Contents Introduction
Rapid development of wireless communication technology and the dramatical innovation of UAV-based manufacturing technology make the cost of UAV continous decreasing. Therefore, many new UAV applications have recently appeared in the civilian market, including weather monitoring systems, forest fire prevention technology, man-v-machine areas and so on. This widespread application of UAV technology attracts extensive attention to explore the UAV relay assisted IoT communication network [1].
A. Background
Recently, UAV assisted communications play an significant role in 5G/B5G networks, which is expected to achieve throughput transmission above 10Gbps, ultra density device connection and millisecond transmission delay [2], [3]. The stringent requirements of the future communications make the network coverage expand to 3D interconnection so as to overcome the large scale fading for the ultra speed transmission. Moreover, the construction of cellular stations in urban hotspots have excessive cost and can be extremely difficult to carry out. Therefore, the UAV relays assisted communication could provide solutions for the IoT network and have several advantages, such as its convenient deployment and lower cost, high-altitude assisted transmission and so on [4], [5]. It is mentioned that the UAV relay assisted communications can reduce the obstruction of buildings, mountains and other obstacles and obtain the higher line of sight (LoS) transmission effect [6], [7]. However, the UAV is a battery-powered terminal, which has limited battery life and its power consumption has a significant influence on the communication process [3], [8], so the transmission protocols should be designed to maximize the system throughput and minimize the transmitting power, therefore, the UAV relay assisted communication protocols in IoT system enhanced with energy harvesting are proposed and analyzed in this paper detailedly.
Due to the actual UAV circuit system is difficult to achieve energy harvesting and information processing simultaneously [9]–[11]. Therefore, the TS and PS schemes are typically applied for UAV relay assisted communication networks. The former strategy depends on the time slots seperation, one is used for energy harvesting and the other for information processing. The latter strategy depends on the power division, one part is utilized to energy harvesting and the other to information processing [7], [12]. The above two schemes can be implemented to realize energy harvesting and information processing by single antenna at the same time.
B. Related Work
The traditional energy harvesting technologies were mostly solar energy or wind energy. With the development of new technology, RF energy harvesting has received widespread attention from academia and industry. Varshney et al. [13] studied amplify and forward (AF) relay network with energy harvesting function through the joint optimization of power control and the distribution of source, moreover, the average rate maximization of delay constraint under Rayleigh fading channel was investigated in detail. Dong et al. [14] derived the outage probability and system capacity of the energy harvesting relay with AF function under delay limited and delay tolerant transmission. Nasir et al. [15] analyzed a multiple sources to relay and destination node with energy harvesting transmission method. Ding et al. [16] investigated the performance of a two hop model for energy harvesting AF relays to transmit energy and information to two downlink sinks. Ji et al. [17] explored PS schemes for nonlinear energy harvesting AF relay systems with direct path and optimized system to maximize system capacity. Bai et al. [18] proposed a new hybrid EH protocol such as a combination of power splitting and time switching. In order to maximize the system throughput performance, Atapattu and Evans [19] analyzed the outage probability expression of a nonlinear energy harvesting and decoding system in the two-hop transmission relay network under the influence of interference through a multiple parameters joint optimization.
Kumar et al. [20] analyzed the average symbol error rate (SER) of a three nodes decode and forward (DF) relay system for energy harvesting in a Nakagami-
Several researches have been explored for UAV relays assisted communications with RF energy harvesting. Hua et al. [28] analyzed the UAV AF relay networks using BS to implement the simultaneous wireless information and power transfer (SWIPT) and derived the multiple parameters joint optimization. Xie et al. [29] proposed the schemes for UAV uplink transmission energy harvesting systems, where the user adopted the harvested energy to send the information to the UAV. Yin et al. [30] established a typical single source and dual target system model and derived the solution of multiple parameters joint optimization with throughput maximization. Lu et al. [31] proposed an energy constrained UAV communication network protocol based on OFDM relay wireless power transmission. Yin et al. [32] gave several surveys for energy constrained UAV cellular network, which acts as a relay for all users to increase the uplink rate through cooperative communication.
C. Innovation and Structure
The interference received by the terminal may be very common in the future ultra-dense wireless communication deployment of 5G/B5G environments. Therefore, this paper explores the multiple UAV relays assisted communication to complete the information transmission between the source and the IoT nodes enhanced with energy harvesting. In order to improve the performance of UAV relay assisted communication systems as well as reduce the power consumption, this paper proposes two methods of opportunistic UAV relay selection and partial UAV relay selection, meanwhile, TS and PS schemes are also utilized for UAV relay assisted energy harvesting protocols. Moreover, the derivations and analysis present the proposed multi-parameter joint optimization of transmitting power, scaling factor and UAV relay selection could effectively improve the system throughput and reduce the system outage probability and BER.
The main contributions of this paper can be seen as follows:
The energy harvesting and information transmission schemes of UAV relay assisted DF networks is designed in detail together with TS and PS protocols;
The closed-form expressions of system outage probability and BER via the Nagamai-
m The system throughput and delay limited state of UAV relay assisted system are analyzed and the joint optimization of multiple parameters is investigated detailedly.
The structure of this paper can be seen as follows: Section II introduces the system model of UAV relay assisted communication networks; Section III elaborates the PS and TS protocols of UAV energy harvesting; Section IV derives the closed form expressions of system outage probability and BER under aggregated interference environment; The simulations in section V verifies the proposed schemes and proves the correctness of theoretical analysis; Section VI concludes this paper.
System Model
The two-hop energy harvesting UAV relay assisted communication model can be considered as shown in Figure 1. Where the UAV relays are half duplex and DF mode, one of the UAV relays are selected as the optimal relay to implement transmission for terminal IoT nodes. As shown in Figure 1, the system consists of source node S, multiple UAV relay nodes R(denoted as 1,2,…) and terminal node D. Typically, the communication between source and terminal node may be prevented or affected by the obstacles or blockage of urban buildings and mountains, then the UAV relays are required to realize assisted communication and the transmission process is divided into three phases. In the first phase, the source node S transmits signal s to the selected UAV relay node
It is assumed that the signal transmission among each UAV relay link is independent identically distributed, the terminal node D may be suffered aggregate interference caused from multi-layer ultra dense network coverage in 5G/B5G environments. It can be denoted that the power of the
UAV Relay Assisted Communication Protocol Enhanced With Energy Harvesting
The UAV relay assisted communication enhanced with energy harvesting can be designed through TS and PS protocols elaborated detailedly in this section as follows.
A. TS Protocol
The TS protocol divides energy harvesting and information processing into three parts according to time intervals allocation, which can be seen as Figure 2, where T is the total time for energy harvesting and information processing. In the entire time block T, the duration of energy harvesting of the UAV relay node is
The TS protocol of UAV relay time allocation scheme is shown in Figure 3. The source node S send signal s to the selected UAV relay node. Accordingly, the received signal of UAV relay can be expressed as:
\begin{equation*} {y_{sr}} = \sqrt {P_{s}} {h_{j}}s + n_{j}^{r}~\tag{1}\end{equation*}
Therefore, the SNR of the UAV relay can be obtained as:\begin{equation*} {\gamma _{S{R_{j}}}} = \frac {{P_{s}{{\left \|{ {h_{j}} }\right \|}^{2}}}}{\sigma _{R}^{2}}~\tag{2}\end{equation*}
Then the energy harvesting by the UAV relay is:\begin{equation*} Eh = \eta {P_{s}}{\left \|{ {h_{j}} }\right \|^{2}}\alpha T~\tag{3}\end{equation*}
\begin{equation*} {P_{r}} = \frac {Eh}{{\frac {(1 - \alpha)T}{2}}} = \frac {{\eta {P_{s}}{{\left \|{ {h_{j}} }\right \|}^{2}}\alpha T}}{{\frac {(1 - \alpha)T}{2}}}~\tag{4}\end{equation*}
Since the terminal node is affected by the aggregate interference caused from the dense network coverage in the environments, the signal received of terminal node can be indicated as:\begin{equation*} {y_{rd}} = \sqrt {P_{r}} {g_{j}}s' + \sum \nolimits _{i = 1}^{M} {\sqrt {P_{i}} {\beta _{i}}{x_{i}}} + n_{j}^{d}~\tag{5}\end{equation*}
\begin{equation*} {\gamma _{R_{j}D}} = \frac {{P_{r}{{\left \|{ {g_{j}} }\right \|}^{2}}}}{{\sum \nolimits _{i = 1}^{M} {P_{i}{{\left \|{ {\beta _{i}} }\right \|}^{2}} + \sigma _{D}^{2}} }} = \frac {{2\eta {P_{s}}{{\left \|{ {h_{j}} }\right \|}^{2}}\alpha {{\left \|{ {g_{j}} }\right \|}^{2}}}}{(1 - \alpha)(\sigma _{D}^{2} + Id)}~\tag{6}\end{equation*}
It can be noted that the SNR needs to be greater than the threshold in the two-hop transmission. Otherwise, the transmission may be interrupted and the terminal node cannot achieve the correct signal.
B. PS Protocols
Unlike the TS protocol, the PS protocol can be mainly divided into two steps. In the first step, the source node S sends signal to the selected UAV relay node, In this process, the energy is divided into two parts according to the power split mode,
The PS protocol of UAV relay power allocation can be shown as Figure 5. The signal received of UAV relay is:\begin{equation*} {y_{sr}} = (1 - \theta)\sqrt {P_{s}} {h_{j}}s + n_{j}^{r}~\tag{7}\end{equation*}
Therefore, the equation (7) shows that the SNR received by the UAV relay is:\begin{equation*} {\gamma _{S{R_{j}}}} = \frac {{(1 - \theta){P_{s}}{{\left \|{ {h_{j}} }\right \|}^{2}}}}{\sigma _{R}^{2}}~\tag{8}\end{equation*}
Since the energy harvesting time of UAV relay node for PS protocol is \begin{equation*} Eh = \frac {{\eta {P_{s}}{{\left \|{ {h_{j}} }\right \|}^{2}}\theta T}}{2}~\tag{9}\end{equation*}
\begin{equation*} {P_{r}} = \frac {Eh}{T/2} = \eta {P_{s}}{\left \|{ {h_{j}} }\right \|^{2}}\theta ~\tag{10}\end{equation*}
The signal received of the terminal node is:\begin{equation*} {y_{rd}} = \sqrt {P_{r}} {g_{j}}s' + \sum \nolimits _{i = 1}^{M} {\sqrt {P_{i}} {\beta _{i}}{x_{i}}} + n_{j}^{d}~\tag{11}\end{equation*}
So the SINR of the terminal node can be expressed as:\begin{equation*} {\gamma _{R_{j}D}} = \frac {{P_{r}{{\left \|{ {g_{j}} }\right \|}^{2}}}}{{\sum \nolimits _{i = 1}^{M} {P_{i}{{\left \|{ {\beta _{i}} }\right \|}^{2}}} + \sigma _{D}^{2}}} = \frac {{\eta {P_{s}}{{\left \|{ {h_{j}} }\right \|}^{2}}\theta {{\left \|{ {g_{j}} }\right \|}^{2}}}}{\sigma _{D}^{2} + Id}~\tag{12}\end{equation*}
Performance Analysis
This section focuses on the performance analysis of the proposed UAV relay assisted communication protocols and derives the closed-form expressions of the system outage probability and BER.
A. Outage Probability
The optimal UAV relay should be selected from the multiple UAV nodes and the SNR maximization selection criterion can be adopted as follows.\begin{equation*} {SNR_{R_{j}}} = \max ({\gamma _{R_{j}D}})~\tag{13}\end{equation*}
\begin{align*} {P_{out2}}=&\Pr \left \{{ {SN{R_{R_{j}}} < {\gamma _{th}}} }\right \} \\[-2pt]=&\Pr \left({\max \left({\frac {{P_{r}{{\left \|{ {g_{j}} }\right \|}^{2}}}}{{\sum \nolimits _{i = 1}^{M} {P_{i}{{\left \|{ {\beta _{i}} }\right \|}^{2}}} + \sigma _{D}^{2}}}}\right) < {\gamma _{th}}}\right) \\[-2pt]=&\prod \limits _{j = 1}^{N} {\Pr \left({\frac {{P_{r}{{\left \|{ {g_{j}} }\right \|}^{2}}}}{{\sum \nolimits _{i = 1}^{M} {P_{i}{{\left \|{ {\beta _{i}} }\right \|}^{2}}} + \sigma _{D}^{2}}} < {\gamma _{th}}}\right)} \tag{14}\end{align*}
Since the channel \begin{equation*} {f_{P_{r}{\left \|{ g }\right \|^{2}}}}\left ({x }\right) = \frac {{{x^{\alpha - 1}}{e^{ - \frac {x}{\beta }}}}}{\beta ^{a}\Gamma \left ({\alpha }\right)} \tag{15}\end{equation*}
The parameters in equation (15) are as follows:\begin{align*} \Gamma (\alpha)=&\int \limits _{0}^\infty {{t^{\alpha - 1}}{e^{ - t}}} dt \tag{16}\\[-2pt] \alpha=&\frac {1}{{{e^{{{({\sigma _{dB}}/8.686)}^{2}}}} - 1}}~\tag{17}\\[-2pt] \beta=&{P_{d}}\sqrt {\frac {\alpha + 1}{\alpha ^{3}}} ~\tag{18}\end{align*}
Since the noise is negligible compared to interference at high SNR so that the noise variance is assumed to zero [36], [37]. The PDF of the aggregate interference \begin{equation*} {f_{Id}}\left ({y }\right) = \frac {{{y^{M - 1}}{e^{ - \frac {y}{\Omega }}}}}{\Omega ^{M}\Gamma \left ({M }\right)} \tag{19}\end{equation*}
\begin{equation*} \Omega = \frac {{2{P_{ave}}}}{M}{\left({\frac {{\Gamma \left({M + \frac {1}{2}}\right)}}{{\Gamma \left({\frac {1}{2}}\right)}}}\right)^{M}}~\tag{20}\end{equation*}
\begin{align*} f\left ({z }\right)=&\int _{0}^\infty {f\left ({y }\right)} {f_{x}}\left ({{zy} }\right)ydy \\=&\int \limits _{0}^\infty {\frac {{{y^{M - 1}}{e^{ - \frac {y}{\Omega }}}}}{\Omega ^{M}\Gamma \left ({M }\right)} \cdot \frac {{{{\left ({{zy} }\right)}^{\alpha - 1}}{e^{ - \frac {zy}{\beta }}}}}{\beta ^\alpha \Gamma \left ({\alpha }\right)}} ydy \\=&\frac {{{z^{\alpha - 1}} \cdot {{\left ({{\frac {1}{\Omega } + \frac {z}{\beta }} }\right)}^{ - \left ({{M + \alpha } }\right)}} \cdot \Gamma \left ({{M + \alpha } }\right)}}{{\Omega ^{M}\Gamma \left ({M }\right){\beta ^\alpha }\Gamma \left ({\alpha }\right)}} \tag{21}\end{align*}
\begin{align*} {P_{{\mathrm{out2}}}}=&\prod \limits _{j = 1}^{N} {\Pr \left({\frac {{{{\mathop {\mathrm{ P}}\nolimits } _{r}}{{\left \|{ {g_{j}} }\right \|}^{2}}}}{{\sum \nolimits _{i = 1}^{M} {P_{i}{{\left \|{ {\beta _{i}} }\right \|}^{2}}} }} < {\gamma _{th}}}\right)} \\=&{\left [{ {\int \limits _{0}^{{\gamma _{th}}} {\frac {{{z^{\alpha - 1}} \cdot {\Omega ^\alpha }{{\left ({{1 + \frac {\Omega z}{\beta }} }\right)}^{ - \left ({{M + \alpha } }\right)}} \cdot \Gamma \left ({{M + \alpha } }\right)}}{{\Gamma \left ({M }\right){\beta ^\alpha }\Gamma \left ({\alpha }\right)}}} } }\right]^{N}}dz \\=&{\left[{\frac {{\Gamma (M + \alpha){\Omega ^\alpha } \cdot {\eta ^\alpha }}}{{\Gamma \left ({M }\right){\beta ^\alpha }\Gamma \left ({{\alpha \!+\! 1} }\right)}}{}_{2}{F_{1}}\left({M \!+\! \alpha, \alpha; 1 \!+\! \alpha; - \frac {\Omega }{\beta }\eta }\right)}\right]^{N}} \\{}\tag{22}\end{align*}
So the outage probability of the UAV relay assisted communication system can be given as:\begin{equation*} {P_{out}} = 1 - (1 - {P_{out1}})(1 - {P_{out2}}) \tag{23}\end{equation*}
B. BER Analysis
This subsection presents the BER closed-form expression of the proposed UAV relay assisted communication systems as follows:\begin{align*} {P_{{\mathrm{BER}}}}=&E\left \{{ {Q\left ({{\sqrt {v\gamma } } }\right)} }\right \} \\=&\frac {\sqrt {v} }{{2\sqrt {2\pi } }}\int \limits _{0}^\infty {\frac {{{e^{ - \frac {v}{2}x}}}}{\sqrt {x} }} \big [\frac {{\Gamma \left ({{M + \alpha } }\right){\Omega ^\alpha } \cdot {x^\alpha }}}{{\Gamma \left ({M }\right){(\beta)^\alpha }\Gamma \left ({{\alpha + 1} }\right)}} \cdot \\&\times \,{}_{2}{F_{1}}\left({M + \alpha, \alpha; 1 + \alpha; - \frac {\Omega }{\beta }x}\right)\big]^{N} \\=&\frac {\sqrt {v} }{{2\sqrt {2\pi } }}{\left [{ {\frac {{\Gamma \left ({{M + \alpha } }\right){\Omega ^\alpha }}}{{\Gamma \left ({M }\right){(\beta)^\alpha }\Gamma \left ({{\alpha + 1} }\right)}}} }\right]^{N}} \cdot \\&\times \,\int \limits _{0}^\infty {\frac {{{x^{N\alpha - \frac {1}{2}}}}}{{{e^{\frac {v}{2}x}}}}{{\left [{ {_{2}{F_{1}}\left ({{M + \alpha, \alpha; 1 + \alpha; - \frac {\Omega }{\beta }x} }\right)} }\right]}^{N}}} dx \\{}\tag{24}\end{align*}
The hypergeometric function can be written as [38]:\begin{align*}&\hspace {-0.8pc}{\left [{ {_{2}{F_{1}}\left ({{M + \alpha, \alpha; 1 + \alpha; - \frac {\Omega }{\beta }x} }\right)} }\right]^{N}} \\&\qquad \qquad \qquad ={\left [{ {\sum \limits _{k = 0}^\infty {\frac {{{{\left ({{M + \alpha } }\right)}_{k}}{\left ({\alpha }\right)_{k}}}}{{{{\left ({{1 + \alpha } }\right)}_{k}}}}} } }\right]^{N}}\frac {{{{\left ({{ - \frac {\Omega }{\beta }} }\right)}^{kN}}}}{{{{\left [{ {k!} }\right]}^{N}}}} \cdot {x^{kN}} \\{}\tag{25}\end{align*}
\begin{equation*} {(q)_{k}} = q(q + 1) \ldots (q + k - 1) = \frac {\Gamma (q + k)}{\Gamma (q)} \tag{26}\end{equation*}
Therefore, the BER closed form expression can be achieved after substituting equation (25) into equation (24) as:\begin{align*} {P_{{\mathrm{BER}}}}=&E\left \{{ {Q\left ({{\sqrt {v\gamma } } }\right)} }\right \} \\=&\frac {\sqrt {v} }{{2\sqrt {2\pi } }}{\left [{ {\frac {{\Gamma \left ({{M + \alpha } }\right){\Omega ^\alpha }}}{{\Gamma \left ({M }\right){{({\beta _{r}})}^\alpha }\Gamma \left ({{\alpha + 1} }\right)}}} }\right]^{N}}\int \limits _{0}^\infty {{e^{ - \frac {v}{2}x}} \cdot {x^{N\alpha - \frac {1}{2}}}} \\&\cdot \, {\left [{ {\sum \limits _{k = 0}^\infty {\frac {{{{\left ({{M + \alpha } }\right)}_{k}}{\left ({\alpha }\right)_{k}}}}{{{{\left ({{1 + \alpha } }\right)}_{k}}}}} } }\right]^{N}} \cdot \frac {{{{\left ({{ - \frac {\Omega }{\beta }} }\right)}^{kN}}}}{{{{\left [{ {k!} }\right]}^{N}}}} \cdot {x^{kN}}dx \\=&{\left [{ {\frac {{\Gamma \left ({{M + \alpha } }\right){\Omega ^\alpha }}}{{\Gamma \left ({M }\right){(\beta)^\alpha }\Gamma \left ({{\alpha + 1} }\right)}}} }\right]^{N}}{\left [{ {\sum \limits _{k = 0}^\infty {\frac {{{{\left ({{M + \alpha } }\right)}_{k}}{\left ({\alpha }\right)_{k}}}}{{{{\left ({{1 + \alpha } }\right)}_{k}}}}} } }\right]^{N}} \cdot \\&\times \,\frac {{{2^{Nd + kN - 1}}}}{{\sqrt \pi {v^{Nd + kN}}}} \cdot \frac {{{{\left ({{ - \frac {\Omega }{\beta }} }\right)}^{kN}}}}{{{{\left [{ {k!} }\right]}^{N}}}} \cdot \Gamma \left ({{N\alpha + kN + \frac {1}{2}} }\right) \\{}\tag{27}\end{align*}
The value of \begin{equation*} {C_{out}} = \frac {{(1 - {P_{out}}){\log _{2}}(1 + {\gamma _{th}})}}{2} \tag{28}\end{equation*}
The acquisition of the optimal throughput depends on the effective time of information transmission in the UAV relay assisted communication system can be obtained into two cases, the system throughput of TS protocol can be given as:\begin{equation*} T = {C_{out}}(1 - \alpha) \tag{29}\end{equation*}
\begin{equation*} T = {C_{out}} \tag{30}\end{equation*}
Simulation and Analysis
This section simulates and analyzes the system performance of the proposed schemes and verifies the correctness of the theoretical analysis. The simulation parameters can be seen as follows: the number of UAV relays is N, the SNR maximization criterion is adopted for the UAV relays selection, the UAV relay energy harvesting of conversion efficiency
The Figure 6 depicts the system throughput varies with different scaling factor and different thresholds under TS and PS schemes and the transmitting power is fixed. Moreover, the relay number
System throughput varies with different scaling factor under different thresholds.
Figure 7 depicts the curves of system outage probability varies with different scaling factor under different thresholds condition. It is obvious that the TS scheme is superior to PS scheme with the corresponding thresholds. The transmitting power required will be smaller with the threshold decreasing, the harvested energy for the UAV relays is smaller with parameters
Figure 8 displays the curves of system throughput versus scaling factor with different number of interference under TS and PS protocols. The UAV relay number
The system throughput versus scaling factor with different number of interference.
In Figure 9 shows the system outage probability versus scaling factor
The system outage probability versus scaling factor with different number of interference.
In Figure 10, the system relay number
In Figure 11, it can be seen that when
Figure 12 displays that the system optimal throughput curves when the interference number
Figure 13 displays the curves of the system BER performance. The system interference number
As can be seen from Figure 14, it is worth noting that the parameter m represents the effect of line of sight. The larger the m is, the better the LoS effect have, then the system performance can be better. Moreover, the system performance of LoS can be superior to the non LoS case.
Conclusion
The paper proposed multi-UAV relay assisted network in Internet of Things (IoT) systems enhanced with energy harvesting, which effectively overcome the large scale fading between source and sink node and reduce the transmitting power to the required minimum value. Moreover, the proposed TS and PS schemes strategies were typically applied for UAV relays and achieved the energy harvesting with information transmission, where the multiple UAV assisted relays were selected via SNR maximization criterion so that the terminal node can obtain the optimal signal. Meanwhile, the outage probability and BER closed expressions were derived with the terminal node disturbed by aggregate interference caused from dense network signaling interaction in the future 5G/B5G systems. Additionally, the throughput and delay limited state of UAV relay assisted transmission were also analyzed thoroughly. The derivations and analysis results showed that the proposed multiple parameters joint optimization can effectively improve the system throughput and reduce the system outage probability and BER. Simulation experiments verified the effectiveness of the proposed schemes and the correctness of theoretical analysis.