Received October 18, 2021, accepted December 6, 2021, date of publication December 8, 2021,

date of current version December 17, 2021.

Digital Object Identifier 10.1109/ACCESS.2021.3134207

Rate-Splitting Multiple Access for URLLC Uplink

in Physical Layer Network Slicing With eMBB
ELÇO JOÃO DOS SANTOS JR. 1, RICHARD DEMO SOUZA 1, (Senior Member, IEEE),
AND JOÃO LUIZ REBELATTO 2 , (Senior Member, IEEE)
1 Department of Electrical and Electronics Engineering, Federal University of Santa Catarina (UFSC), Florianópolis, Santa Catarina 88040-900, Brazil
2 Department of Electrical Engineering, Federal University of Technology—Paraná (UTFPR), Curitiba, Paraná 80230-901, Brazil
Corresponding author: Elço João dos Santos Jr. (e.joaojr@gmail.com)
This work was supported in part by CNPq, Brazil; in part by Print CAPES-UFSC ‘‘Automation 4.0;’’ and in part by RNP/MCTIC under
Grant 01245.010604/2020-14 (6G mobile communications systems).

ABSTRACT In this paper, we investigate the problem of heterogeneous service coexistence in the scope
of 5G and beyond (B5G) networks, where multiple ultra-reliable low-latency communication (URLLC) and
enhanced mobile broadband (eMBB) users are connected to a common base station (BS), sharing physical
network resources. In contrast to the orthogonal multiple access (OMA) and non-orthogonal multiple
access (NOMA) usually adopted in literature, in this work we employ rate splitting multiple access (RSMA)
for URLLC transmission, where a URLLC device splits its message into two sub-messages with partial
transmission power, which are potentially recovered at the BS by means of successive interference can-
cellation (SIC). To study the performance of such methods in the presence of eMBB users, we consider
both orthogonal and non-orthogonal network slicing approaches to share the network resources between
heterogeneous user profiles with diverse requirements. As a result, we show that, in general, RSMA presents
an improved performance in terms of sum-rate and reliability, even when transmitting concurrently with
eMBB users. Finally, our results also show that the URLLC sum-rate can be increased by properly adjusting
the rate splitting factor based on the average signal-to-noise ratio (SNR), not being necessary instantaneous
channel state information (CSI).

INDEX TERMS Beyond 5G, heterogeneous users, rate splitting multiple access, ultra-reliable and low
latency communications.

I. INTRODUCTION network softwarization and virtualization, being considered

As the 5G technology deployment around the world evolves, the main enabler of Resource as a Service (RaaS) for beyond-
it becomes clear how challenging are the three generic ser- 5G (B5G) [5]. In the path to B5G and 6G wireless commu-
vices encompassed by such technology, namely enhanced nication systems, it is reasonable to assume that the three
mobile broadband (eMBB), ultra-reliable and low latency heterogeneous services could be divided into sub-services [6]
communications (URLLC), and massive machine type com- or even combined, emerging new service classes [7]. Such
munications (mMTC). To allow the coexistence of these services require robust multiple access methods that can com-
heterogeneous services with diverse requirements within the bine higher spectral efficiency with strict delay and reliability
same Radio Access Network (RAN) architecture, the concept requirements to attend applications like fully automated driv-
of network slicing has been proposed [1], which slices the ing, where cooperation among cars for collision avoidance is
network in logical and physical sub-networks usually with vital [8], [9].
customized requirements in terms of latency, energy effi- To face the massive connectivity problem, some methods
ciency, mobility, massive connectivity and throughput [2], have been proposed in the past few years to replace the
aiming at guaranteeing minimum performance requirements traditional orthogonal multiple access (OMA). One of them is
and isolation [3], [4]. This can be performed thanks to non-orthogonal multiple access (NOMA), a promising tech-
nology that usually exploits the power domain to allow multi-
The associate editor coordinating the review of this manuscript and ple users to share the same resource block along the spectrum,
approving it for publication was Muhammad Awais Javed . time and/or code, increasing the spectral efficiency [10].

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
163178 VOLUME 9, 2021
E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

In order to recover the overlapped signals, the receiver of the power allocation factor that splits the messages of the
a NOMA-based communication system can apply the suc- near user can be fixed or dynamically designed based on CSI,
cessive interference cancellation (SIC) algorithm, a method respectively. This work is then extended in [29], adopting
whose performance depends on the different power levels cyclic prefixed single carrier transmissions. In both works,
between the overlapped incoming signals [11]. To this end, rate splitting has been shown to achieve superior outage
two approaches are commonly used to guarantee such power performance when compared to NOMA.
distinctiveness: (i) user pairing; and (ii) power allocation. In [30], an exhaustive-search rate splitting algorithm was
In (i), users with distinct channel gains are separated in groups proposed to guarantee max-min fairness in single-input
and paired [12], [13]. It is intuitive that the complexity of multiple-output (SIMO) NOMA networks, aiming at max-
such technique increases with the number of users, turning its imizing the minimum data rate and reduce the scheduling
implementation unbearable in terms of latency in scenarios process. The receiver combines minimum mean squared
with a massive number of users. In (ii) power allocation error (MMSE) with SIC to identify the optimal detection
methods separate users [14], [15], even if random pairing is order based on CSI. Results showed that rate splitting has
applied. This implies, in some cases, the need of channel state higher minimum data rate and lower transmission latency
acquisition to adapt the power of transmission, which entails than SIMO-OMA and SIMO-NOMA. The use of rate split-
extra latency and a potential loss in terms of reliability. ting in user cooperation networks is proposed in [31]. Each
The rate-splitting multiple access (RSMA) method has user transmits its signal and receives the transmitted signal
gained significant attention recently, since it enables the of the other user in the first mini-slot and, at the second
achievement of the entire capacity region with successive mini-slot, relays the other user’s message with amplify-
decoding [16], [17], providing superior performance over and-forward protocol. The rate is split between mini-slots,
NOMA and OMA methods, like Space Division Multiple generating space diversity at the uplink and consequently
Access (SDMA) [18], [19]. In uplink RSMA, each user cre- increasing reliability. At the receiver, maximum ratio com-
ates virtual users by splitting its transmission in two sub- bining (MRC) is used to combine the received signals and
messages. Although this procedure entails extra rounds in SIC is applied to decode the superposed signal. Results prove
the SIC procedure, it automatically creates different arriving that cooperative RSMA outperforms cooperative OMA and
power levels among users, thus significantly reducing the NOMA.
implementation complexity. One of the main advantages of In scenarios with spectrum sharing among URLLC and
RSMA is the increased number of possible decoding orders, eMBB services, several works compared OMA and NOMA
which makes it viable to reach higher capacity regions when network slicing [32]–[37]. However, none of the aforemen-
compared to NOMA. This is also a big challenge in practical tioned works consider multiple concurrent URLLC users in
RSMA deployments and must be optimized, since the decod- the same resource block. In [38], URLLC users are assumed
ing order affects the achievable rate. to share time and frequency resources through NOMA,
in both OMA and NOMA slicing with eMBB service. It was
A. RELATED WORK shown that NOMA can leverage the URLLC sum-rate in
Recently, several works studied different RSMA implementa- some cases, considering that the SIC process is capable
tions in downlink wireless networks [20]–[24], showing that of attending the communication latency. Authors from [39]
RSMA can improve downlink rate and quality of service, apply RSMA to URLLC in the downlink, showing its supe-
achieving better performance than both NOMA and SDMA. rior performance in terms of latency, allowing shorter block
For uplink RSMA systems, authors from [25], [26] study the lengths. However, no interference from other services is
problem of maximizing the sum-rate under proportional rate considered.
constraints for all users, by setting users transmission power
and optimizing the decoding order at the BS through exhaus- B. NOVELTY AND CONTRIBUTION
tive search. As a result, they show that RSMA achieves better Motivated by the above literature, in this work we focus
performance than NOMA and OMA techniques, such as on increasing the URLLC spectral efficiency, allowing
frequency division multiple access (FDMA) and time division non-orthogonal sharing of frequency and time resources
multiple access (TDMA). However, the proposed strategy through rate-splitting for URLLC users, which we refer
requires a priori channel state information (CSI), not being in to U-RSMA. In the proposed scheme, we combine the
general applicable to URLLC users due to latency constraints. benefits of RSMA, SIC decoding and frequency diversity,
In [27], the authors propose the use of RSMA to reduce in both OMA and NOMA slicing with eMBB. The pro-
the scheduling complexity of NOMA, since the transmission posed U-RSMA scheme is then compared to the so-called
splitting by default diversifies the arriving power at the BS, U-NOMA and U-OMA schemes, where the multiple access
avoiding the need of user pairing. In [28], the authors apply between URLLC devices is performed by means of NOMA
rate splitting to a pair of users under power-domain NOMA, and OMA, respectively. To characterize the performance of
considering that one of them is near the BS, while the other eMBB and URLLC users, we evaluate each service sum-rate
is far from the BS. Two techniques are studied, namely, fixed in different scenarios. To the best of our knowledge, this work
rate splitting (FRS) and adaptive rate splitting (ARS), where is the first to apply RSMA to URLLC uplink transmission in

VOLUME 9, 2021 163179

 E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

TABLE 1. List of symbols. device is uncorrelated from another due to the assumption
that all devices have a large enough spatial separation. Fur-
thermore, the fading is considered constant during one time
slot (TS), i.e., a block fading model where the TS is con-
sidered to be within the channel coherence time since its
length is fairly small [40]. As we assume that the average
transmission power of all devices and the noise power at
the BS are normalized to one, the received power equals the
signal-to-noise ratio (SNR) for each device. Moreover, the
channel fading realization for user i ∈ {B, U } in channel f is
Hi,f ∼ CN (0, 0̄i ), following a circular-symmetric complex
Gaussian distribution, where 0̄i corresponds to the average
SNR, being Gi,f , |Hi,f |2 the channel gain, and where
subscripts B and U refer to eMBB and URLLC devices,
respectively. The number of channels allocated to user i is
Fi ≤ F, with i ∈ {B, U }. Moreover, each TS is divided into
S mini-slots, as considered in low latency scenarios [41].
In accordance to [32], we assume that an eMBB user is
active with probability aB and during the entire TS, occupying
one random frequency channel f among FB available chan-
nels. Furthermore, we model only the transmission phase,
assuming that radio access and competition among eMBB
devices have been resolved prior to the considered time slot,
as usual in wireless cellular networks. Thus, the number of
eMBB devices able to transmit in such TS is equal to the
number of channels FB . Moreover, we suppose that the eMBB
devices and the BS have CSI as currently implemented in
wireless standards such as LTE and 5G New Radio [42]–[44].
Although channel estimation errors can occur in practice, for
simplicity we consider a perfect CSI scenario in this work,
as widely considered in the literature [32], [45]. In contrast,
an URLLC device spreads its transmission over FU ≤ F
channels to increase the reliability with the aid of frequency
diversity, and sends, with some activation probability aU , the
entire information in only one mini-slot (the smallest time
a network slicing scenario, showing that RSMA can outper- unit in our model) that was pre-assigned to meet latency
form OMA and NOMA methods for URLLC service even in requirements. We also consider that the protocol block length,
the presence of eMBB interference, specially for very strict which should be considered finite given the short transmis-
reliability levels. sions, is long enough to justify an asymptotic information-
The rest of this paper is organized as follows. Section II theoretic formulation [46]. Moreover, in each mini-slot we
presents the system model. Section III introduces the outage have a maximum number of nU users that share the resources
formulation for eMBB and URLLC (for U-OMA, U-NOMA, following three distinct methods: orthogonal (U-OMA), non-
and U-RSMA cases), for both orthogonal and non-orthogonal orthogonal (U-NOMA) or through rate splitting (U-RSMA)
network slicing approaches. Numerical results illustrating multiple access.
the performance trade-offs between the services are given in Different from eMBB users, we assume that the BS has
Section IV. Finally, Section V concludes the paper. no knowledge about the URLLC channel, given the high
Notation: For convenience, the list of symbols adopted in latency requirement which does not allow the exchange of
this work is summarized in Table 1. reference signals for CSI acquisition. However, we do con-
sider in U-RSMA that the BS sends (e.g., in a synchronization
II. SYSTEM MODEL mini-slot transmitted at the end of each TS), the optimal
We evaluate the uplink of multiple eMBB and URLLC users power splitting factor based on 0̄U from a look-up table,
when communicating to a common Base Station (BS) in a which results in power adaptation for the user that performs
single-cell network with shared radio resources. The band- the splitting. Despite that, the overall transmission power is
width is divided into F channels of index f ∈ {1, . . . , F} the same as in U-OMA and U-NOMA cases.
subject to independent and identically distributed (i.i.d.) A time-frequency grid is illustrated in Fig. 1, consid-
Rayleigh fading. The fading realization observed by each ering that the heterogeneous URLLC and eMBB traffics

163180 VOLUME 9, 2021

E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

radio resource f ∈ {1, . . . , FB }, if the instantaneous channel

gain is greater than a threshold SNR Gmin B,f . This decision is
made based on CSI. The outage probability of a point-to-point
(single channel) communication is then [32]
Z Gmin
B,f
P(EB ) = Pr[GB,f < Gmin
B,f ] = pGB,f (x)dx, (1)
0

where pGB,f (x) is the probability density function (PDF) of

GB,f , which, due to the Rayleigh fading, is given by

 e−x/0̄B
, if x > 0

pGB,f (x) =
 0̄B (2)
 0, otherwise

The eMBB outage probability is then obtained as [32]

Z Gmin
B,f e−x/0̄B
P(EB ) = dx
0 0̄B
1 Gmin
× −0̄B × e−x/0̄B
B,f
=
0̄B 0

= − e−GB,f /0̄B − e−0/0̄B

min
 

min /0̄
= 1 − e−GB,f B
. (3)
FIGURE 1. System model with F = 4 channels and S = 4 mini-slots,
composed by eMBB and URLLC users. Services are sliced in (a) and
(c) Orthogonal and (b) and (d) Non-Orthogonal multiple access schemes.
Imposing the reliability condition P(EB ) = B , one can
obtain the threshold SNR from (3) as
are sliced in an OMA (Figs. 1(a) and 1(c)), and NOMA
 
min 1
(Figs. 1(b) and 1(d)) fashion. In this example, S = 4 is the GB,f = 0̄B ln . (4)
1 − B
quantity of mini-slots in the time domain, whereas F = 4 is
the total number of channels available in the bandwidth. The main objective of eMBB is to maximize its data rate,
Considering the OMA scenario, two channels are allocated subject to the reliability requirement B and the average
to URLLC (FU = 2) and two for eMBB (FB = 2). There are power constraint E[PB (GB,f )] = 1, where PB (GB,f ) is the
nU = 2 URLLC active users, UU ,1 and UU ,2 , in each mini- instantaneous transmission power, selected using the power
slot that spread their transmission over one channel, in the inversion scheme from [47] based on GB,f , i.e.
case of U-OMA, or over two channels when considering 
tar
U-NOMA or U-RSMA, without interference from eMBB  GB,f
, if GB,f ≥ Gmin

PB (GB,f ) = GB,f B,f (5)
users. On the eMBB band, there are also two users, UB,1 
and UB,2 , connected to the BS. When considering NOMA, 
0, otherwise
all four channels are available for both services (F = FU =
FB = 4), which implies a multi-service interference, turning This means that the eMBB device will not transmit in
the detection at the BS more complex and prone to errors. The every slot allocated to it because of outage situations, then
frequency diversity gain for URLLC users is higher in this it is possible to increase the instantaneous power when the
case, and, as this device type does not necessarily transmit transmission occurs, so that the long-term average power
PB (GB,f ) = 1 is achieved. The target SNR Gtar


at every TS, the spectrum efficiency should increase because B,f is then
eMBB users can occupy a radio resource that might be unused obtained by imposing the average power constraint to the
for long periods, which is represented with the inclusion of expected value of the function PB (GB,f ) of the random vari-
new eMBB users UB,3 and UB,4 . able GB,f . This is calculated using
Z ∞
III. OUTAGE FORMULATION AND SLICING SCHEMES  
E PB (GB,f ) = pGB,f (x)PB (x)dx = 1. (6)
In this section, we discuss the achievable rates of the different Gmin
B,f
services and slicing schemes.
After replacing (2) in (6), one has
A. EMBB tar
∞ e−x/0̄B GB,f
Z
A given eMBB device transmits, with a certain instantaneous  
E PB (GB,f ) = dx = 1
power and data rate, in the randomly allocated dedicated Gmin
B,f
0̄B x

VOLUME 9, 2021 163181

 E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

Gtar
B,f
Z ∞ e−x/0̄B where σn,f is defined as
= dx = 1, (7)
0̄B Gmin x
B,f GU ,n,f
| {z } σn,f = PnU . (14)
1+ j>n GU ,j,f
 min 
GB,f
−Ei − 0̄
B
The decoding procedure starts with the strongest among all
B,f /0̄B
−Gmin the active users in the current mini-slot. If correctly decoded,


where −Ei is obtained from the integral and can
be classified as the upper incomplete gamma function 0(·, ·) it is removed from the received signal and the operation
B,f /0̄B > 0. Then, (7) can be rewritten as
for Gmin continues, until an user cannot be decoded (event that occurs
  with probability U ) or all users have been properly decoded.
Gtar Gmin We consider that the BS is capable of decoding the nU users
E PB (GB,f ) = 0̄B,f 0 0, 0̄B,f = 1.
 
(8)
B B within the mini-slot period, since each transmission carries a
The target SNR of eMBB user Gtar different message and the procedure must attend the latency
B,f is then obtained
from (8), resulting in requirement. The outage probability of the u-th user is
 
0̄B 1
FU
Gtar .
X
B,f =  (9) PU-NOMA (EU ) = Pr  log2 (1 + σn,f ) < rU ,n  .
Gmin FU
0 0, 0̄B,f f =1
B
(15)
Finally, one can obtain the eMBB rate as
  The target rate rU ,n is numerically obtained by imposing
rBorth = log2 1 + Gtar
B,f . (bits/s/Hz) (10)
the requirement PU-NOMA (EU ) ≤ U to (15). Thus, the
sum-rate of the URLLC service is
B. URLLC
nU
1) U-OMA X
rUU-NOMA = rU ,n . (16)
The FU channels available for URLLC are divided in nU
n=1
orthogonal slices with FU0 channels reserved to each UU ,n
user, with n ∈ {1, . . . , nU }. The outage probability of UU ,n , 3) U-RSMA
in the absence of interference from other services, is [32] Either under U-OMA or U-NOMA, URLLC users directly
FU0 transmit their data to the BS once they are active. However,
 
1 X
PU-OMA (EU ) = Pr  0 log2 (1 + σn,f ) < rU ,n  , (11) in U-RSMA, an user may first split its information into two
FU sub-messages, creating the concept of ‘‘virtual users’’. Each
f =1
sub-message has transmission power defined by the so-called
where σn,f , the Signal-to-Interference-plus-Noise Ratio
splitting factor α ∈ [0, 1].
(SINR) of the n-th active user in frequency channel f , equals
As an example, let us consider the case with nU = 2. In this
GU ,n,f , since for the moment there is no interference from
two-user scenario, we assume that only one user, say UU ,1 ,
other users. The target rate rU ,n is numerically obtained by
splits its message,1 creating two virtual users referred to as
imposing the outage probability requirement PU-OMA (EU ) ≤
UU ,1,1 and UU ,1,2 . Without loss of generality, we consider
U to (11). Thus, the sum-rate of the URLLC service is given
that UU ,1,1 is always decoded before UU ,1,2 . In this scenario,
by
we have three possible decoding orders at the BS, namely:
nU
X (i) UU ,1,1 → UU ,2 → UU ,1,2 ; (ii) UU ,1,1 → UU ,1,2 →
rUU-OMA = rU ,n . (12) UU ,2 ; and (iii) UU ,2 → UU ,1,1 → UU ,1,2 , such that the
n=1 proper decoding order is chosen based on the sum of mutual
information from (13), similarly to U-NOMA.
2) U-NOMA
While the decoding orders (ii) and (iii) achieve the same
In U-NOMA, URLLC users share the FU channels available
results of U-NOMA with UU ,1 → UU ,2 and UU ,2 → UU ,1 ,
in each mini-slot and the BS performs SIC to decode the
respectively [48], is has been shown that (i) represents the
multiple messages, which outperforms other techniques of
optimal decoding order of RSMA [49]. Thus, in the SIC
multi-user detection, such as puncturing and erasure decod-
process, the receiver first attempts to decode a (virtual) user
ing [32], and is a general receiver structure for non-orthogonal
while regarding all the remaining messages as noise. Once the
uplink [10]. As an user occupies more than one channel,
decoding is successful, its interference is removed out of the
we cannot simply define the decoding order in terms of the
superimposed received signal, and the receiver then attempts
channel gain magnitude. Instead, the BS can order the users
to decode the next message following the pre-established
according to their mutual information [38]
decoding order. Upon adopting the decoding order from (i),
FU
X
Isum
n = log2 (1 + σn,f ), (13) 1 Following [16], only one out of the two users needs to split its message
f =1 in order to achieve the capacity region.

163182 VOLUME 9, 2021

E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

the SINR of the virtual user UU ,1,1 is connected (SU = 0) in that particular TS; or (ii) there are
α GU ,1,1,f URLLC transmissions (SU > 0), but they were decoded and
σ1,1,f = . (17) removed from the signal by the SIC decoder. In case (ii),
1 + GU ,2,f + (1 − α)GU ,1,2,f
either all URLLC messages are properly decoded (event ĒU )
If UU ,1,1 is correctly decoded and canceled from the or they are all incorrectly decoded (event EU ), since interfer-
received signal, the SINR of UU ,2 becomes ence from eMBB users are constant over all mini-slots. Thus,
GU ,2,f the eMBB outage probability in the NOMA scenario depends
σ2,f = . (18) on whether it is subjected to interference of URLLC service
1 + (1 − α)GU ,1,2,f or not, i.e.
Finally, the SINR of the remaining virtual user UU ,1,2 ,
PB = Pr(SU = 0) Pr(EB |SU = 0)
subject to the correct decoding of the previous users, is
+ Pr(SU > 0) Pr(EU |SU > 0) Pr(EB |EU , SU > 0)
σ1,2,f = (1 − α)GU ,1,2,f . (19)
+ Pr(ĒU |SU > 0) Pr(EB |ĒU , SU > 0) ,


(22)
Then, the achievable rates of U-RSMA can be calculated where EB is the event of eMBB not being correctly decoded
from (15), by substituting σn,f with the SINRs of U-RMSA and SU ∼ Bin(nU S, aU ) is a random variable that represents
presented in (17)-(19). The final rate of user UU ,1 is rU ,1 = the number of URLLC transmissions during the TS. The only
rU ,1,1 + rU ,1,2 . Thus, the sum-rate of the two-user U-RSMA source of outage for eMBB when there is no URLLC signal
URLLC service finally obtained as interfering is when the SNR value is below the threshold
rUU-RSMA = rU ,1 + rU ,2 . (20) SNR (GminB,f ), which implies that the term Pr(EB |SU = 0)
from (22) equals the outage probability for the orthogonal
case 1 − aB , where aB = exp[−Gmin B,f /0̄B ] for simplifica-
It is worthy mentioning that, when compared to U-NOMA,
U-RSMA requires an extra round in the SIC procedure, tion purposes. Moreover, we also consider a simplified and
increasing the complexity of the decoding process. worst case scenario where the eMBB user is in outage when
the URLLC message is incorrectly decoded, i.e., such that
C. ORTHOGONAL NETWORK SLICING Pr(EB |EU , SU > 0) = 1. Besides that, the correct decoding
In Sections III-A and III-B we present, respectively, the and subtraction of URLLC signal has the same performance
achievable rates of eMBB and URLLC services when effect of the case when URLLC is not transmitting, thus,
operating in standalone mode, without slicing the network Pr(EB |ĒU , SU > 0) = Pr(EB |SU = 0) = 1 − aB . Under
resources. When such slicing between the heterogeneous these assumptions,
eMBB and URLLC services is designed in a orthogonal
fashion, they are ‘‘isolated’’ from each other, thus for URLLC PB ≤ (1 − aU )nU S (1 − aB )
h i
+ 1 − (1 − aU )nU S U + (1 − U )(1 − aB ) .


the only source of interference are the nU users active with (23)
probability aU in certain mini-slot occupying all FU ≤ F
channels, whereas eMBB experiences an interference-free By imposing the eMBB reliability constraint PB ≤ B , one
scenario since users are allocated orthogonally within the can rewrite (23) as
remaining FB = F − FU channels. The OMA performance is 1 − B
aB ≥ . (24)
measured in terms of the sum-rate pair (rBsum , rUsum ), where 1 − U

1 − (1 − aU )nU S
rBsum can be defined as [32]
Having in mind that aB = exp[−Gmin B,f /0̄B ], it is possible
rBsum = FB rBorth , (21) to isolate the threshold SNR Gmin from (24), resulting in
B,f
where rBorth comes from (10) and rUsum is computed as pre- !
min 1 − B
sented in Section III-B for each particular multiple access GB,f ≤ −0̄B ln  . (25)
1 − U 1 − (1 − aU )nU S

method adopted by the URLLC service.
The target SNR Gtar
B,f is obtained similarly to (9) as
D. NON-ORTHOGONAL NETWORK SLICING
In non-orthogonal slicing, eMBB and URLLC services 0̄B
simultaneously share all the F available channels, i.e., FB = Gtar
B,f ≤  . (26)
Gmin
FU = F. Due to latency and reliability constraints, it is 0 0, 0̄B,f
B
assumed that the BS always attempts to decode the nU
active URLLC devices first, through SIC, while treating However, in the non-orthogonal case, Gmin B,f is bounded
the eMBB traffic as interference. Therefore, the interfer- by (25). Therefore, the maximum achievable rate of an eMBB
ence from URLLC transmissions into eMBB (and vice-versa) device in NOMA is rBn-orth = log2 (1 + Gtar
B,f ).
needs to be considered. The threshold from (25) indicates that the impact of
An eMBB message would not be affected by URLLC URLLC transmissions in the eMBB decoding should be min-
interference in two cases: (i) there are no URLLC devices imal, due to the fact that, by definition, U  B , which

VOLUME 9, 2021 163183

 E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

implies that aB is close to 1−B . On the other hand, the eMBB TABLE 2. Simulation parameters.
interference in the URLLC traffic is supposed to be more
critical, since URLLC is decoded prior to eMBB. As in [32]
the outage probability of URLLC under NOMA is

PNOMA (EU )
 
FU
!
1 X σn,f
= Pr  log2 1 + < rU ,n  , (27)
FU
f =1
1 + Gtar
B,f

where it is assumed that the interference of eMBB is always

present in the URLLC decoding. The value of σn,f depends on
the multiple access technique used by URLLC users, as dis-
cussed in Section III-B. The URLLC achievable sum-rate
rUsum is then numerically obtained by imposing the reliability
constraint PNOMA (EU ) ≤ U , where the rates are separately
calculated for all nU transmitting URLLC users.

IV. NUMERICAL RESULTS

In this section, we present some numerical results aiming at
comparing the sum-rate performance of U-OMA, U-NOMA
and U-RSMA under both OMA and NOMA network slic-
ing strategies. These results were generated using Monte
Carlo simulations in MATLAB, where, for each particular
scenario, we average a number of 107 independent random
runs. Herein, we consider only the case of nU = 2, as having
several SIC iterations would probably violate the latency
constraint of a URLLC service. In U-RSMA, user UU ,1
splits its transmission according to α (which is optimized
in each simulation step), creating two virtual users, namely
FIGURE 2. Sum-rate region in OMA and NOMA scenarios with URLLC
UU ,1,1 and UU ,1,2 . Furthermore, users that belong to the same under U-OMA, U-NOMA, and U-RSMA schemes. 0̄U = 20 dB, 0̄B = 10 dB,
service have the same average SNR, since we consider they U-OMA = 10−5 ,  U-NOMA =  U-RSMA = 5 × 10−6 ,  = 10−3 , a = 1,
U U U B U
are running identical applications. We consider that in each F = 8, S = 5, nU = 2, and optimized α.

mini-slot there are always two URLLC users connected, i.e.,

aU = 1 for each one of them, thus FU0 = FU /2. Also, that, U-RSMA achieves higher rates, presenting almost the
the number of eMBB users is FB , equaling the number of same results of U-NOMA for high rBsum values.
channels available for the service. Moreover, one TS is com- Fig. 3 shows the URLLC sum-rate for different values of
posed by S = 5 mini-slots and the bandwidth is divided into power splitting factor α. Note that, as expected, in U-OMA
F = 8 channels. The reliability requirement of eMBB service and U-NOMA we obtain constant values, since there is no
is B = 10−3 . For URLLC under U-OMA, the reliability is message splitting. For U-RSMA, on the other hand, it is
UU-OMA = 10−5 , however, as for U-NOMA and U-RSMA the possible to observe that, as α increases, rUsum also increases,
receiver employs SIC, we follow [39] and set the reliability reaching the highest value when α = 0.8 for NOMA and
target as UU-NOMA = UU-RSMA = 5 × 10−6 to ensure that α ≈ 0.75 for OMA slicing.
the overall reliability does not exceed 10−5 . Unless stated The rates of users UU ,1 and UU ,2 when operating under
otherwise, we set 0̄U = 20 dB and 0̄B = 10. Table 2 U-RSMA are presented in Fig. 4, for both OMA and NOMA
summarizes the simulation parameters. slicing. We see that UU ,1 , the user that performs rate split-
In Fig. 2 we plot the sum-rate pair (rBsum , rUsum ) for ting, is capable of reaching higher rates when compared to
OMA and NOMA network slicing with URLLC operating UU ,2 . Also, NOMA slicing is the best choice for this setup,
under U-OMA, U-NOMA, and U-RSMA schemes. Compar- achieving higher rates.
ing the NOMA slicing curves, U-OMA presents the high- We consider that, during one TS, each eMBB user has
est rate pair values until rBsum ≈ 7 bits/s/Hz, from where the same target rate, since the channel gain is constant dur-
U-RSMA outperforms the other methods. Interestingly, rUsum ing this period over all channels. However, for URLLC,
remains almost constant as we increase rBsum in U-RSMA not imposing this requirement is beneficial, since different
and U-NOMA, making these methods a good option when decoding orders provided by U-RSMA enable UU ,1 to reach
we want to achieve greater eMBB rates. In OMA slicing, higher rates, contributing to leverages the overall sum-rate,
U-OMA is the best method until rBsum ≈ 2.3 bits/s/Hz, after as shown in Figs. 5(c) and 5(d), where we plot the URLLC

163184 VOLUME 9, 2021

E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

FIGURE 3. URLLC sum-rate under U-OMA, U-NOMA, and U-RSMA

schemes in OMA and NOMA scenarios for α ∈ {0, . . . , 1}, 0̄U = 20 dB,
U-OMA = 10−5 ,  U-NOMA =  U-RSMA = 5 × 10−6 ,  = 10−3 ,
0̄B = 10 dB, U U U B
aU = 1, F = 8 (FU = FB = 4 in OMA), S = 5, and nU = 2.

FIGURE 5. URLLC sum-rate and per-user rate under U-OMA, U-NOMA, and
U-RSMA schemes in OMA and NOMA scenarios for 0̄U ∈ {0, . . . , 20} dB,
U-OMA = 10−5 ,  U-NOMA =  U-RSMA = 5 × 10−6 ,  = 10−3 ,
0̄B = 10 dB, U U U B
aU = 1, F = 8 (FU = FB = 4 in OMA), S = 5, nU = 2, and optimized α.

FIGURE 4. URLLC per-user-rate under U-RSMA in OMA and NOMA slicing

U-OMA = 10−5 ,
for α ∈ {0, . . . , 1}, 0̄U = 20 dB, 0̄B = 10 dB, U
U-NOMA =  U-RSMA = 5 × 10−6 ,  = 10−3 , a = 1, F = 8 (F = F = 4 in
U U B U U B
OMA), S = 5, and nU = 2.

per-user rate for 0̄U ∈ {0, . . . , 20} dB. Comparing U-RSMA FIGURE 6. URLLC sum-rate under U-OMA, U-NOMA, and U-RSMA
schemes in OMA and NOMA scenario for F ∈ {1, . . . , 12}, 0̄U = 20 dB,
and U-NOMA sum-rates in Figs. 5(a) and 5(b), we see that U-OMA = 10−5 ,  U-NOMA =  U-RSMA = 5 × 10−6 ,  = 10−3 ,
0̄B = 10 dB, U U U B
the former is capable of operating with less performance aU = 1, S = 5, nU = 2, and optimized α.
degradation as the SNR increases, due to the fact that it
is capable of handling the interference better, while the V. FINAL COMMENTS
latter saturates as the SIC procedure fails to eliminate the In this paper, we considered the problem of radio resource
interference. slicing between eMBB and multiple URLLC devices.
From Fig. 6, considering the case of NOMA slicing, We evaluated the sum-rate performance of three multiple
we conclude that U-OMA needs more bandwidth to outper- access methods for URLLC, namely U-OMA, U-NOMA,
form other methods, which is a limiting factor. Moreover, and U-RSMA, when operating under both OMA and NOMA
U-RSMA is the better choice for smaller chunks of spectrum, network slicing strategies. Our results show that U-RSMA
resulting in higher spectral efficiency since we can transmit is capable of achieving higher rates when the power split-
more data with less bandwidth. In OMA, U-RSMA is better ting factor is properly configured, even with strict reliabil-
than other methods in all the evaluated range. ity requirements. Moreover, we show that non-orthogonal

VOLUME 9, 2021 163185

 E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

network slicing is capable of reaching the highest pair of rates [16] B. Rimoldi and R. Urbanke, ‘‘A rate-splitting approach to the Gaus-
for URLLC and eMBB simultaneously. This leads us to show sian multiple-access channel,’’ IEEE Trans. Inf. Theory, vol. 42, no. 2,
pp. 364–375, Mar. 1996.
another interesting scenario in which combining U-RSMA [17] D. Tse and P. Viswanath, Fundamentals of Wireless Communication.
and NOMA is a powerful tool for attending 6G demands. Cambridge, U.K.: Cambridge Univ. Press, 2005.
The practical implementation of RSMA is still evolving, how- [18] B. Clerckx, H. Joudeh, C. Hao, M. Dai, and B. Rassouli, ‘‘Rate splitting
for MIMO wireless networks: A promising PHY-layer strategy for LTE
ever, for our scenario, limiting the number of users in each evolution,’’ IEEE Commun. Mag., vol. 54, no. 5, pp. 98–105, May 2016.
mini-slot is a good strategy to reduce the decoding complexity [19] Y. Mao, B. Clerckx, and V. O. K. Li, ‘‘Energy efficiency of rate-splitting
and delay. Also, it is necessary to add a few bits of informa- multiple access, and performance benefits over SDMA and NOMA,’’
in Proc. 15th Int. Symp. Wireless Commun. Syst. (ISWCS), Aug. 2018,
tion in the message intended to the user that splits its data, pp. 1–5.
sent in the synchronization slot to set the power allocation [20] J. Cao and E. M. Yeh, ‘‘Asymptotically optimal multiple-access commu-
factor based on the average SNR. As some future research nication via distributed rate splitting,’’ IEEE Trans. Inf. Theory, vol. 53,
topics, the impact of imperfect CSI and applying rate-splitting no. 1, pp. 304–319, Jan. 2007.
[21] H. Joudeh and B. Clerckx, ‘‘Robust transmission in downlink multiuser
among users of different services could be investigated. MISO systems: A rate-splitting approach,’’ IEEE Trans. Signal Process.,
vol. 64, no. 23, pp. 6227–6242, Dec. 2016.
[22] H. Joudeh and B. Clerckx, ‘‘Sum-rate maximization for linearly precoded
REFERENCES downlink multiuser MISO systems with partial CSIT: A rate-splitting
[1] H. Zhang, N. Liu, X. Chu, K. Long, A. Aghvami, and V. C. M. Leung, approach,’’ IEEE Trans. Commun., vol. 64, no. 11, pp. 4847–4861,
‘‘Network slicing based 5G and future mobile networks: Mobility, resource Nov. 2016.
management, and challenges,’’ IEEE Commun. Mag., vol. 55, no. 8, [23] Y. Mao, B. Clerckx, and V. O. K. Li, ‘‘Rate-splitting for multi-user multi-
pp. 138–145, Aug. 2017. antenna wireless information and power transfer,’’ in Proc. IEEE 20th Int.
[2] GSMA Association. An Introduction to Network Slicing. Accessed: 2017. Workshop Signal Process. Adv. Wireless Commun. (SPAWC), Jul. 2019,
[Online]. Available: https://www.gsma.com/futurenetworks/wp-content/ pp. 1–5.
uploads/2017/11/GSMA-An-Introduction-to-Network-Slicing.pdf [24] B. Clerckx, Y. Mao, R. Schober, and H. V. Poor, ‘‘Rate-splitting unifying
[3] X. Li, H. A. Chan, D. Bhamare, L. Gupta, C. Guo, and R. Jain, ‘‘Network SDMA, OMA, NOMA, and multicasting in MISO broadcast channel:
slicing for 5G: Challenges and opportunities,’’ IEEE Internet Comput., A simple two-user rate analysis,’’ IEEE Wireless Commun. Lett., vol. 9,
vol. 21, no. 5, pp. 20–27, Sep. 2017. no. 3, pp. 349–353, Mar. 2020.
[4] A. E. Kalor, R. Guillaume, J. J. Nielsen, A. Mueller, and P. Popovski, ‘‘Net- [25] Z. Yang, M. Chen, W. Saad, W. Xu, and M. Shikh-Bahaei, ‘‘Sum-rate maxi-
work slicing in Industry 4.0 applications: Abstraction methods and end-to- mization of uplink rate splitting multiple access (RSMA) communication,’’
end analysis,’’ IEEE Trans. Ind. Informat., vol. 14, no. 2, pp. 5419–5427, in Proc. IEEE Global Commun. Conf. (GLOBECOM), Dec. 2019, pp. 1–6.
Dec. 2018. [26] Z. Yang, M. Chen, W. Saad, W. Xu, and M. Shikh-Bahaei, ‘‘Sum-rate
[5] F. Tariq, M. R. A. Khandaker, K.-K. Wong, M. A. Imran, M. Bennis, maximization of uplink rate splitting multiple access (RSMA) communi-
and M. Debbah, ‘‘A speculative study on 6G,’’ IEEE Wireless Commun., cation,’’ IEEE Trans. Mobile Comput., early access, Nov. 16, 2020, doi:
vol. 27, no. 4, pp. 118–125, Aug. 2020. 10.1109/TMC.2020.3037374.
[27] Y. Zhu, X. Wang, Z. Zhang, X. Chen, and Y. Chen, ‘‘A rate-splitting non-
[6] N. H. Mahmood, S. Böcker, A. Munari, F. Clazzer, I. Moerman,
orthogonal multiple access scheme for uplink transmission,’’ in Proc. 9th
K. Mikhaylov, O. Lopez, O. S. Park, E. Mercier, H. Bartz, and R. Jäntti,
Int. Conf. Wireless Commun. Signal Process. (WCSP), Oct. 2017, pp. 1–6.
‘‘White paper on critical and massive machine type communication
towards 6G,’’ 2020, arXiv:2004.14146. [28] H. Liu and K. S. Kwak, ‘‘Adaptive rate splitting for uplink non-orthogonal
multiple access systems,’’ in Proc. 11th Int. Conf. Ubiquitous Future Netw.
[7] W. Saad, M. Bennis, and M. Chen, ‘‘A vision of 6G wireless systems:
(ICUFN), Jul. 2019, pp. 158–163.
Applications, trends, technologies, and open research problems,’’ IEEE
[29] H. Liu, T. A. Tsiftsis, K. J. Kim, K. S. Kwak, and H. V. Poor, ‘‘Rate splitting
Netw., vol. 34, no. 3, pp. 134–142, Oct. 2020.
for uplink NOMA with enhanced fairness and outage performance,’’ IEEE
[8] P. Popovski, J. J. Nielsen, C. Stefanovic, E. De Carvalho, E. Strom, Trans. Wireless Commun., vol. 19, no. 7, pp. 4657–4670, Jul. 2020.
K. F. Trillingsgaard, A. S. Bana, D. M. Kim, R. Kotaba, J. Park, and
[30] J. Zeng, T. Lv, W. Ni, R. P. Liu, N. C. Beaulieu, and Y. J. Guo, ‘‘Ensuring
R. B. Sorensen, ‘‘Wireless access for ultra-reliable low-latency commu-
max–min fairness of UL SIMO-NOMA: A rate splitting approach,’’ IEEE
nication: Principles and building blocks,’’ IEEE Netw., vol. 32, no. 2,
Trans. Veh. Technol., vol. 68, no. 11, pp. 11080–11093, Nov. 2019.
pp. 16–23, Mar./Apr. 2018.
[31] O. Abbasi and H. Yanikomeroglu, ‘‘Rate-splitting and NOMA-enabled
[9] H. Chen, R. Abbas, P. Cheng, M. Shirvanimoghaddam, W. Hardjawana, uplink user cooperation,’’ in Proc. IEEE Wireless Commun. Netw. Conf.
W. Bao, Y. Li, and B. Vucetic, ‘‘Ultra-reliable low latency cellular net- Workshops (WCNCW), Mar. 2021, pp. 1–6.
works: Use cases, challenges and approaches,’’ IEEE Commun. Mag.,
[32] P. Popovski, K. F. Trillingsgaard, O. Simeone, and G. Durisi, ‘‘5G wire-
vol. 56, no. 12, pp. 119–125, Dec. 2018.
less network slicing for eMBB, URLLC, and mMTC: A communication-
[10] Y. Liu, Z. Qin, M. Elkashlan, Z. Ding, A. Nallanathan, and L. Hanzo, theoretic view,’’ IEEE Access, vol. 6, pp. 55765–55779, 2018.
‘‘Nonorthogonal multiple access for 5G and beyond,’’ Proc. IEEE, [33] A. Anand, G. De Veciana, and S. Shakkottai, ‘‘Joint scheduling of URLLC
vol. 105, no. 12, pp. 2347–2381, Dec. 2017. and eMBB traffic in 5G wireless networks,’’ in Proc. IEEE Conf. Comput.
[11] S. M. R. Islam, N. Avazov, O. A. Dobre, and K.-S. Kwak, ‘‘Power-domain Commun. (IEEE INFOCOM), Apr. 2018, pp. 1970–1978.
non-orthogonal multiple access (NOMA) in 5G systems: Potentials and [34] R. Abreu, T. Jacobsen, G. Berardinelli, K. Pedersen, N. H. Mahmood,
challenges,’’ IEEE Commun. Surveys Tuts., vol. 19, no. 2, pp. 721–742, I. Z. Kovacs, and P. Mogensen, ‘‘On the multiplexing of broadband traffic
2nd Quart., 2017. and grant-free ultra-reliable communication in uplink,’’ in Proc. IEEE 89th
[12] X. Chen, A. Benjebbour, A. Li, and A. Harada, ‘‘Multi-user proportional Veh. Technol. Conf. (VTC-Spring), Apr. 2019, pp. 1–6.
fair scheduling for uplink non-orthogonal multiple access (NOMA),’’ in [35] M. Alsenwi, N. H. Tran, M. Bennis, A. K. Bairagi, and C. S. Hong,
Proc. IEEE 79th Veh. Technol. Conf. (VTC Spring), May 2014, pp. 1–5. ‘‘EMBB-URLLC resource slicing: A risk-sensitive approach,’’ IEEE Com-
[13] A. Rauniyar, P. Engelstad, and O. N. Österbö, ‘‘An adaptive user pair- mun. Lett., vol. 23, no. 4, pp. 740–743, Apr. 2019.
ing strategy for uplink non-orthogonal multiple access,’’ in Proc. IEEE [36] P. K. Korrai, E. Lagunas, S. K. Sharma, S. Chatzinotas, and B. Ottersten,
31st Annu. Int. Symp. Pers., Indoor Mobile Radio Commun., Aug. 2020, ‘‘Slicing based resource allocation for multiplexing of eMBB and URLLC
pp. 1–7. services in 5G wireless networks,’’ in Proc. IEEE 24th Int. Workshop Com-
[14] M. A. Sedaghat and R. R. Müller, ‘‘On user pairing in uplink NOMA,’’ put. Aided Modeling Design Commun. Links Netw. (CAMAD), Sep. 2019,
IEEE Trans. Wireless Commun., vol. 17, no. 5, pp. 3474–3486, May 2018. pp. 1–5.
[15] I. Azam, M. B. Shahab, and S. Y. Shin, ‘‘User pairing and power allocation [37] R. Kassab, O. Simeone, P. Popovski, and T. Islam, ‘‘Non-orthogonal
for capacity maximization in uplink NOMA,’’ in Proc. 42nd Int. Conf. multiplexing of ultra-reliable and broadband services in fog-radio archi-
Telecommun. Signal Process. (TSP), Jul. 2019, pp. 690–694. tectures,’’ IEEE Access, vol. 7, pp. 13035–13049, 2019.

163186 VOLUME 9, 2021

E. J. dos Santos Jr. et al.: Rate-Splitting Multiple Access for URLLC Uplink in Physical Layer Network Slicing With eMBB

[38] E. N. Tominaga, H. Alves, R. D. Souza, J. L. Rebelatto, and M. Latva-Aho, RICHARD DEMO SOUZA (Senior Member,
‘‘Non-orthogonal multiple access and network slicing: Scalable coexis- IEEE) received the D.Sc. degree in electri-
tence of eMBB and URLLC,’’ in Proc. IEEE 93rd Veh. Technol. Conf. cal engineering from the Federal University
(VTC-Spring), Apr. 2021, pp. 1–6. of Santa Catarina (UFSC), Brazil, in 2003.
[39] O. Dizdar, Y. Mao, Y. Xu, P. Zhu, and B. Clerckx, ‘‘Rate-splitting multiple From 2004 to 2016, he was with the Federal Uni-
access for enhanced URLLC and eMBB in 6G: Invited paper,’’ in Proc. versity of Technology—Paraná (UTFPR), Brazil.
17th Int. Symp. Wireless Commun. Syst. (ISWCS), Sep. 2021, pp. 1–6. Since 2017, he has been with UFSC, where he
[40] D. Tse and P. Viswanath, Fundamentals of Wireless Communication.
is a Professor. His research interests include in
Cambridge, U.K.: Cambridge Univ. Press, 2005.
the areas of wireless communications and signal
[41] H. Ji, S. Park, J. Yeo, Y. Kim, J. Lee, and B. Shim, ‘‘Ultra-reliable and low-
latency communications in 5G downlink: Physical layer aspects,’’ IEEE processing. He was a co-recipient of the 2014
Wireless Commun., vol. 25, no. 3, pp. 124–130, Jun. 2018. IEEE/IFIP Wireless Days Conference Best Paper Award, the supervisor of
[42] K. Takeda, H. Harada, and R. Osawa. (Accessed: Sep. 2021). NR the awarded Best Ph.D. Thesis in Electrical Engineering in Brazil, in 2014,
Physical Layer Specifications in 5G. Accessed: 2019. [Online]. Available: and a co-recipient of the 2016 Research Award from the Cuban Academy of
https://www.nttdocomo.co.jp/english/binary/pdf/corporate/technology/rd/ Sciences. He has served as an Editor or an Associate Editor for the Journal of
technical_journal/bn/vol20_3/vol20_3_007en.pdf Communication and Information Systems (SBrT), the IEEE COMMUNICATIONS
[43] A. Zaidi and Z. Wang. (Accessed: Sep. 2021). Designing for the LETTERS, the IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, and the IEEE
Future the 5G NR Physical Layer. Accessed: 2017. [Online]. Available: INTERNET OF THINGS JOURNAL.
https://www.ericsson.com/en/reports-and-papers/ericsson-technology-
review/articles/designing-for-the-future-the-5g-nr-physical-layer
[44] 3GPP. (Accessed: Sep. 2021). 5G, NR, Physical Layer Procedures
for Data. Accessed: 2018. [Online]. Available: https://www.etsi.org/
deliver/etsi_ts/138200_138299/138214/15.02.00_60/ts_138214v150200p.
pdf
[45] A. A. Zaidi, R. Baldemair, M. Andersson, S. Faxér, V. Molés-Cases, and
Z. Wang, ‘‘Designing for the future: The 5G NR physical layer,’’ Eric-
sson Technol. Rev. Articles, Stockholm, Sweden, Tech. Rep., Jul. 2017.
[Online]. Available: https://www.ericsson.com/en/ericsson-technology-
review/archive/2017/designing-for-the-future-the-5g-nr-physical-layer
[46] W. Yang, G. Durisi, T. Koch, and Y. Polyanskiy, ‘‘Quasi-static multiple-
antenna fading channels at finite blocklength,’’ IEEE Trans. Inf. Theory,
vol. 60, no. 7, pp. 4232–4265, Jun. 2014.
[47] G. Caire, G. Taricco, and E. Biglieri, ‘‘Optimum power control over
fading channels,’’ IEEE Trans. Inf. Theory, vol. 45, no. 5, pp. 1468–1489,
Jul. 1999.
[48] H. Liu, T. A. Tsiftsis, K. J. Kim, K. S. Kwak, and H. V. Poor, ‘‘Rate splitting
for uplink NOMA with enhanced fairness and outage performance,’’ IEEE JOÃO LUIZ REBELATTO (Senior Member,
Trans. Wireless Commun., vol. 19, no. 7, pp. 4657–4670, Jul. 2020.
IEEE) was born in Lapa, State of Paraná,
[49] Z. Yang, M. Chen, W. Saad, W. Xu, and M. Shikh-Bahaei, ‘‘Sum-rate
maximization of uplink rate splitting multiple access (RSMA) communi-
Brazil. He received the B.Sc. degree in electri-
cation,’’ IEEE Trans. Mobile Comput., early access, Nov. 16, 2020, doi: cal engineering from the Federal University of
10.1109/TMC.2020.3037374. Technology—Paraná (UTFPR), Curitiba, Brazil,
in 2006, and the D.Sc. degree in electrical engi-
neering from the Federal University of Santa
ELÇO JOÃO DOS SANTOS JR. was born in Catarina (UFSC), Florianópolis, Brazil, in 2010.
Florianópolis, State of Santa Catarina, Brazil, From 2009 to 2010, he was a visiting Ph.D. student
in 1994. He received the B.Sc. degree in elec- with The University of Sydney, NSW, Australia.
tronics engineering and the M.Sc. degree in elec- From 2011 to 2012, he held a postdoctoral position at UFSC. Since 2011,
trical engineering from the Federal University of he has been with the Department of Electronics, UTFPR, where he is
Santa Catarina (UFSC), Brazil, in 2019 and 2020, currently an Associate Professor. His research interests include in the area
respectively. He is currently a Researcher at the of coding and information theory, with applications to wireless communi-
SENAI Institute of Innovation in Embedded Sys- cations systems. He was a co-recipient of the 2011 Brazilian Symposium
tems (ISI-SE) and a Ph.D. student at UFSC. His on Telecommunications (SBrT) Best Paper Award, the 2014 IEEE/IFIP
research interests include wireless communica- Wireless Days Conference Best Paper Award, and the 2016 Research Award
tions, multiple-access strategies, URLLC, random access protocols, and from the Cuban Academy of Sciences.
signal processing.

VOLUME 9, 2021 163187