System Model

The system model used in this thesis is described as follows, and it is simplified from the transmitter for uplink physical layer described in 2.1.

2.2.1 Transmitter Model

A K-user asynchronous turbo-coded CDMA system with QPSK modulation is considered.

The transmitter model used in this thesis is shown in Fig. 2-19 where the spreader for the k-th user is shown in Fig. 2-20. Fig. 2-19 is the simplified version of Fig. 2-1 while Fig.

2-20 is simplified from uplink DPDCH/DPCCH in WCDMA systems shown in Fig. 2-12.

The turbo encoder used is described in Fig. 2-4. The information bit stream u_k={u_k[i]}, }

1 , 0 { ] [i ∈

u_k , 30<i≤M_k + , is encoded with the rate-R=1/3 turbo encoder where Mk is the CB size. The internal interleaved version of information bit stream is u^'_k={u_k[i]},

} 1 , 0 { ]

'[i ∈

u_k , 30<i≤M_k + . The parity bits can be generated from encoder E1 or encoder E2 and are defined as u_k^px={u_k^px[i]}, }u_k^px[i]∈{0,1 , 30<i≤M_k + . The resultant coded bit

stream b^'_k={b_k[n′]} at the output of P/S block is as described in 2.1.2. The Mapper maps {0,1}→{-1,1} and the interleaver ∏ is the second interleaver described in 2.1.2 which reorders the coded bits and obtains b_k={b_k[n]} where {b_k[n]}=∏({b_k[n′]}), 0<n≤N_b,k

where N_b,k = 3M_k +n_tail. n =12 is the number of tail bits generated from the encoder. The _tail equivalent complex baseband representation of the transmitted signal of the k-th user at the mobile station is given by

) ( )}

( ) ( )

( ) ( {

)

(t P _dC _, t B t j C _, t A t C t

s_k = _k β _{Od k} + β_{p Op pilot k}

(2-1)

„ Pk is the signal power, and power distribution of all users in the system follows a power distribution ratio (PDR) where PDR=Pk/Pk+1≧1 for 1≤k ≤ K −1 when there are K users in the system . For the sake of justice, total transmit power of all users is K with any given PDR, and PK =K(PDR -1)/|(PDR )^K-1|.

„ βd and βp are the traffic-channel gain and pilot-channel gain, respectively. βd +βp=1, and βc=βp/βd denotes the pilot-to-traffic amplitude ratio.

„ Bk(t) =Σbk[n]pb(t-nTb) are the traffic-channel signals, where bk is the binary data signal taking the values ±1 from the Interleaver output.

„ Apilot(t) =Σapilot[npi]ppi(t-npiTpi) are the uncoded pilot signals modulated at Q -channel and has the same characteristic as Bk(t) but with a symbol period equal to Tpi.

„ Ck(t) =Σck[nc]pc(t-ncTc) =Σ(ck,I[nc]+ck,Q[nc])pc(t-ncTc) are the complex scramble sequences Sdpch in Fig. 2-12, where {ck,I[nc]} and {ck,Q[nc]} are the Gold sequences with period equal to the length of one frame and composed of N(SF) chips described in 2.1.5, where N is the bit number in a frame.

„ CO,x(t) =ΣcO,x[nc]pc(t-ncTc), where cO,x are the OVSF codes as shown in 2.1.5 with period equal to SF, where SF= Tb/Tc, and ΣcO,x1[nc]cO,x2[nc]=0 when x1 ≠x2. In the following, cO = cO,d, CO = CO,d, and cO,p is neglected since it is an all one sequence.

„ pc, pb and ppi are unit power pulses with duration Tc, Tb and Tpi, respectively. SFpilot = Tpi/Tc is the SF for pilot signal, and Fsf = Tpi/Tc.

2.2.2 Fading Channel Model

In the wireless environments, the information-bearing signals encounter fading when they are transmitting through the channel. Small scale fading, or simply fading, is used to describe the rapid fluctuation of the amplitude of a radio signal over a short period of time or distance. The multipath propagation, mobile and surrounding objects movement, and bandwidth of transmitted signal influence the fading. Based on the Doppler spread, the fading channel can be classified as either fast fading or slow fading. When mobile station and other object are moving, the channel characteristics are changing according to the moving speed. If the channel changes so fast that the symbol period is no longer less than the coherence time, this channel is a time selective (fast) fading channel. Based on the multipath

time delay spread, the fading channel can be classified as either flat fading or frequency-selective fading. When the transmission bandwidth is broader than the channel bandwidth, the multipath can be resolved and the channel is called frequency-selective channel. In CDMA systems, the transmission bandwidth is usually wider than the coherence bandwidth. Consequently, the information-bearing signals encounter multipath fading channels. According to the wide-sense stationary uncorrelated scattering (WSSUS) model [56], each delay path p of user k has independent attenuation and phase shift expressed by

)

,_p(t

αk in multipath fading channel. The envelope of the output is Rayleigh distributed while the phase of the output is uniformly distributed. Fig. 2-21 shows the multipath fading channel model with four paths. In this thesis, the fading profile of each path is modeled as Jakes model in which a U-shaped Doppler spectrum with given maximum Doppler frequency fd defines the characteristics of fading. The U-shaped spectrum is defined as

⎪⎩

⎪⎨

⎧ ≤

= −

elsewhere f f f

f f

S ^d

d f f

0 )

( ^{2 2}

2

π σ

where σ²_f is the average power of the faded carrier, and fd is the maximum Doppler frequency caused by the mobile motion where fd =v/λ, v is the vehicular speed and λ is the wavelength of the carrier. To generate fading channel coefficients, in the Jakes’ model [38], the narrowband time-dispersive multipath fading is basically a combination of oscillators with different frequency. The outputs of the oscillators for real part and imaginary part can be described as xc(t) and xs(t), respectively, and are defined as follows.

t t

t

x _D

N

n n n

c

n cosβ cosω 2cosαcosω

2 ) (

1

+

=

∑

=

t t

t

x _D

N n

n n s

n sinβ cosω 2sinαcosω

2 ) (

1

+

=

∑

=

where ωD =2π/fd is the maximum Doppler frequency in radian, Nn is the number of other lower frequency oscillators, Generally speaking, N0 equal to or larger than 8 gives good approximation and ωn =ωDcos(2πn/N) while n=1, 2, …, Nn and N=(2Nn+1)/2. In order to normalize average power to unity, we should set α=0, βn=π/(Nn+1) that let <xc(t)xs(t)>=0 and <xc(t)²>=Nn <xs(t)²>=Nn+1. Thus we have xc(t)and xs(t) as follows.

t N t

t

x _D

N n

n n n

c

n β ω cosαcosω

2 cos 1

2 cos )

(

1

+

=

∑

=

∑

=

= + ^Nn

n

n n n

s t

t N x

1

cos 1 sin

) 2

( β ω

The complex baseband output of the Jakes model is then given by

( )y t_f =x t_c( )+ jx t_s( )

where the output yf(t) is a baseband signal with Rayleigh fading characteristics.

After the transmitted signal ( )s t of the k-th user in (2-1) passing through a P-ray _k Rayleigh fading channel, assuming that the channel variation in a symbol interval is constant, the asynchronous received signal is represented as

) ( ) (

) ( )

(

1

, ,

1

t n t

s t t

r

K

k

p k k p k P

p

+

−

=

∑∑

= =

τ

α

(2-2) where n(t) is the complex AWGN with zero mean and one-sided power spectral density N0. τk,p and αk,p(t) are the delay and the complex channel gain of the k-th user at the p-th path, respectively. τk,p are uniformly distributed random variables in [0, Tb) for asynchronous systems. αk,p(t) is still a zero-mean complex-valued Gaussian random variable without loss of generality when the carrier phase shift part is absorbed in it [88]. For simplicity, we assume

that the τk,p are perfectly estimated for all users. αJ,p(t) are constant in a symbol interval, i.e. first CB of the first frame is the bit and the user of interest.

2.2.3 RAKE Receiver

The matched filter (MF) is the optimal receiver when all interferences are viewed as background noise with Gaussian distributed characteristics. The RAKE receiver is considered as the optimal receiver in frequency-selective fading channel and combines multiple resolvable paths at receiver end to improve the performance [1]. Among combination criteria, the maximal ratio combining (MRC) maximizes the instantaneous signal-to-noise power ratio (SNR). The MRC RAKE for uplink dedicated channel is shown in Fig. 2-22. The received signal r(t) is directly sent into the correlator followed by a summation of all fingers’ outputs. The real part of the RAKE output is given by

⎭⎬ fingers, and in this thesis we assume that F=P. Thus, the denominator ^{( )}_, ²

1

where

In (2-4), it is shown that the second and the fourth interference terms come from the pilot-channel signal of multipath and other users, respectively. The data decision is made by taking sign bit of RAKE output Y^ˆ_J^{( )n} , i.e. sgn{Y^ˆ_J^{( )n}}, where sgn{.} is the sign function.

The conventional (RAKE) receiver is easy to implement. However, MAI from both pilot signal and data signal of other users due to non-orthogonality are viewed as background noise. This results in dramatically reduction in system capacity. Additionally, in (2-4) it is also shown that users with large channel gain result in larger MAI than those with small channel gain. In mobile environments, large channel gain usually comes from users who are located close to the base station. Users with weaker received signal are overwhelmed by users with stronger received signal. This is the so called near-far effect. The system capacity is expected to be larger if the MAI in RAKE output Y^ˆ_J^{( )n} in (2-3) can be eliminated.

Table 2-1 Main WCDMA parameters [33]

Table 2-2 Mapping from zn(i) to cshort,1,n(i) and cshort,2,n(i), i = 0, 1, …, 255

zn(i) cshort,1,n(i) cshort,2,n(i) 0 +1 +1 1 -1 +1

2 -1 -1

3 +1 -1

Table 2-3 UL reference measurement channel (64 kbps)

Parameter Level Unit

Information bit rate 64 Kbps

DPCH 240 Kbps

Power control Off

TFCI On

Repetition 19 %

Physical channel mapping 2nd interleaving CRC attachment

TrBk concatenation/

CB segmentation

Radio frame segmentation 1st interleaving

Radio frame equalization Channel coding

Physical channel segmentation TrCH multiplexing

Rate matching Rate matching

Physical channel mapping 2nd interleaving

PhCH#1 PhCH#2 CCTrCH

...

... ...

Fig. 2-1 TrCH multiplexing structure for uplink

Fig. 2-2 TrCH multiplexing structure for downlink

Output 0 G0 = 557 (octal) Input

D D D D D D D D

Output 1 G1 = 663 (octal) Output 2 G2 = 711 (octal) Output 0 G0 = 561 (octal) Input

D D D D D D D D

Output 1 G₁ = 753 (octal) (a) Rate 1/2 convolutional coder

(b) Rate 1/3 convolutional coder

Fig. 2-3 Rate 1/2 and rate 1/3 convolutional coders

Fig. 2-4 Structure of rate 1/3 Turbo coder (dotted lines apply for trellis termination only)

D C C H U p lin k

D T C H

24 0 k b ps D P D C H T u rb o C o d e R = 1 /3

1 2 • • • • 1 5 1 2 • • • • 1 5 1 2 • • • • 1 5 1 2 • • • • 1 5

R a dio fra m e F N = 4 N + 1 R a dio fra m e F N = 4 N + 2 R a d io fram e F N = 4 N + 3 R a dio fra m e F N = 4 N

In form atio n d ata C R C d e tec tio n

R ate m a tc h in g

2n d in te rlea vin g

2 4 0 0

1 6 0 1 6 0

1 2 1 5

• • • •

1 2 1 5

• • •

•

1 2 1 5

• • • •

1 2

1 6 0 1 5

• • • •

2 2 9 3 1 0 7 2 2 9 3 1 0 7 2 2 9 3 1 0 7 2 2 9 3

# 1 2 2 9 3 # 2 2 2 9 3 #3 2 2 9 3 # 4 2 2 9 3 1 0 7 1 0 7 1 0 7 1 0 7

2 4 0 0 2 4 0 0 2 4 0 0

7 7 4 0 7 7 4 0 2 5 7 6

T erm in a tio n 1 2 C R C 1 6

2 5 6 0 2 5 6 0

1 0 7 3 6 0

3 6 0 1 1 2

T a il8 1 0 0

H e a d e r 1 6

C R C 1 2 p a d d in g M a x. 8 0

1s t in te rle a v in g R a d io F ram e se gm en tatio n

s lot se gm e n ta tio n

C R C d e tec tio n L ay e r 3 L A C h ea d e r,p ad d in g d is c ard

T a il bit dis c a rd

V ite rbid ec o din g R = 1 /3 1 s t in terlev in g

# 1 1 9 3 5 # 2 1 9 3 5 # 3 1 9 3 5 # 4 1 9 3 5 9 0 9 0 9 0 9 0

Fig. 2-5 Channel coding for the UL reference measurement channel (64 kbps)

Dedicated Physical Data Channel (DPDCH) Dedicated Physical Control Channel (DPCCH) Fractional Dedicated Physical Channel (F-DPCH) E-DCH Dedicated Physical Data Channel (E-DPDCH) E-DCH Dedicated Physical Control Channel (E-DPCCH) E-DCH Absolute Grant Channel (E-AGCH)

E-DCH Relative Grant Channel (E-RGCH) E-DCH Hybrid ARQ Indicator Channel (E-HICH) Physical Random Access Channel (PRACH)

Common Pilot Channel (CPICH)

Primary Common Control Physical Channel (P-CCPCH) Secondary Common Control Physical Channel (S-CCPCH)

Synchronisation Channel (SCH) Acquisition Indicator Channel (AICH) Paging Indicator Channel (PICH)

MBMS Notification Indicator Channel (MICH)

High Speed Physical Downlink Shared Channel (HS-PDSCH) HS-DSCH-related Shared Control Channel (HS-SCCH)

Dedicated Physical Control Channel (uplink) for HS-DSCH (HS-DPCCH) Transport Channels

DCH

RACH

BCH FACH PCH

Physical Channels

HS-DSCH E-DCH

Fig. 2-6 Summarizing the mapping of TrCH s onto PhCHs

c

_long,1,n

c

long,2,n

MSB LSB

Fig. 2-7 Configuration of uplink scrambling sequence generator

0

7 4

+ mod n addition

1 d(i) 2 3 5

6

2

mod 2

0

7 4

1 b(i) 2 3 5

6

2

mod 2

+

mod 4 multiplication

z_n(i)

0

7 6 5 4 3 2 1

+ mod 4

Mapper

cshort,1,n(i)

a(i)

+ + +

+ + +

+ + +

3 3

3 2

c_short,2,n(i)

Fig. 2-8 Uplink short scrambling sequence generator for 255 chip sequence

I Q

1

1 0

0 2

2 3

3 4

4 5

5 6

6 7

7 8

8 9

9 17

17 16

16 15

15 14

14 13

13 12

12 11

11 10

10

Fig. 2-9 Configuration of downlink scrambling code generator

S F = 1 S F = 2 S F = 4 Cc h ,1 ,0 = (1 )

Cc h ,2 ,0 = (1 ,1 )

Cc h ,2 ,1 = (1 ,-1 )

Cc h ,4 ,0 = (1 ,1 ,1 ,1 )

Cc h ,4 ,1 = (1 ,1 ,-1 ,-1 )

Cc h ,4 ,2 = (1 ,-1 ,1 ,-1 )

Cc h ,4 ,3 = (1 ,-1 ,-1 ,1 )

Fig. 2-10 Code-tree for generation of OVSF codes

Pilot Npilot bits

TPC NTPC bits Data

Ndata bits

Slot #0 Slot #1 Slot #i Slot #14

Tslot = 2560 chips, 10 bits

1 radio frame: Tf = 10 ms DPDCH

DPCCH FBI

NFBI bits TFCI

NTFCI bits

Tslot = 2560 chips, Ndata = 10*2^k bits (k=0..6)

Fig. 2-11 Frame structure for uplink DPDCH/DPCCH

Σ

I

j

cd ,1 βd

I+ jQ D P D C H1

Q

cd ,3 βd

D P D C H3

cd ,5 βd

D P D C H5

cd ,2 βd

D P D C H2

cd ,4 βd

cc βp

D P C C H

Σ

Sd p c h

D P D C H4

cd ,6 βd

D P D C H6

Fig. 2-12 The uplink spreading of DPCCH and DPDCHs

Σ

Sd p ch ,n

I+ jQ Sd p ch

Shs -d p c c h

S

Se -d p ch

Spreading Spreading Spreading D P C C H

D P D C H s

H S -D P C C H

E -D P D C H s E -D P C C H

Fig. 2-13 Spreading for uplink dedicated channels

S

Im{S}

Re{S}

cos(ωt)

Complex-valued chip sequence from spreading operations

-sin(ωt) Split

real &

imag.

parts

Pulse-shaping

Pulse-shaping

Fig. 2-14 Uplink modulation

One radio frame, Tf = 10 ms TPC

NTPC bits

Slot #0 Slot #1 Slot #i Slot #14

Tslot = 2560 chips, 10*2^k bits (k=0..7) Data2 Ndata2 bits DPDCH TFCI

NTFCI bits

Pilot Npilot bits Data1

Ndata1 bits

DPDCH DPCCH DPCCH

Fig. 2-15 Frame structure for downlink DPCH

I

downlink physical channel

S

→ P

Cch,SF,m

j

S_dl,n

Q

I+jQ S

Modulation Mapper

Fig. 2-16 Spreading for all downlink PhCHs except SCH

Different downlink Physical channels (point S in Fig. 2-16)

Σ

G1

G2

GP

GS

S-SCH

P-SCH

Σ

(point T in

Fig. 2-18)

Fig. 2-17 Combining of downlink PhCHs

T

Im{T}

Re{T}

cos(ωt)

Complex-valued chip sequence from summing operations

-sin(ωt) Split

real &

imag.

parts

Pulse-shaping

Pulse-shaping

Fig. 2-18 Downlink modulation

Fig. 2-19 Transmitter model

j

Fig. 2-20 Spreader for the k-th user

,1( )

k t α

,2( )

k t

α

k( ) s t

k( ) r t )

,_P(t αk

Fig. 2-21 Multipath fading channel model

( )

r t 1

2β_{d b}T

1 2β_{d b}T

1 2β_{d b}T

* ) (

, )

(αˆ_Jⁿ_F )

( )

( _J,F ^*_{J J,F}

O t C t

C −τ −τ

) ( )

( _J,2 ^*_{J J,2}

O t C t

C −τ −τ

) ( )

( _J,1 ^*_{J J,1}

O t C t

C −τ −τ

* ) (

2

, )

(αˆ_Jⁿ

* ) (

1

, )

(αˆ_Jⁿ

)

ˆ(ⁿ

YJ

Fig. 2-22 Structure of the MRC RAKE with F finger combining

Chapter 3

Pilot-Channel Aided Successive

Interference Cancellation

在文檔中 應用引導通道協助於寬頻分碼多重進接系統 (頁 44-60)