• No results found

Direct Transmission Cellular Network

4.6 Economic Efficiency

4.7.2 System Level Performance

5.2.1.2 Direct Transmission Cellular Network

Here, only the BSs participate in data transmission. The BS will transmit directly to the K UEs of each sector in a TDMA arrangement whereby each UE is allocated a

BS signal RS signal RS–RS cooperative link BS (a) RS1 RS2 (b) B sys Relay transmission Direct transmission tDT Relay phase RS1, RS2 UE1,...,UEL … BSRS1, RS2, UEL Broadcast phase BSRS1, RS2, UE1 BSUE1 … BSUE

TDMA direct transmission

tRT

(1 – τr)tRT τrtRT

L

Direct

G GRelay

Figure 5.2: (a) The relaying structure and (b) the transmission protocol of a RACN employing relay cooperation with M = 2 RSs per sector.

transmission duration of KT.

5.2.2

Power Consumption Model

Let PBS be the RF transmit power allocated to the BS in each cell. Assuming that

PBS is equally allocated among the sectors and the BSs utilise full bandwidth to

transmit, the effective RF transmit power at each sector is Pb = NPBSSec. Full bandwidth

is employed during BS transmission as it was shown in Section4.7.2 to deliver better performance. Also, let PRS be the RF transmit power allocated to each RS. We

assume that equal power is assigned to each sub-channel during relay transmission. Consequently, each RS will utilise a fraction, ηr(0 ≤ ηr ≤ 1), of its allocated RF

transmit power for relaying in its designated sub-channel(s). Thus, the effective RF transmit power of each RS is Pr= ηrPRS.

In modelling the circuit power consumption, we assume that the circuit power con- sumption of the BS and RSs is proportional to Pb and Pr, respectively, as described

in Section2.5.2.1. Let Pc,ref be the circuit power consumption at a given RF transmit

power Pref. Therefore, the circuit power consumption of the BS and RS is defined as

Pc,b =

PbPc,ref

Pref and Pc,r =

PrPc,ref

Pref , respectively.

When measuring total power consumption, we consider the operational power of the system which includes both the RF transmit power and the circuit power consumption of the PA and SP modules, respectively. Considering the aggregate effect of the duplexer/feeder losses and the efficiency of the antenna/amplifier modules, let the effective operational efficiency of the BS and RS be given as αb and αr, respectively.

Therefore, the operational power per sector of a relay transmission is

Pop,relay = (1 − τr) αbPb+ M τrαrPr+ Pc,b+ M Pc,r (5.1)

while the operational power per sector of a direct transmission is given as

A RACN consists of both direct and relay transmissions. Thus, the total operational power per sector of the RACN is

Pop,totalRelay = tD(Pop,direct+ M Pc,r) + tRPop,relay (5.3)

while that of the DTCN is given as

Pop,totalDirect = Pop,direct. (5.4)

From (5.1) and (5.3), we observe that in a relay transmission the circuit power con- sumption of the M RSs is the additional power cost that must be accommodated. This additional power cost can quickly become substantial in a network architecture that employs many transmission nodes.

Energy Consumption Ratio

As described in Section2.5.2.2, the ECR is used as a performance metric for the energy efficiency of a system. It is proportional to the ratio of the average total operational power to the average capacity of the system under consideration. Thus, the ECR is

ECRsys =

EPop,totalsys

Bsys· E {Csys}

(5.5)

where Pop,totalsys can be either (5.3) or (5.4) and Csys is the spectral efficiency of the

system under consideration in bits/s/Hz. Therefore, the ECR has units of J/bit.

5.3

Interference Analysis

When all BSs are actively transmitting at full bandwidth, the set of interference sources X experienced by the RSs during the broadcast phase of the relay transmission and by the UEs during direct transmission are from the BSs transmitting to all sectors of all cells except the base sector, that is, X = {(i, j) |(i, j) ∈ C × S, (i, j) 6= (1, 1) }.

Assuming the interference sources are independent, the interference covariance matrix for JDEC is a block diagonal matrix given as

RJ DECBC = diag (Um|m = 1, · · · , M ) (5.6) where Um = P (i,j)∈X Pb Nb  Hb(i,j),r(1,1,m)HHb(i,j),r(1,1,m) 

. For IDEC, the interference co- variance matrix at the mth RS is given as

R(m)BC = Um (5.7)

while for the kth UE, the interference covariance matrix is given as

R(k)D = X

(i,j)∈X

Pb

Nb

Hb(i,j),u(1,1,k)HHb(i,j),u(1,1,k). (5.8)

When all the RSs are actively transmitting, the interference at the kth UE which is receiving at frequency fu(1,1,k) is from the surrounding RSs, other than the base sector,

that are relaying at frequency fr(i,j,m)= fu(1,1,k). Thus, the set of RSs interfering the

kth UE is Pu(1,1,k) =(i, j, m)

(i, j, m) ∈ X × M, fr(i,j,m) = fu(1,1,k) . Following that,

the interference covariance matrix at the kth UE is given by

R(k)R = X

(i,j,m)∈Pu(1,1,k)

Pr

Nr

Hr(i,j,m),u(1,1,k)HHr(i,j,m),u(1,1,k) . (5.9)