therefore, in reference [
16], the authors introduced two networking strategies for maximizing communication performance and improving networking efficiency. The economic strategy uses fewer UAVs while maintaining the required data rate threshold, and the performance strategy uses a higher number of UAVs to enhance the quality of communication (at least one U-IP and one U-EP). They used a greedy algorithm to discover the best hovering point for the area so that the location of the UAV in the network could be determined as an alternative to the conventional local search techniques, such as clustering and genetic algorithms.
In reference [
17], researchers proposed a wireless communication system that uses UAV-powered energy harvesting to improve network connectivity and transfer energy during a natural disaster. Furthermore, in reference [
18], RF energy harvesting-based power allocation systems were proposed. Researchers investigated a UAV equipped with a pico base station scenario, which might minimize network congestion and traffic overload while enhancing wireless coverage. They used various clustering techniques to overcome energy harvesting challenges contributing to power supply constraints.
In reference [
19], a SWIPT approach is suggested to improve energy efficiency (EE) performance and to use radio frequency (RF) signals to harvest energy while functioning with limited battery capacity. A stable matching EH technique is utilized to overcome the problem of resource allocation under the reuse of the spectrum. However, this study does not enhance the EE for D2D communications and cellular networks by improving the CH selection, power transfer, and power splitting ratio.
To address the energy performance constraint, the authors in reference [
20] investigated D2D communications based on energy harvesting to maximize the energy efficiency based on the transmit power control and time slot allocation. Thus, a practical resource distribution based on D2D energy harvesting (D2D-EH) was used to improve channel connection quality and reduce the probability of communication outages after disasters.
An integrated strategy for optimal energy harvesting between functional and dysfunctional areas (UAV, CH, and D2D communications) was used. In reference [
21,
22], UAVs having multiple antennas function as relay nodes to transfer power and transmit information to the UEs located outside the network coverage area. D2D communication within the cluster utilizes an unlicensed spectrum to enhance the system spectrum efficiency for communication between CH and CMs [
23]. However, it is challenging to utilize CHs to re-transmit the UAV’s wireless signal to the UEs within its cluster during disasters.
The energy harvesting strategy introduced in this paper could support keeping the wireless network operating during disasters by increasing the UE’s battery life. Thus, the improved clustering technique (CFT) with D2D communication can sustain communication when cellular infrastructure becomes partially or fully dysfunctional. In other words, in a situation where the D2D communications model or the UAV deployment model, which is assisted by the clustering technique to recover cellular networks, could not be utilized due to distance limitations. The proposed framework can be considered optimal when UEs are distributed widely in the disaster area.
5. Simulation Results and Discussion
This section presents the simulation results of the proposed framework, which considers the disaster scenario illustrated in
Figure 1 to demonstrate its effectiveness. The simulation assumes that the operational BS transmits its signal at maximum power to extend signal coverage to the disaster area by using the RS as a relay at the edge of its coverage area. To connect the MCH in the disaster area, the RS transmits its signal power by employing
= 9 W and a bandwidth of
= 10 MHz. Moreover,
Table 1 details the simulation parameters that were used.
Notably, the received signal power at the destination UE is affected by path loss, which impacts the UE’s ability to harvest energy. As a result, it is critical to place the RS in the best possible location to receive good signal power at the MCHs. Consequently, we will discuss the optimal location of the RS, the wireless network efficiency in harvesting energy effectively, and the signal coverage outage probability.
Figure 5 shows the signal success probability when the RS transmits its signal using differing power values and is received at the furthest distance MCH from the different RS locations. The four candidate locations of the RS are
,
,
, and
, which are based on the greatest SINR achieved from the BS. As can be seen in
Figure 5,
at the location of
achieved appropriate results in terms of received signal strength. Therefore, it is ideal for establishing a wireless communication link with MCHs, transmitting data, and transferring energy.
The outage probability vs. the MCHs distance from the RS is evaluated in
Figure 6 to ensure appropriate power is used to transfer energy from the RS to the MCHs considering the optimal location of the RS
. The investigation considers that the RS transmits its signal using 3 W, 5 W, and 9 W to transfer its energy to MCHs and utilizes 700 MHz as an operational frequency. Thus, the simulation results showed that the clustered UEs’ total outage probability increases as the distance to the RS increases. In the scenario where the RS transmits
= 9 W, the outage probability gradually increases up to
at the most distant cluster, whereas it reached
when the RS transmits
= 3 W, which is to be expected because the more transmitting power, the lower the probability of an outage. Generally, path loss occurs when the signal travels farther from the transmitter, so the outage occurs. The proposed RS-assisted wireless network strategy and the UAV-assisted wireless network approach are examined in the context of the disaster area illustrated in
Figure 4, where 12 clusters have formed, and 12 UEs have been assigned as MCHs. The UAV altitude has been set at 100 m and it is transmitting its signal power using 5 W, which is assumed to be at the center of the disaster area. By assessing the outage probability of establishing links between MCHs using these two approaches, we can ensure the proposed strategy’s efficiency in terms of the energy harvesting outage probability at the 1st hop (RS to MCHs and UAV to MCHs).
Figure 7 assessed the EH outage probability vs. time switching factor
. The simulation findings reveal that as the number of time switching factors increases, the overall EH outage probability decreases. To put it another way, the higher the number of
, the lower the probability of EH failure during the transmission block time
.
The RS model’s EH outage probability performance is significantly superior to the UAV model’s EH outage probability due to the enhanced channel quality associated with the RS model and the high transmitting power employed by the RS to establish links between the RS and MCHs. Furthermore, the UAV interferes with MCH signals, increasing the probability of an EH outage within its range.
The transmission signal power that the RS uses to establish a successful connection with the MCH from its optimal location is investigated. As a result,
Figure 8 illustrates the received signal at MCHs while the RS transmits at various power levels to ensure the quality of service of each cluster. Therefore, according to reference [
31], the MCHs’ reference sensitivity is estimated in dBm based on the BS noise figure (NF), the bandwidth, and SNR required to reach the threshold throughput, which could be estimated as:
The result shown in
Figure 8 indicates that as the distance between the MCHs and RS increases, the signal strength decreases, even though we are including the weakest signal that MCHs could be able to identify and process. Furthermore, when the RS transmits the signal with
= 9 W, the farthest MCH receives a signal power of
= −68 dBm, whereas −74 dBm is received when the RS transmits with 3 W. As a result, it is essential to emphasize that higher received power at the MCH improves system capacity and gains more efficient energy harvesting.
The path loss affects signals transmitted by the RS, even when MCHs are located in line of sight (LoS) with the RS, which has been observed from the strength of the received signal at the MCHs that are shown in
Figure 8, which are affected vastly by the RS location and its signal transmitting power. Therefore, we investigated the impact of the RS’s location on the efficiency of energy harvesting by MCHs.
Figure 9 showed the harvested energy vs. time switching ratio
in the 1st hop (RS to MCHs) when the RS transmitted its signal power from four different locations (
,
,
, and
). According to the simulation result, EH increases as
increases where
. However, in practice,
cannot be set to 1 because it means that no communication data is transmitted. Thus, the RS at the location
is the optimal location where it gives better signal coverage, as shown in
Figure 5, and efficient energy harvesting.
Since the received signal at the MCHs is affected by RS transmission signal power when the large-scale path loss considers the distances between the MCHs and RS while the bandwidth is fixed, so thus the energy harvested at the MCHs is also affected. Therefore, the RS transmission signal power is expected to affect EH performance since higher transmit power is required to compensate for increasing the distances between and more hops between .
Accordingly, we investigated the probability of efficient energy harvesting for MCHs utilizing the RS’s optimal location, where
Figure 10 shows an examination of energy harvesting at MCHs when the RS transmits its signal by different power values at its optimal location. The EH steadily decreases as a function of the MCHs as the distance increases. When
= 9 W, EH decreased from 1.32 to 0.62 joules as the distance increased from 152 to 945 m, while EH decreased from 0.84 to 0.49 joules as the distance increased from 152 to 945 m when
= 3 W, demonstrating that the RS transmission signal power had an impact on the EH since as the distance increased, a higher transmit power would be necessary.
Furthermore, we investigated the amount of energy harvested by MCHs at various time switching ratios
when the RS is in its optimal location and transmits its signal at different power levels, as shown in
Figure 11.
The EH steadily increases as the value increases. However, when = 9 W, EH increases from 0.652 to 1.323 joules as the increases from 0.02 to 0.9. In comparison, EH increases from 0.49 to 0.84 joules as the increases from 0.02 to 0.9 when = 3 W, demonstrating that the RS transmission signal power had an impact on the EH as the time switching ratio increased. Thus, a higher transmit power would be necessary to harvest a more significant amount of energy.
The performance of EH versus energy harvesting efficiency
for RS and UAV communication is simulated in
Figure 12. According to the simulation result, the RS scenario obtained 0.52 joule at
= 0 and gradually increased as the
value increased to achieve 1.49 joule when
= 1, whereas the UAV scenario achieved 0.12 joule and gradually increases as
value increases to achieve 1.18 joule when
= 1. Hence, The RS-assisted wireless link scenario maximizes EH approximately 20.8% better than the UAV link scenario through MCHs.
As a result, the EH performance in the RS scenario is superior to that in the UAV communication situation. The aforementioned is attributed to the slight loss of signal power received at the UEs and the considerable propagation path gain between the RS and MCHs. Further, EH in the UAV communication scenario is lower than the RS scenario due to the size of the disaster area compared to the UAV maximum cover range and the random-distributed UEs at spaces apart. The suggested framework ensures the continuity of wireless signal coverage in the disaster area.
Additionally, the framework utilizes wireless communication links assisted by an CFT and D2D communication, used to reduce UEs’ required transmission power to transfer their information, extend the network coverage area, and improve network spectral efficiency.
The network’s spectral efficiency performance with various MCH densities is shown in
Figure 13. Due to the varying MCH densities and the efficient reuse of radio resources, as the MCHs’ numbers increase, the spectral efficiency increases, affecting network coverage. However, higher MCH densities improved spectral efficiency in the network scenario under investigation. The spectral efficiency increases from 0.14 bps/Hz to 1.1 bps/Hz when the MCHs are increased from 1 to 12 at MCH density
.
Similarly, at MCH densities and , the spectral efficiency improves from 0.15 bps/Hz to 1.38 bps/Hz and from 0.16 bps/Hz to 1.5 bps/Hz, respectively. Thus, a higher spatial density of MCHs, which is based on cluster formation, can serve more while maintaining the same spectral efficiency of the system. Furthermore, unlike the UAV model, the proposed communication system’s performance is unaffected by an increase in the number of clusters since the signal power used for the communication to cover the disaster area is stable and fixed.
Figure 14 illustrated the energy harvesting performance for different time slots in two-hop EH strategies,
and
. According to the simulation results, it is noticeable that the EH in the second-hop connection (within the cluster) is less than in the first-hop connection (RS to MCHs) due to the transmission signal power used by the RS, which was as expected.
Energy harvested versus the energy harvesting efficiency is estimated when the D2D distance between MCMs/SCH is 20, 30, 40, and 50 m. As demonstrated in
Figure 15, generally, EH increases as the EH efficiency increases. Whereas the EH amount is less as the UE’s sparsity distance increases since the received SINR is affected by interference. In other words, due to decreased UE density, as the sparsity distance increases, D2D communication interference affected the received signal power, which affected the EH at the MCMs/SCH. The simulation result shows that when the distance between MCH and MCMs/SCH exceeds 30 m, EH performance is affected as the EH efficiency is set to less than 0.9.