1 Introduction
The fifth generation (5G) of cellular networks for beyond 2020 envisages to handle two new use cases in MachineType Communications (MTC), namely UltraReliable Low Latency Communications (URLLC) and massive MTC (mMTC) [1]. In MTC, MTC devices autonomously communicate with minimum human cooperation [2], [3]. 5G communication technology should be flexible enough to support ultrareliable low latency communications by guaranteeing reliability greater than [4]. Key challenges and requirements of 5G technology such as latency, data rate, energy and cost issues are discussed in more details in [1], [5], [6].
In recent years, MTC has gained much attention from the mobile network operators, equipment vendors and academic researchers due to such novel communication paradigm, the capability of exchanging short data messages and also being costeffective, energy efficient[7], [8]; reliable and within a stringent delay requirement. MTC takes advantage of several distinctive properties such as groupbased communications, low mobility, timecontrolled, timetolerant and secure connection which are at the same time challenging tasks since technically advanced solutions are needed to deliver the required tasks. Within the application requirements, hence, opening up different research areas is currently being carried out in academia, industry, and standards bodies [9]. Current technologies cover a small range of applications and services while the upcoming MTC should be able to cover a broad range of services with multiple forms of data traffic in order to deal with different service requirements as data rate, latency, reliability, energy consumption and security [6], [10]. Future MTC improvements will be conspicuous in healthcare, logistics, process automation, transportation, e.g [11], [6]. In mMTC a huge number of devices in a specific domain are connected to the cellular network with lowrate and lowpower connectivity, different qualityofservice (QOS) requirements and high reliability to support demanding situations, e.g. smart meters, actuators [12], [13].
Moreover, MTC services have to met stringent timing constraints from few seconds to even excessively low endtoend deadlines in mission critical communications [14], connection between vehicles, remote control of robots in addition to an extreme low endtoend latency in the scope of less than a millisecond which is a key enabler in several services including cloud connectivity, industrial control, road safety [1], [11], [15]. Latency refers to the time duration between transferring the message from the transmitter and receiving correctly at the receiver where some messages drop due to the buffer overflows, unsuccessful synchronizations, unsuccessful decoding which result in unlimited delay [1]. Hence, we can define the reliability as the probability of successful transmission under the predetermined delay constraint [1], [16]. In URLLC, high probability of successful transmission indicates low outage probability (or packet drop) while the opposite does not always hold as the reliability is restricted to a specific latency budget due to the limited amount of channel uses [1]. Hence, one of the major requirements of URLLC is a extremely low outage probability under a very demanding latency budget where retransmissions are not always available. In the use of short messages under URLLC, new robust channel codes are needed; otherwise, the performance of the system will be even further away from the Shannon limit with long data packets [17].
Under Shannon’s channel coding theorem, errorfree communication is attained when the blocklength goes to infinity [18]. For instance, authors in [19], provide a tight approximation of achievable coding rate under finite blocklength (FB) regime and indicate a noticeable performance loss compared to the Shannon coding. This motivates us to analyze the performance of MTC under FB regime since in URLLC, due to the equal packet length of metadata and information bits, an unsuccessful encoding of the metadata decreases the system efficiency [15]. In the past few years, several works have studied different aspects of FB coding since majority of the theoretical results assume infinite blocklength (IFB). For instance, authors in [20], examine some possible FB coding schemes which may be applied in 5G technology. They show that novel coding schemes with better minimum distance between the codewords, improve the efficiency of system at the cost of more sophisticated decoders. Moreover, the performance of spectrum sharing networks with FB codes are studied in [21]. The blocklength of information bits highly affects the system quality where an optimal power allocation technique improves the system efficiency with short message transmissions. Furthermore, authors in [22], propose a new power allocation technique, socalled modified waterfilling in order to maximize the lower bound of the coding rate with short packet transmission compared to the common waterfilling method. In addition, performance of ARQ protocol in terms of throughput and average latency is studied in [23]. Authors determine the optimal lengths of the codeword which minimize the latency and maximize the throughput peruser for an specific number of information bits. They illustrate that with optimal codes, the shorter the codeword is, the lower outage probability attains.
1.1 On the Impact of Cooperative Diversity
Cooperative diversity provides the possibility of high data rate; while improving the reliability. In cooperative networks, intermediate nodes transfer the message from the source to destination [24]. Cooperative technique exploits the spatial diversity gain to reduce the impact of wireless fading from multipath propagation. The major advantage of this technique is that the several independent copies of a signal arrive at the destination without installing collocated antennas at the source or receiver in addition of a better signal quality, better coverage, greater capacity and lower transmit power [25], [26]. The most conventional cooperative scheme is decodeandforward (DF), where the auxiliary node, namely relay, decodes, encodes and retransmits the message [25]. Cooperative schemes are categorized as fixed, adaptive and feedback schemes [25]. In the fixed protocol, relay always forwards the message to destination while in adaptive protocol, the relay retransmits the message under a predefined threshold rule which enables that to communicate independently or not. In the feedback protocol, if the destination requests, the cooperation takes place [25]. During the past few years, the efficiency of cooperative networks has been investigated in several system and channel models. Authors in [27], propose a method that meet the high reliability and latency requirements through taking the advantage of cooperative relaying technique. Moreover, authors in [26], provide a comprehensive study regarding the exiting cooperative schemes and analyze the performance of each scheme. Relaying performance of quasistatic Rayleigh channels where the channel gains of the direct link and relaying are combined at the destination, is studied in [18]. They indicate that the performance loss increases if the outage probability of the sourcetorelay link is higher than the overall outage probability. The efficiency of multirelay DF scenario under the assumption of perfect channelstateinformation (CSI) and partial CSI is provided in [28]. Authors show that with perfect CSI, the throughput of IFB is smaller than the throughput with FB coding. Authors in [29], examine the throughput of a multihop relaying network under FB and IFB regimes with two assumptions: target overall outage probability constant coding rate. They illustrate that there is different but optimal number of hops which maximize the throughput for either FB or IFB assumptions. In addition, they indicate that the FBthroughput is quasiconcave in the overall outage probability and coding rate. Furthermore, authors illustrate that the multihop network is less affected by the blocklength under the constant coding rate assumption compared to the target overall outage probability scenario. Moreover, [51] studies the performance of DF relay network in dissimilar Rayleigh fading channels. Although authors attain the closed form expression of the outage probability, but they do not consider the impact of finite blocklength coding. Furthermore, authors in [50], study the achievable coding rate and ergodic capacity of nonorthogonal amplyandforward (AF) multi relay network subject a total average power constraint (TAPC) and an individual average power constraint (IAPC). They indicate that the ergodic capacity can be attained by an iterative waterfillingbased algorithm. In addition, they show that in a multi relay NAF network, the transmit power at the source should be equally allocated in all broadcasting phases to cover the capacity at sufficiently high SNRs.
Moreover, in our previous work [30], introduces relaying as means to achieve ultrareliable. We study the performance of cooperative relaying protocols, supposing Rayleigh fading channels. We show that relaying technique improves the reliability and how we can meet the ultrareliable communication requirements. We examine the impact of coded blocklength and number of information bits on the probability of successful transmission. In addition, it is shown that relaying requires less transmit power compared to the direct transmission (DT) to enable ultrareliable under FB regime. We also provide an approximation to the outage probability in closed form. We extend our work in [30], by considering ultrareliable MTC with incremental relaying technique in [31]. We define the overall outage probability in each studied relaying scheme, assuming Nakagami fading. We investigate the impact of fading severity and power allocation factor on the outage probability. We also provide the outage probability in closed form. Our works in [31] and [30] show that cooperative diversity os useful to meet URLLC requirements.
1.2 Energy Efficiency of Cooperative Communication
Another key characteristic of wireless communications that highly affect the performance of 5G networks, is the energy efficiency (EE) due to the limited energy resources in energyconstraint networks [32], [33]. EE which has been widely studied recently literatures, is defined as the ratio of successfully transmitted bits to the total consumed energy [34],[35]. Hence, reducing the amount of energyperbit, improves EE at low SNR regime [33], particularly in wireless networks where the batteries which are not rechargeable or easy to charge, supply the wireless components [34]. The reason which motivates us to study EE in the context of URLLC is that,URLLC is achieved at the cost of high transmit power [36], [37], but we aim to show that cooperative diversity alleviates these demands.
In early works, authors in [35], examine EE in tactile Internet under queuing and transmission delays to design an energy efficient resource allocation strategy. They propose an optimal resource allocation strategy where the average total consumed power under stringent latency constraint equals to that with unlimited queuing latency requirement with plenty of transmit antennas. Moreover, [34], provides a comprehensive overview of energyefficient networks, and determines the tradeoff between energy efficiency and spectrum efficiency and their applications in 5G networks.
1.3 Our Contribution
In this work, we further study three cooperative protocols, namely DF, selection combining (SC) and MRC. Furthermore, we indicate the superiority of MRC over SC and DF protocols in terms of coding rate and reliability. We also show the optimal value of power allocation at the source and relay in each of studied protocols. Moreover, we examine the minimum latency and energy efficiency in cooperative schemes under two different power allocation constraints.
The following are considered the contributions of this work.

We provide the general expression of the outage probability for each relaying scheme studied in this work.

We extend the work in [18], by proposing the closed form expression for the outage probability.

We provide the asymptotic analysis of studied cooperative schemes.
The rest of this paper is organized as follows. Section 2 presents the system model. Section 3 discusses the cooperative diversity and examines the outage probability in three cooperative schemes considered in this work, and Section 4.1 presents some numerical results regarding the performance of studied cooperative schemes under URR. Section 4.2 investigates the energy efficiency of considered cooperative schemes and presents some numerical results. Finally, Section 5 concludes the paper. The important abbreviation and symbols are listed in Table 1.
bpcu  Bit per Channel Use 

CSI  Channel State Information 
DF  Decode and Forward 
DT  Direct Transmission 
EE  Energy Efficiency 
EPA  Equal Power Allocation 
FB  Finite Blocklength 
IFB  Infinite Blocklength 
mMTC  Massive MachineType Communication 
MRC  Maximum Ratio Combining 
MTC  MachineType Communication 
OPA  Optimal Power Allocation 
Probability Density Function  
QOS  Quality of Service 
RV  Random Variable 
SC  Selection Combining 
SH  Single Hop 
SNR  Signal to Noise Ration 
URLLC  UltraReliable Low Latency Communication 
URR  Ultra reliable Region 
Probability Density Function  
Expectation  
Function  
Shannon Capacity  
Channel Dispersion  
Inverse of Function  
Exponential Euler’s Number  
Received Signal  
Fading Channel  
AWGN Noise  
Information Bits  
Transmitted Signal  
Number of Channel Uses  
Logarithm to the Base 2  
Number of Channel Uses for Source  
Number of Channel Uses for Relay  
Distance of SourceDestination Link  
Distance of SourceRelay Link  
Distance of RelayDestination Link  
Energy Efficiency  
Total Power  
Power of Relay  
Power of Source  
Power of Transmitter  
Power of Receiver  
Power of Amplifier  
Probability of Successful Transmission  
Maximum Total Power  
Power of Noise  
Instantaneous SNR  
Average SNR  
Average Power Constraint  
Power Allocation Factor  
Maximum Coding Rate  
Drain Efficiency  
Outage Probability 
2 Preliminaries
2.1 System Model
Fig. illustrates a DF relaying scenario including a source , destination and a decodeforward relay . We normalize the distance of to as m, and that can move in a straight line between and , while the distance between and is denoted by and the distance of the relaying link is denoted by . The links denoted by the following random variables , and represent the ,  and  links respectively, and each transmission uses channel uses where . This means that channel uses in the broadcasting phase and channel uses for the relaying phase. In this scenario, first sends the message to the and in the broadcasting phase and if successfully decodes the message, forwards it to the in the relaying phase [18]. The received signals in the broadcasting phase are denoted as and , and only if collaborates with , the received signal at is as follow [18]
(1) 
where is the transmitted signal with power and is the AWGN noise with power where . Quasistatic Rayleigh fading channels in the ,  and  links are denoted as , and , respectively. In this work, we consider two distinct power constraints, namely ) EPA where equal powers are allocated to and and ) OPA where total power of and is equal to the maximum power. In a DFbased relaying protocol, the instantaneous SNR depends on the total power constraint , which is given by , and , where is the power allocation factor considered to provide a fair comparison between DT and cooperative transmissions and with EPA. Hence, the average SNR in each link is , and .
2.2 Performance Analysis of SingleHop Communication under the Finite Blocklength Regime
In this section we revisit the concept of FB coding. In a single hope communication, first information bits are mapped to a sequence, namely codeword including
symbols. Afterwards, the created codeword passes the wireless channels and channel outputs map into the estimate of the information bits. Thus, for a singlehop communication with blocklength
, outage probability and the average power constraint , where holds, the maximum coding rate of AWGN channel in bits is calculated as(2) 
where, is the positive channel capacity and is the channel dispersion [15]. According to (2), the outage probability is given by
(3) 
which holds for the AWGN channels where the channel coefficient is equal to one. While for quasistatic fading channels, we attain the outage probability as follow [15]
(4) 
Note that (4) is accurate for , as proved for AWGN channels [19, Figs. and ], as well as for fading channels as discussed in [38]. In addition, in the relaying schemes, we assume that can encode information bits into channel uses, while uses channel uses. Hence, and could employ more sophisticated encoding technique than [18], [39].
2.3 ClosedForm Expression of the Outage Probability
Unfortunately, (4) does not have a closedform expression, but it can be tightly approximated as we shall see next.
Lemma 1.
The outage probability is approximated as
(5) 
where and , where .
Proof.
Remark.
Moreover, we compare the accuracy of linearized Qfunction in (6) to original Qfunction in (3) as indicated in Fig.. The difference between these two plots does not have a noticeable impact on the outage probability since we find the approximated outage probability in (5) via integrating over the SNR range and due to the symmetric property of the function as evinced by Fig., regions that show the difference between the original and linearized Qfunction, cancel each other and so, this difference becomes negligible as illustrated in Fig.. Thus, we can notice that error defined by is approximately equal to zero which shows the accuracy of the linearized Qfunction applied in the closed form expression of the outage probability. Fore example, the maximum error over the entire SNR range is about . Similar conclusion holds for other values of .
3 The Proposed Method
In this section, we investigate the outage probability of cooperative DF, SC and MRC protocols under the FB regime. The direct transmission model is used here as the basis of the comparison.
3.1 Direct Transmission
The source sends the message directly to the destination, where , with average SNR , where the outage probability is calculated as in (5) but with and with .
3.2 Dual Hop DecodeandForward (DF)
In this scheme, since the  distance is too large, it assumes that the direct link is in the outage; thus, always collaborates with the source. Hence, sends the message to both and in the broadcasting phase. Then, transfers the message to [41]. The overall outage probability is given by
(7) 
where and are calculated according to (5). Notice that we update with , and with , , respectively. This scenario can be analyzed as selection combining (SC) or maximum ratio combining (MRC) depending on how the destination combines the original transmitted signal and the retransmitted signal.
3.3 Selection Combining (SC)
In this protocol, starts to collaborate with if the destination confirms that the source transmission was unsuccessful and so, the destination requests for retransmission from to receive the frame correctly. Cooperation occurs if decodes the received message from correctly and so, transfers the message to . Thereafter, if confirms that the transmission from is also failed, requests for the next subsequent message from . Thus, the outage probability happens only if both  and  links are in outage [42], [43]. The overall outage is given by
(8) 
where is equal to (5) where is updated with and with .
3.4 Maximum Ratio Combining (MRC)
In this scenario, relay always collaborates with the source and so, the channel gains of  and  links are combined at the receiver. Thus, the aggregated SNR is bigger than the primary attempted transmission rate as the  transmission failed. In addition, the outage probability occurs if  or  transmission fails. Hence, the instantaneous SNR is [42], [43]. The outage probability is [42]
(9) 
where is the outage probability of the sourcetorelayto destination link, notice that the term refers to the probability that D was not able to decode S message alone. In order to calculate the (9), first we need to attain the PDF of , and then we calculate the outage probability as proposed in proposition 1. To do so, let denote the sum of two independently distributed exponential random variables (RV), and . Then, is [42]
(10) 
Since the RVs are independent, the proof is straightforward solution of [44].
Proposition 1.
The outage probability of the MRC of the  and  links , is equal to
(11) 
(12) 
(13) 
where, and are specified in (6), and , , , , , and .
3.5 Asymptotic Analysis
The outage probability in (5) can be defined as as the SNR goes to infinity, where . Thus, the approximated asymptotic outage probability per link in Rayleigh fading channels is , where is a function of and [36, §10]. Thereafter, we resort to Taylor series as approaches zero as , and so, and attain an asymptotic expression as [36, §11]. The asymptotic expression of after maximum ratio combining of and transmissions is given in [42, §7]. Therefore, the outage probability is approximated as , resorting to series expansion as .
4 Numerical Results and Discussion
4.1 URLLC via Cooperative Diversity
In this section we show some numerical results of cooperative relaying transmission under FB regime. First, we show the impact of coding rate on the probability of successful transmission where MRC protocol outperforms DF, SC and DT in terms of reliability. We also indicate the minimum latency required to support URLLC. Thereafter, we show the optimal value of power allocation factor for each of studied protocols. In addition, we compare the performance of cooperative relaying to DT in terms of power consumption and blocklength to perform under the UR region (URR). We verify the accuracy of our analytical model through the MonteCarlo simulations. Unless stated; otherwise, assume maximum transmit power per link as dB, , , and is in between and , with . The URR is shaded purple area in the following plots, and its most loose constraint is denoted with a red line where the outage probability is , thus reliability is feasible.
4.1.1 Reliability vs. Coding Rate
Fig. compares the probability of successful transmission () as a function of coding rate in URR. We can clearly see that MRC supports URLLC with higher coding rates compared to DT, DF and SC schemes which is more evident with short packet lengths under the FB regime. Hence, MRC is less affected by the coding rate growth under the URR. For instance, with and , MRC covers reliability, while SC provides equal reliability as MRC but with lower coding rate as and . In addition, with , reliability decreases to and with and for DF and DT schemes, respectively. Thus, URLLC is feasible via the cooperative schemes and we can apply each of these schemes based on our requirements such as reliability, packet length and number of transmitted information bits.
In Fig. we examine the impact of power allocation factor on the probability of successful transmission. As mentioned earlier in Section 2, in order to provide a fair comparison between DT and cooperative schemes, we allocate powers to and according to the power allocation factor. In DF, outage probability is minimized via equal power allocation strategy while in SC and MRC, we exploit additional diversity of the direct link; thus, less power should be allocated to as shown in Fig.. We also illustrate that URLLC is feasible through the cooperative schemes, particularly with SC and MRC where the outage probability is minimized to and , respectively. As we indicated in our previous work in [30], these results holds for other values of SNR and coding rate.
In Fig. we compare the ultrareliable performance of cooperative schemes to DT in terms of transmit power under equal power allocation constraint. We can clearly see the power gain attained via the cooperative protocols at high SNR regime where there is huge performance gap between cooperative schemes and DT. In addition, we indicate that MRC and SC protocols perform closely in the entire SNR range and consumes less transmit power to communicate under the URR in comparison to DF and DT.
In addition, we indicate the possibility of using asymptotic expressions in ultrareliable region. In other words, at high SNR regime, the maximum achievable coding rate (2) converges the asymptotically long codewords as , where . In Fig. we show that the asymptotic expressions approach the analytical results as the transmit power increases.
Fig. indicates the performance advantage of cooperative schemes over DT. Cooperative schemes exploit diversity gain which decreases the outage probability remarkably. As we expected, the outage probability decreases in blocklength. In addition, SC and MRC protocols are able to support URLLC under FB regime with very short packet lengths.
Fig. indicates the total minimum latency required for URLLC under the FB regime with two distinct power constraints as ) EPA: , where , and ) OPA: . The choices of the minimum latency and optimal powers are in such a way that minimizes the outage probability constraint to a specific interval of interest and holds the power constraints which gives the optimal values of and , and is a nonlinear optimization problem as follows ^{1}^{1}1We solve the optimization problem numerically with the Matlab function . Interior point algorithm is used to solve the nonlinear optimization problem [46]..
subject to  
We set the minimum blocklength to 100 since (4) is accurate for , as proved for AWGN channels [19, Figs. and ] as well as for fading channels as discussed in [38], and to a maximum of so to reduce the search range, and to be within URLLC boundaries.
It shows that DT is not able to cope with the stringent latency constraint and need a large tolerance of delay; thus, we resort to cooperative protocols in order to reduce the latency in URR. It can be clearly seen that SC works highly better than DF and performs closely to MRC in the entire range but with higher latency requirements when we allocate equal powers to the and . We also indicate that under OPA constraint, SC outperforms MRC in terms of channel uses and is more energy efficient than MRC as we discuss about it in the following section, while with equal power allocation strategy, MRC requires less channel uses and consumes less transmit power as we can see in Fig.. Here, with equal power allocation strategy, the total power of and may be less than the maximum total power ( dB) but in Fig. we force and to be equal with total power of . Therefore, according to the simulations, when bpcu and dB, higher reliability is feasible in Fig. compared to Fig..
4.2 Energy Efficiency Analysis
Energy efficiency (EE) determines the tradeoff between throughput gains and total energy consumed. Let us first define the total energy consumption per bit of each scenario. The total power consumed includes power of transmission with no dependency on the distance of relay nodes, consumed power in radio frequency(RF) circuitry and also coding rate [47], [48]. Here, we ignore the baseband processing consumption since its value is negligible in comparison to the energy consumption of RF circuitry [49].
Then the total energy consumption per bit of a singlehop transmission is
(15) 
where is the power amplifier consumption for a singlehop transmission and is the drain efficiency of the amplifier, and are the power consumed for transmitting and receiving in the internal circuitry, respectively. In a similar way, we can also find the total power consumption of multihop schemes by determining the outage probability on  link in each cooperative schemes.
4.2.1 Cooperative Transmissions
The total power consumption for DF protocol depends on the outage probability of  link as follow
(16) 
where the first term indicates that the consumed energy on the  link, while the second term shows that could decode the message correctly and send the packet to .
In the case of SC and MRC, the total power consumption is formulated as follow
(17) 
where and is calculated by (5) accordingly to each method. The additional in each term of (17) compared to the (16), corresponds to the transmission of , which is heard by both and and destination decodes  and  transmissions, simultaneously.
Hence, the EE for each protocol is formulated as
(18) 
Furthermore, as observed from Fig., is nonconvex in the SNR, while the outage probability is monotonically decreasing in the SNR and energy consumption is monotonically increasing, which is observed in Figs. and , respectively. We maximize the energy efficiency as follow
subject to  
This problem is equivalent to minimize the outage probability with respect to , and blocklength . Since we aim to compare the performance of cooperative schemes, we do not focus on the proposal of a particular solution, but we resort to numerically efficient algorithm. Therefore, we resort to implemented in Matlab and use interior point algorithm to solve the nonlinear optimization problem as detailed in [46]. We consider outage probability threshold in an interval of interest as . At each outage probability value, we numerically determine , and blocklength that maximize the energy efficiency. We apply the numerical optimization due to the nonlinear constraint on the outage probability .
Fig. compares the energy efficiency of cooperative schemes in terms of probability of successful transmission under two distinct power constraints. In this paper, we assume mW, mW and the drain efficiency according to the power consumption values reported in [49]. Under EPA strategy, MRC is the most energy efficient scenario among other cooperative scenarios as it consumes less transmit power shown in Fig., and has lower latency in URR while under OPA, SC becomes the most energy efficient protocol as we show in Fig. it reduces the latency and the total power consumption is less than that of MRC. Since Fig. indicates that in order to perform in URR, we should allocate more power to the source where is equal to and for SC and MRC, respectively. Hence, more power is allocated to the source of MRC than that of SC; thus, MRC becomes less energy efficient compared to SC under OPA strategy.
Fig. compares the total consumed energy in each of studied cooperative scenarios under EPA and OPA strategies. Under EPA, as we expected MRC is superior and consumes less transmit power compared to DF and SC protocols, while with OPA, SC outperforms MRC and becomes most energy efficient protocol. In addition, with the maximum transmit power of dB, and no feasible solutions are found for DF protocol under stringent reliability requirements, which evinces the need for more sophisticated cooperative protocols. Feasible solutions are found if the transmit power increases, but it would be spectrally and energy insufficient.
5 Conclusions
In this paper, we assess the relay communication under the finite blocklength regime under Rayleigh fading. Performance of three relaying scenario, namely DF, SC and MRC is compared to direct transmission under two distinct power constraints socalled EPA and OPA. Based on the outage probability analysis of each transmission protocol, we show that relaying improves the probability of a successful transmission and guarantees ultrahigh reliability with FB codes. MRC protocol is less affected by the coding and provide higher reliability compared to DT and two other relaying scenarios. In addition, we numerically show the optimal power allocation for the relaying protocols under study so to operate in URR. Our results shows that operation at URR is feasible by allocating more power to the source; however, relay node is considered to provide additional diversity gain compared to the DT which is more evident at high SNR regime. We compare the studied cooperative schemes in terms of latency and energy efficiency under the two distinct power constraints. According to the results, with equal power allocation at source and relay, MRC is the most energy efficient protocol with lower latency and power consumption compared to the other scenarios while SC has the highest energy efficiency and lowest latency under optimal power allocation strategy. Moreover, we provide the outage probability in closed form and prove the accuracy and appropriateness of our analytical model through numerical results. Finally, in our future work, we will focus on the impact of imperfect channel state information on URLLC.
Declarations
Availability of data and material
The manuscript is selfcontained. Simulations description and parameters are provided in details in Section 4.
Competing interests
The authors declare that they have no competing interests.
Funding
This work has been partially supported by Finnish Funding Agency for Technology and Innovation (Tekes), Huawei Technologies, Nokia and Anite Telecoms, and Academy of Finland (under Grant no. 307492)
Author’s contributions
All authors have contributed to this manuscript and approved the submitted manuscript.
Acknowledgments
No applicable.
Author details
Centre for Wireless Communications (CWC), University of Oulu, Finland
References
 [1] P. Popovski, J. J. Nielsen, C. Stefanovic, E. de Carvalho, E. Ström, K. F. Trillingsgaard, A.S. Bana, D. M. Kim, R. Kotaba, J. Park et al., “Ultrareliable lowlatency communication (urllc): Principles and building blocks,” arXiv preprint arXiv:1708.07862, 2017.
 [2] H. Shariatmadari, R. Ratasuk, S. Iraji, A. Laya, T. Taleb, R. Jäntti, and A. Ghosh, “Machinetype communications: current status and future perspectives toward 5G systems,” IEEE Communications Magazine, vol. 53, no. 9, pp. 10–17, 2015.
 [3] Y. Mehmood, C. Görg, M. Muehleisen, and A. TimmGiel, “Mobile m2m communication architectures, upcoming challenges, applications, and future directions,” EURASIP Journal on Wireless Communications and Networking, vol. 2015, no. 1, p. 250, 2015.
 [4] A. E. Kalør, R. Guillaume, J. J. Nielsen, A. Mueller, and P. Popovski, “Network slicing for ultrareliable low latency communication in industry 4.0 scenarios,” arXiv preprint arXiv:1708.09132, 2017.
 [5] J. G. Andrews, S. Buzzi, W. Choi, S. V. Hanly, A. Lozano, A. C. Soong, and J. C. Zhang, “What will 5G be?” IEEE Journal on selected areas in communications, vol. 32, no. 6, pp. 1065–1082, 2014.
 [6] Z. Dawy, W. Saad, A. Ghosh, J. G. Andrews, and E. Yaacoub, “Toward massive machine type cellular communications,” IEEE Wireless Communications, vol. 24, no. 1, pp. 120–128, 2017.
 [7] B. Lee, S. Park, D. J. Love, H. Ji, and B. Shim, “Packet structure and receiver design for lowlatency communications with ultrasmall packets,” in Global Communications Conference (GLOBECOM), 2016 IEEE. IEEE, 2016, pp. 1–6.
 [8] T. Taleb and A. Kunz, “Machine type communications in 3GPP networks: potential, challenges, and solutions,” IEEE Communications Magazine, vol. 50, no. 3, 2012.
 [9] M. Condoluci, M. Dohler, G. Araniti, A. Molinaro, and K. Zheng, “Toward 5G densenets: architectural advances for effective machinetype communications over femtocells,” IEEE Communications Magazine, vol. 53, no. 1, pp. 134–141, 2015.
 [10] J. F. Monserrat, G. Mange, V. Braun, H. Tullberg, G. Zimmermann, and Ö. Bulakci, “Metis research advances towards the 5g mobile and wireless system definition,” EURASIP Journal on Wireless Communications and Networking, vol. 2015, no. 1, p. 53, 2015.
 [11] C. Bockelmann, N. Pratas, H. Nikopour, K. Au, T. Svensson, C. Stefanovic, P. Popovski, and A. Dekorsy, “Massive machinetype communications in 5G: Physical and maclayer solutions,” IEEE Communications Magazine, vol. 54, no. 9, pp. 59–65, 2016.
 [12] P. Popovski, “Ultrareliable communication in 5G wireless systems,” in 1st International Conference on 5G for Ubiquitous Connectivity (5GU), 2014 . IEEE, 2014, pp. 146–151.
 [13] B. Singh, O. Tirkkonen, Z. Li, M. A. Uusitalo, and R. Wichman, “Selective multihop relaying for ultrareliable communication in a factory environment,” in 27th Annual International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC), 2016 IEEE . IEEE, 2016, pp. 1–6.
 [14] A. Biral, H. Huang, A. Zanella, and M. Zorzi, “On the impact of transmitter channel knowledge in energyefficient machinetype communication,” in Globecom Workshops (GC Wkshps), 2016 IEEE. IEEE, 2016, pp. 1–7.
 [15] G. Durisi, T. Koch, and P. Popovski, “Toward massive, ultrareliable, and lowlatency wireless communication with short packets,” Proceedings of the IEEE, vol. 104, no. 9, pp. 1711–1726, 2016.
 [16] O. Yilmaz, “Ultrareliable and lowlatency 5G communication,” in European Conference on Networks and Communications (EuCNC), vol. 2016, 2016.
 [17] M. Sybis, K. Wesolowski, K. Jayasinghe, V. Venkatasubramanian, and V. Vukadinovic, “Channel coding for ultrareliable lowlatency communication in 5G systems,” in 84th Vehicular Technology Conference (VTCFall), 2016 IEEE . IEEE, 2016, pp. 1–5.
 [18] Y. Hu, J. Gross, and A. Schmeink, “On the performance advantage of relaying under the finite blocklength regime,” IEEE Communications Letters, vol. 19, no. 5, pp. 779–782, 2015.
 [19] Y. Polyanskiy, H. V. Poor, and S. Verdú, “Channel coding rate in the finite blocklength regime,” IEEE Transactions on Information Theory, vol. 56, no. 5, pp. 2307–2359, 2010.
 [20] O. Iscan, D. Lentner, and W. Xu, “A comparison of channel coding schemes for 5G short message transmission,” in Globecom Workshops (GC Wkshps), 2016 IEEE. IEEE, 2016, pp. 1–6.
 [21] B. Makki, T. Svensson, and M. Zorzi, “Finite blocklength analysis of spectrum sharing networks,” in International Conference on Communications (ICC), 2015 IEEE . IEEE, 2015, pp. 7665–7670.
 [22] J.H. Park and D.J. Park, “A new power allocation method for parallel awgn channels in the finite block length regime,” IEEE Communications Letters, vol. 16, no. 9, pp. 1392–1395, 2012.
 [23] R. Devassy, G. Durisi, P. Popovski, and E. G. Strom, “Finiteblocklength analysis of the arqprotocol throughput over the gaussian collision channel,” in 6th International Symposium on Communications, Control and Signal Processing (ISCCSP), 2014 . IEEE, 2014, pp. 173–177.
 [24] S. Ikki and M. H. Ahmed, “Performance analysis of cooperative diversity wireless networks over nakagamim fading channel,” IEEE Communications letters, vol. 11, no. 4, 2007.
 [25] E. Zimmermann, P. Herhold, and G. Fettweis, “On the performance of cooperative relaying protocols in wireless networks,” Transactions on Emerging Telecommunications Technologies, vol. 16, no. 1, pp. 5–16, 2005.
 [26] F. Mansourkiaie and M. H. Ahmed, “Cooperative routing in wireless networks: A comprehensive survey,” IEEE Communications Surveys & Tutorials, vol. 17, no. 2, pp. 604–626, 2015.
 [27] V. N. Swamy, S. Suri, P. Rigge, M. Weiner, G. Ranade, A. Sahai, and B. Nikolić, “Cooperative communication for highreliability lowlatency wireless control,” in International Conference on Communications (ICC), 2015 IEEE . IEEE, 2015, pp. 4380–4386.
 [28] Y. Hu, A. Schmeink, and J. Gross, “Relaying with finite blocklength: Challenge vs. opportunity,” in Sensor Array and Multichannel Signal Processing Workshop (SAM), 2016 IEEE. IEEE, 2016, pp. 1–5.
 [29] F. Du, Y. Hu, L. Qiu, and A. Schmeink, “Finite blocklength performance of multihop relaying networks,” in International Symposium on Wireless Communication Systems (ISWCS), 2016. IEEE, 2016, pp. 466–470.
 [30] P. Nouri, H. Alves, and M. Latvaaho, “On the performance of ultrareliable decode and forward relaying under the finite blocklength,” in European Conference on Networks and Communications (EuCNC), 2017. IEEE, 2017, pp. 1–5.
 [31] P. Nouri, H. Alves, R. Demo Souza, and M. Latvaaho, “Ultrareliable short message cooperative relaying protocols under nakagamim fading,” in International Symposium on Wireless Communication Systems (ISWCS), 2017. IEEE, 2017.
 [32] D. Qiao, M. C. Gursoy, and S. Velipasalar, “Energy efficiency of fixedrate wireless transmissions under queueing constraints and channel uncertainty,” in Global Telecommunications Conference, 2009. GLOBECOM 2009. IEEE. IEEE, 2009, pp. 1–6.
 [33] M. C. Gursoy, “On the capacity and energy efficiency of trainingbased transmissions over fading channels,” IEEE Transactions on Information Theory, vol. 55, no. 10, pp. 4543–4567, 2009.
 [34] G. Wu, C. Yang, S. Li, and G. Y. Li, “Recent advances in energyefficient networks and their application in 5G systems,” IEEE Wireless Communications, vol. 22, no. 2, pp. 145–151, 2015.
 [35] C. She and C. Yang, “Energy efficient design for tactile internet,” in International Conference on Communications in China (ICCC), 2016 IEEE/CIC . IEEE, 2016, pp. 1–6.
 [36] E. Dosti, M. Shehab, H. Alves, and M. Latvaaho, “Ultra reliable communication via ccharq in finite blocklength,” in European Conference on Networks and Communications (EuCNC), 2017 . IEEE, 2017, pp. 1–5.
 [37] E. Dosti, U. L. Wijewardhana, H. Alves, and M. Latvaaho, “Ultra reliable communication via optimum power allocation for typei arq in finite blocklength,” arXiv preprint arXiv:1701.08617, 2017.
 [38] W. Yang, G. Durisi, T. Koch, and Y. Polyanskiy, “Quasistatic multipleantenna fading channels at finite blocklength,” IEEE Transactions on Information Theory, vol. 60, no. 7, pp. 4232–4265, 2014.
 [39] Y. Hu, A. Schmeink, and J. Gross, “Blocklengthlimited performance of relaying under quasistatic Rayleigh channels,” IEEE Transactions on Wireless Communications, vol. 15, no. 7, pp. 4548–4558, 2016.
 [40] B. Makki, T. Svensson, and M. Zorzi, “Finite blocklength analysis of the incremental redundancy harq,” IEEE Wireless Communications Letters, vol. 3, no. 5, pp. 529–532, 2014.
 [41] J. N. Laneman, D. N. Tse, and G. W. Wornell, “Cooperative diversity in wireless networks: Efficient protocols and outage behavior,” IEEE Transactions on Information theory, vol. 50, no. 12, pp. 3062–3080, 2004.
 [42] H. Alves, R. D. Souza, G. Fraidenraich, and M. E. Pellenz, “Throughput performance of parallel and repetition coding in incremental decodeandforward relaying,” Wireless Networks, vol. 18, no. 8, pp. 881–892, 2012.
 [43] H. Alves, R. D. Souza, G. Brante, and M. E. Pellenz, “Performance of typei and typeii hybrid arq in decode and forward relaying,” in 73rd Vehicular Technology Conference (VTC Spring), 2011 IEEE . IEEE, 2011, pp. 1–5.
 [44] P. Athanasios, “Probability, random variables, and stochastic processes,” 2017.
 [45] I. S. Gradshteyn and I. M. Ryzhik, Table of integrals, series, and products. Academic press, 2014.
 [46] R. A. Waltz, J. L. Morales, J. Nocedal, and D. Orban, “An interior algorithm for nonlinear optimization that combines line search and trust region steps,” Mathematical programming, vol. 107, no. 3, pp. 391–408, 2006.
 [47] G. G. de Oliveira Brante, M. T. Kakitani, and R. D. Souza, “Energy efficiency analysis of some cooperative and noncooperative transmission schemes in wireless sensor networks,” IEEE Transactions on Communications, vol. 59, no. 10, pp. 2671–2677, 2011.
 [48] H. Alves, R. D. Souza, and G. Fraidenraich, “Outage, throughput and energy efficiency analysis of some half and full duplex cooperative relaying schemes,” Transactions on Emerging Telecommunications Technologies, vol. 25, no. 11, pp. 1114–1125, 2014.
 [49] S. Cui, A. J. Goldsmith, and A. Bahai, “Energyconstrained modulation optimization,” IEEE transactions on wireless communications, vol. 4, no. 5, pp. 2349–2360, 2005.
 [50] Tran, Tuyen X and Tran, Nghi H and Bahrami, Hamid Reza and Sastry, Shivakumar, “On achievable rate and ergodic capacity of NAF multirelay networks with CSI,” IEEE transactions on communications, vol. 62, no. 5, pp. 14901502, 2014.
 [51] Beaulieu, Norman C and Hu, Jeremiah, “A closedform expression for the outage probability of decodeandforward relaying in dissimilar Rayleigh fading channels,” IEEE Communications Letters, vol. 10, no. 12, 2006.
Comments
There are no comments yet.