Enhanced Power Choice Barring Scheme For Massive MTCs With Grant-Free NOMA

Liang Wu,Xiaorui Tang,Zaichen Zhang,*,Jian Dang

1 National Mobile Communications Research Laboratory,Southeast University,Nanjing 210096,China

2 Purple Mountain Laboratories,Nanjing 211111,China

Abstract:Massive machine-type communication(mMTC)is a typical application scenario of the fifth generation(5G)mobile communications.To keep the mMTC reliable and minimize the energy consumption of the mMTC devices,this paper proposes an enhanced power choice barring(EPCB)scheme based on the distributed layered grant-free non-orthogonal multiple access(NOMA)framework,where the cell is divided into different layers according to a predetermined power levels.The proposed EPCB scheme not only combines the grant-free strategy with the sleep mode to reduce the energy consumption,but also designs a power level choosing strategy to increase the access success probability.Simulation results show that when compared with existing schemes,the proposed EPCB scheme has better performance in the aspects of the access success probability and energy efficiency.

Keywords:grant-free;massive machine-type communication;NOMA;power level choosing

I.INTRODUCTION

With the rapid development of the fifth generation(5G)mobile communication and the internet of things(IoT)technology[1],massive machine-type communication(mMTC)has gained more and more attention,and will be widely used in smart grids[2],ehealth[3],internet of vehicles(IoV)[4],etc.MMTC is different from human-to-human communications.The key challenge of mMTC is that the number of mMTC devices(mMTCDs)is quite large such that it may cause a lot of collisions during data transmissions[5—7].Furthermore,because mMTCDs are usually deployed in the dangerous places,where stable energy can not be provided,mMTCDs are required to be energy efficient.Besides,mMTC system has the properties of small data packet and low latency[8,9].

To address these challenges,many researches have been carried out in both physical layer(PL)and data link layer(DLL).The PL mainly specifies the physical links and transmission methods to ensure reliable data transmissions.By employing edge computing to support the implementation of Industrial Internet of Things,a learning-based channel selection framework with service reliability awareness,energy awareness,backlog awareness,and conflict awareness was proposed in[10].Non-orthogonal multiple access(NOMA)technologies are also being investigated in PL,including power-domain NOMA(PD-NOMA)[11—14]and code-domain NOMA(CDNOMA)[15].For the PD-NOMA,mMTCDs transmit information with different power levels in the same time-frequency resource,and multiuser signals are detected and separated by employing successive interference cancellation(SIC)at the base station(BS)during the uplink transmission[16].Sparse-code multiple access(SCMA)[17],pattern-division multiple access(PDMA)[18,19]and multi-user shared access(MUSA)[20]are CD-NOMA schemes.Both PDNOMA and CD-NOMA exploit the degrees of freedom for channel sharing such that more mMTCDs are able to get a unique resource to achieve a high access success probability.Besides,NOMA schemes can be applied both in downlink and uplink transmissions[21,22].

The DLL is the protocol layer that transfers data between nodes on a network segment across the PL.It determines the access mode,and is responsible for the error control,flow control,and link management.The access schemes in the DLL are divided into three categories including contention based scheme,contentionfree scheme,and grant-free scheme[23,24].The contention based schemes including carrier sense multiple access(CSMA)and ALOHA can save a lot of frequency-time resources and are suitable for dynamic networks.However,in the mMTC scenario,the contention based schemes will cause a lot of collisions and high energy consumption because of numerous mMTCDs and the retransmission mechanism.For the contention-free scheme,which includes time division multiple access(TDMA)and frequency division multiple access(FDMA),the system assigns a fixed resource to each mMTCD such that it can guarantee the high access success probability.However,it requires a large number of communication resources and the resource allocation process consumes a lot of energy.Both the contention based scheme and the contentionfree scheme are the request-grant scheme,which is designed to provide reliable access for a small number of users with large data packets.Different from request-grant schemes,grant-free schemes allow users to transmit information without the need of the BS to grant them communication resources,which simplifies access process such that a lot of energy can be saved.However,the transmission without authorization leads to a low access success probability,which can be improved with the PL technologies such as the PD-NOMA technology.Power level choosing strategy is critical in the PD-NOMA.On scheme is that the BS calculates the power allocation strategy with full channel state information(CSI),and then sends the power level choosing strategy to mMTCDs for uplink transmissions[22,25].A distributed NOMA scheme was proposed in[26]to improve the throughput,and the mMTCDs chooses the power level randomly.Based on the distributed NOMA,[27]proposed a low-complexity distributed layered grant-free NOMA framework,where the mMTCD calculates its transmission power independently based on its location.When there are both grant-free users and grantbased users,a semi-grant-free NOMA scheme was proposed to enhance network connectivity and spectral efficiency[28].

When the channel resource is limited,grant-free NOMA can reduce signaling overhead and transmission delay,and support a large number of random access mMTCDs.Besides,grant-free NOMA is suitable for future energy-efficient radio access networks[29].Grant-free PD-NOMA emerges as a promising technology to enable massive access by accommodating multiple users on one resource block[27,30—32].However,the collisions within each resource block are still severe since users randomly select resource blocks with no knowledge of the contention status at each resource block[30].

In order to improve the access success probability and reduce the energy consumption of the grantfree NOMA,this paper proposes an enhanced power choice barring(EPCB)scheme based on the distributed layered grant-free NOMA[27],which is a grant-free scheme by combing the spatial domain resource with the power domain resource.The proposed EPCB scheme can improve the access success probability and reduce the energy consumption at the same time by optimizing the power level of each mMTCD.According to the theoretical analysis and the simulation results,the proposed EPCB scheme is suitable for the mMTC and performs better than reference schemes.The main contributions of this paper are summarized as follows.

·This paper proposes an EPCB scheme based on the distributed layered grant-free NOMA framework to improve the access success probability and minimize the energy consumption of mMTCs.

·In the proposed EPCB scheme,the power level choosing strategy is not only based on the locations of mMTCDs,but also based on a predetermined probability,which permits the mMTCDs located in the low layers to transmit signals based on the energy requirements of the high-power levels.

·For ease of implementation and theoretical analysis,the numbers of mMTCDs in different power levels are designed as an arithmetic progression,which is verified as an efficient strategy through simulations.

·Theoretical analysis and simulation results of the proposed EPCB scheme in terms of the access success probability and energy consumption are provided in detail.

The rest of the paper is organized as follow.In Section II,we introduce the system model and the traditional distributed layered grant-free NOMA.In Section III,the proposed EPCB scheme is presented in detail.Section IV presents the theoretical analysis of the proposed EPCB scheme.Simulation results of the access success probability and the energy consumption are provided in Section V.Section VI concludes this paper.

II.TRADITIONAL DISTRIBUTED LAYERED GRANT-FREE NOMA

As shown in figure 1,there is a BS andNmMTCDs in a single cell,the radius of which isR.The BS is located in the center of the cell.It is assumed that the mMTCDs are uniformly distributed in the cell.In addition,the mMTC system setsLreceived power levels under the framework of the distributed layered NOMA[27].The predetermined received power of different power levels are denoted byvl,(l= 1,2,...,L),respectively.The relationship amongvlisv1＞v2＞...＞vL ＞0.Note that in the power level division strategy,high power level has a large power level indexl,but with a low receive powervl.Therefore,the cell is divided intoLlayers,which are the concentric rings.The nearby users located in the low layers and the distant users located in the high layers.Each concentric ring has the same coverage area to ensure the equal number of mMTCDs in each layer.Therefore,the outside radius of thelth layer,Rl,satisfies that[27]

If the distance between thenth mMTCD and the BS,dn,satisfies thatdn ∈(Rl－1,Rl],thenth mMTCD belongs to thelth layer.

When the BS is equipped withMTantennas,the channel from thenth mMTCD to the BS is hn,which is aMT ×1 vector.The received signal at the BS is expressed as

wherexnis the transmit signal of thenth mMTCD,and w is the noise vector,each element of which is a Gaussian random variable with zero mean and varianceσ2w.For thenth user,the unitary detection vector is defined as gn,that is,|gn|= 1.Therefore,the processed received signal for the first mMTCD is given by

where(·)Hstands for conjugate transpose.It is assumed that

Since there areLpower levels,and it is assumed thatLmMTCDs employ NOMA in one time slot.For thenth mMTCD,when the perviousn-1 mMTCDs are decoded correctly by employing SIC strategy,the received signal to interference-plus-noise ratio(SINR)is expressed as

In addition to the power-domain resource,the timedomain resource is also utilized to increase the degree of freedom of the communication system.The channel is divided into serval frames,and each frame is evenly divided intoMtime slots.The resource block division strategy is shown in figure 2.

The access process of the traditional distributed layered grant-free NOMA is shown in figure 3.When the mMTCD enters the network at the first time,it gets the broadcast packet,which is sent from the BS in the broadcast channel and contains the information of the current network such as the number of the power levelsL,the number of the time slotsM,and the SINR threshold Γ.After receiving and decoding the broadcast packet,the mMTCD will set the parameters correspondingly and determine the power level based on its location,which is calculated according to the power of the received broadcast packet.When a mMTCD needs to send information to the BS,it will transform from sleep mode to the active mode[35].Then,the mMTCD calculates the transmission powerbased on(9)and its location.After that,the mMTCD chooses a time slot randomly to transmit information without the retransmission mechanism.Once the information is sent,the mMTCD is back to the sleep mode.In this scheme,the mMTCD stays in the sleep mode most of the time to save energy.Furthermore,the mMTCD sends information with no authorization and retransmission mechanism to simplify the access process and reduce the energy consumption.

III.PROPOSED EPCB SCHEME FOR MMTC

In the traditional distributed layered grant-free NOMA scheme,the choice of the power level is only determined by the location of the mMTCD.According to(9),if the mMTCD choses a power level greater than that determined by its location,the energy consumption will decrease,and it will be more energy efficient.However,the access success probability will decrease.Therefore,this paper proposes the EPCB scheme to attain a high energy efficiency and a high access success

Table 1.The increased number of each power level when L is even.

Table 2.The increased number of each power level when L is odd.

The access process of the proposed EPCB scheme is shown in figure 5.The major difference between the proposed EPCB scheme and the traditional distributed layered grant-free NOMA scheme lies in the power level choosing strategy.In the proposed scheme,by permitting that the mMTCDs located in the low layers transmit signal based on the energy requirements of the high power levels with a predetermined probability,the energy consumption will reduce according to(9).Besides,there are fewer mMTCDs in the lowpower levels.The collision of the mMTCDs at the low power levels will decrease.It facilitates SIC process because the low-power level mMTCDs are detected first in the SIC process.Therefore,the energy efficiency and the access success probability of the proposed EPCB scheme increase.

IV.THEORETICAL ANALYSIS OF PROPOSED SCHEME

4.1 Analysis of Access Success Probability

The access success probability,which can reflect the reliability of the network,is defined as the probability that the mMTCD is decoded successfully at the BS.There are two main factors that affect the performance of the access success probability.The first factor is the power level collision.When multiple mMTCDs,which are active at the same time slot,choose the same power level,the collision occurs and the signals can not be decoded correctly.The second factor is related to the SIC in NOMA.The process of SIC is not an independent event and the signal decoded first will affect the success probability of the subsequent decoding.For example,in a time slot,if the power level collision happens at thelth power level,the decoding of the signals at levels 1,2,...,l -1 will not be affected,but the signals at levelsl+1,l+2,...,Lcan not be decoded successfully,even if these high power level mMTCDs have no power level collision.Next,we will analyze the access success probability of the proposed EPCB scheme.

For thelth power level,the probability that only thenlth mMTCD is active at themth time slot is given by

According to the above theoretical analysis,we can find that the factors affect the access success probability including the number of the time slotsMand the number of the power levelsL.In the theoretical analysis of the access success probability,it is assumed that the mMTCDs of previous power levels are decoded correctly,which is guaranteed by the preset SINR threshold.Therefore,the analytical result of the access success probability in the high SNR region will be more accurate.Besides,in the high SNR region,the number of power levels can be increased,and the access success probability will increase too.If the desired signal power and the interference signal power increase with the same proportion in the high SNR region,the SINR keeps the same,and the access success probability tends to be stable.

4.2 Analysis of Energy Consumption

The energy consumption is also an important issue for the mMTC system.The proposed EPCB scheme is a grant-free scheme combining with the sleep mode control.The energy consumption is calculated based on the transmission power.

According to(9),the average energy consumption in thelth power level is given by

V.SIMULATION RESULTS

Simulations are carried out to evaluate the access success probability and the energy consumption of different schemes.The simulation results are calculated by Monte Carlo method with different realizations of the mMTCD distribution.The simulation parameters are provided in the table 3[26,27],and it is assumed that the BS is equipped with one antenna.Besides,R2/G0is normalized to be one in the simulation[26].

Figure 6 shows the access success probability.In the simulation,when the imperfect SIC occurs,the access of the mMTCD is not successful.It can be seen from figure 6(a)that the access success probability decreases,when the index of the power levels increases.That is because the mMTCDs of the higher power levels are affected by the mMTCDs of the lower power levels.When the number of power levels increases,the access success probability increases too.In addition,compared with the distributed layered grant-free NOMA scheme in[27],the proposed EPCB scheme performs better,especially at the low power levels.WhenL= 20 and the index of power level is greater than 18,the scheme in[27]achieves higher success probability than the proposed scheme.The reason is that there are more mMTCDs transmitting signals based on the high power levels in the proposed scheme.Figure 6(b)shows that the system access success probability of the proposed EPCB scheme is higher than that of the grant-free NOMA scheme in[26]and the distributed layered grant-free NOMA scheme in[27].The access success probability of the proposed EPCB scheme is increased about 5%when compared with the distributed layered grantfree NOMA scheme in[27].It can be seen from figure 6(b)that when the numbers of the power levels increases,the access success probability increases.That is because the average number of mMTCDs in each power level decreases,and the collision probability also decreases.

Figure 1.Cellular architecture of the distributed layered NOMA.

Figure 2.Resource block division strategy.

Figure 3.Access process of the traditional distributed layered grant-free NOMA scheme[27].

Figure 4.The choice of power levels in the proposed EPCB scheme.

Figure 5.Access process of the proposed EPCB scheme.

Figure 6.(a)Access success probability of different schemes versus the index of the power levels with different values of L and M = 100;(b)System access success probability of different schemes versus the number of power levels L,where M =100.

WhenL= 10,the relationship between the access success probability and the number of the time slots is shown in figure 7.It can be seen from figure 7(a)thatthe access success probability increases with the increase of the number of the time slots.That is because the probability of the power level collision decreases with the increase of the number of the available time slots.Figure 7(b)shows that with different number of the time slots,the system access success probability of the proposed EPCB scheme is always better than those of the reference schemes.

Figure 7.(a)Access success probability of different schemes at each power level with different values of M and L = 10;(b)System access success probability of different schemes versus the number of time slots M,where L=10.

Figure 8.(a)Average energy consumption in each level of different schemes with different values of L and M = 200;(b)System average energy consumption of different schemes with different values of L,where M =200.

Table 3.Simulation parameters.

Note that in figure 6(a)and figure 7(a),whenL= 10 andM= 100,the basic tolerance number is ΔN= 0.In this case,the proposed EPCB scheme is equivalent to the scheme in[27],and they achieve the same performance.

Therefore,the proposed EPCB scheme performs better than the existing schemes in the aspect of the access success probability.In addition,according to the theoretical analysis and simulation results,it can be seen that when the number of the times slots increases,that is there are more times slots,the probability of power level collision decreases so that the access success probability increases.When the number of power level increases,the average number of mMTCDs in each power level decreases,which decreases the probability of power level collision.

Based on the parameters provided in table 3,the simulation of energy consumption is carried out and the results are shown in figure 8,whereNL=180 andM=200.

It can be seen from figure 8(a)that the average energy consumption of the high power level is less than that of the low power level,which can be proved by

When the number of the power levels increases,the energy consumption of each power level increases too.For the first half power levels,the average energy consumption of the proposed EPCB scheme is the same as that of the traditional distributed layered grant-free NOMA scheme.However,for the last half power levels,the average energy consumption of the proposed EPCB scheme is lower than that of the traditional distributed layered grant-free NOMA scheme.The reason is that the number of the mMTCDs,which choose the first half power levels to send information,becomes small in the proposed EPCB scheme.It can be seen from figure 8(b)that the system average energy consumption of the proposed EPCB scheme is lower than that of the reference schemes,and the energy consumption of the proposed EPCB scheme is only about one tenth of the traditional distributed layered grantfree NOMA scheme,which means that the transmit power of the proposed EPCB scheme can be reduced about 10dB.

VI.CONCLUSION

To improve the access success probability and reduce the energy consumption of the mMTC system,this paper has proposed an EPCB scheme based on the distributed layered grant-free NOMA framework.The proposed EPCB scheme uses the power-domain resources to promote the degree of freedom of the communication system,and a new power level choosing strategy has been designed such that more mMTCDs can access to the network.Simulation results show that the access success probability increases about 5%when compared with the traditional distributed layered grant-free NOMA scheme.Besides,the energy consumption of the proposed EPCB scheme is only about one tenth of the traditional distributed layered grantfree NOMA scheme.Through the choice of the number of the power levels,we can get the best trade-off between the access success probability and the energy consumption to meet the requirements of the mMTC system.

ACKNOWLEDGEMENT

This work was supported by the NSFC projects(6217011870,and 61971136),Zhishan Youth Scholar Program of SEU,the Fundamental Research Funds for the Central Universities,and Young Elite Scientist Sponsorship Program by CAST(YESS20160042).