Implementation of Pipelined Architecture for Physical Downlink Channels of 3GPPLTE

International Journal of Next-Generation Networks (IJNGN) Vol.5, No.1, March 2013
DOI : 10.5121/ijngn.2013.5101 1
IMPLEMENTATION OF PIPELINED ARCHITECTURE
FOR PHYSICAL DOWNLINK CHANNELS OF 3GPP-
LTE
S. Syed Ameer Abbas#1a
, M. Beril Sahaya Mary#1
, J. Rahumath Nisha#1
,
S. J. Thiruvengadam#2b
a
Assistant Professor, b
Professor
#
Department of Electronics and Communication Engineering
1
Mepco Schlenk Engineering College, Sivakasi- 626005, India
1
abbas_mepco@yahoo.com
2
Thiagarajar College of Engineering, Madurai- 625015, India
2
sjtece@tce.edu
ABSTRACT
LTE (Long Term Evolution) is a high data rate, low latency and packet optimized radio access technology
designed to support roaming Internet access via cell phones and handheld devices in 3G and 4G networks.
This paper mainly focuses on to improve the processing speed and decrease the maximum delay of the
downlink channels using the pipelined buffer controlled technique. This paper proposes Pipelined buffer
controlled Architecture for both transmitter and receiver for Physical Downlink channels of 3GPP-LTE.
The transmitter architecture comprises Bit Scrambling, Modulation mapping, Layer mapping, Precoding
and Resource element mapping modules. The receiver architecture comprises Demapping from resource
elements, Decoding, Comparing and Detection, Delayer mapping and Descrambling modules as described
in LTE specifications. In addition to these, buffers are included in both transmitter and receiver
architectures. Modelsim is used for simulation, synthesis and implementation are achieved using
PlanAhead13.2 tool on Virtex-5, xc5vlx50tff1136-1 device board is used. Implemented results are discussed
in terms of RTL design, FPGA editor, Power estimation and Resource estimation.
KEYWORDS
LTE, SISO, MISO, MIMO, PBCH, PDSCH, PCFICH, PHICH, PDCCH, PMCH
1. INTRODUCTION
The Long Term Evolution of UMTS (Universal Mobile Telecommunication Systems) is one of
the recent advancements in today’s mobile telecommunications systems. Though GSM
technology has improved in a certain manner by connecting communities and individuals in
remote regions where fixed-line connectivity was nonexistent, the successor of GSM, developed
by the 3GPP (Third Generation Partnership Projects) LTE is framed to increase the speed and
capacity of voice as well as data signals. There are many issues relating to the implementation of
LTE. The hardware architecture in the physical layer is one of the key research topics for the
VLSI Engineers. The main objective of this paper is to bring the pipelined buffer controlled
architecture of downlink data and control channels for transmitter and receiver.

2
There are two types of radio frame structures for LTE: Type‐1 frame structure is applicable to
Frequency Division Duplex and type‐2 frame structure is related to Time Division Duplex. This
paper is framed for type 1 Frame structure which is shown in Figure 1. Each radio frame is
ms
10
307200 s
f =
⋅
= T
T long and consists of 20 slots of length ms
5
.
0
T
15360 s
slot =
⋅
=
T , numbered
from 0 to 19. A subframe is defined as two consecutive slots where subframe i consists of slots
i
2 and 1
2 +
i .For FDD, 10 subframes are available for downlink transmission and 10 subframes
are available for uplink transmissions in each 10 ms interval. Uplink and downlink transmissions
are separated in the frequency domain. In half-duplex FDD operation, the UE (User Equipment)
cannot transmit and receive at the same time while there are no such restrictions in full-duplex
FDD.
Figure 1Frame structure type 1.
This paper is organized as follows: Section 2 discusses about the LTE Physical downlink
channels and their functions. Section 3 describes the system model of transmitter and receiver for
the Physical downlink channels based on 3GPP specifications. Section 4 discusses the proposed
architecture for the SISO, MISO and MIMO transmitters and receivers. Section 5 gives the
simulated and implemented results for the proposed system of transmitter and receiver. Finally
this paper is concluded with section 6.
2. LTE PHYSICAL DOWNLINK CHANNELS
The LTE downlink physical channels include three control channels and three data channels. The
control channels PDCCH (Physical Downlink Control Channel), PCFICH (Physical Control
Format Indicator Channel) and PHICH (Physical Hybrid Indicator Channel) are essential for the
successful reception, demodulation and decoding of the data channels PDSCH (Physical
Downlink Shared Channel), PBCH (Physical Broadcast Channel) and PMCH (Physical Multicast
Channel). The signals for the control channels are transmitted at the start of each subframe in a
LTE grid.
The PDSCH (Physical Downlink Shared Channel) is the main data-bearing downlink channel in
LTE. It is used for all user data, as well as for broadcast system information which is not carried
on the PBCH, and for paging messages – there is no speciﬁc physical layer paging channel in the
LTE system. The PDSCH is utilized basically for data and multimedia transport. [1] The PBCH
(Physical Broadcast Channel) carries the basic system information which allows the other
channels to be conﬁgured and operated in the LTE grid. PBCH information is divided into two
categories. They are Master Information Block (MIB) and System Information Block (SIB).
QPSK is the only modulation used [1]. The PMCH (Physical Multicast channel) physical channel
structure is defined for future use. MBMS enables a set of eNBs to transfer information
simultaneously for a given duration. MBFSN appears to UE (User Equipment) as a transmission
from a single large cell and improves the signal-to-interference noise ratio. Since the operation of
data transmission is different from that of MBFSN, UE should have a separate channel estimate
for MBSFN reception; therefore normal reference signal is not mixed with the MBSFN reference

3
signal in the same subframe and transmission of PDSCH and PMCH in the same subframe is not
possible [1].
The PDCCH (Physical Downlink Control Channel) is to carry mainly scheduling information of
different types such as Downlink resource scheduling, Uplink power control instructions. The
control information carried by PDCCH is DCI (Downlink Control Information) which is
transmitted as an aggregation of Control Channel Elements (CCEs). Each CCE consists of 9
Resource Element Groups (REGs) and each REG has 4 resource elements (Res). Each RE carries
2 bits. A PDCCH can transmit 1, 2, 4 or 8 CCEs. QPSK is the only available modulation format.
[2]The PCFICH is transmitted on the first symbol of every sub-frame and carries a Control
Format Indicator (CFI) field. The CFI contains a 32 bit code word that represents 1, 2, 3 or 4.
CFI-4 is meant for future use [2]. The value of the PCFICH informs the UE (User Equipment)
about the number of OFDM symbols used for the transmission of control channel (PDCCH)
information in a subframe. The PCFICH uses QPSK modulation. The PHICH channel is used to
report the Hybrid ARQ (Hybrid Automatic Repeat Request) status which indicates to the UE
whether the uplink user data is correctly received or not. The HARQ indicator is 1 bit long - "0"
indicates ACK, and "1" indicates NACK. BPSK modulation is used in PHICH channel. [3]
3. SYSTEM MODEL
In LTE, the data which is given to the transmitter should experience the following channel
processing steps. Figure 2 shows the channel processing steps of transmitter. Figure 3 shows the
channel processing steps of receiver.
Figure 2 Modules of Transmitter
Figure 3 Modules of Receiver
3.1 CHANNEL PROCESSING STEPS OF TRANSMITTER
3.1.1 Scrambling
For each codeword q, the block of bits 1)
(q)
bit
(M
(q)
b
(0),...,
(q)
b − , where
)
(
bit
q
M is the number of
bits in codeword q transmitted on the physical channel in one subframe, shall be scrambled
according to (1) for all the five channels except PHICH, where c(q)
(i) is a Gold sequence of length
“i”. The modulation and scrambling of PHICH is done according to (2) [4].
( ) 2
mod
)
(
)
(
)
(
~ )
(
)
(
)
(
i
c
i
b
i
b q
q
q
+
= ................ (1)
3
possible [1].
3. SYSTEM MODEL
3.1.1 Scrambling
(q)
bit
(M
(q)
b
(0),...,
(q)
b − , where
)
(
bit
q
M is the number of
( ) 2
mod
)
(
)
(
)
(
~ )
(
)
(
)
(
i
c
i
b
i
b q
q
q
+
= ................ (1)
3
possible [1].
3. SYSTEM MODEL
3.1.1 Scrambling
(q)
bit
(M
(q)
b
(0),...,
(q)
b − , where
)
(
bit
q
M is the number of
( ) 2
mod
)
(
)
(
)
(
~ )
(
)
(
)
(
i
c
i
b
i
b q
q
q
+
= ................ (1)

4
( ) ( )  
( )
1
,......
0
)
(
2
1
mod
)
( PHICH
SF
PHICH
SF
−
=
⋅
−
⋅
=
symb
M
i
N
i
z
i
c
N
i
w
i
d
............. (2)
Where s
PHICH
SF
symb M
N
M ⋅
= and



=
prefix
cyclic
extended
2
prefix
cyclic
normal
4
PHICH
SF
N
3.1.2 Modulation
In general LTE follows four different types of modulation techniques such as BPSK, QPSK, 16
QAM and 64 QAM. The physical downlink channels and their corresponding modulation
techniques are shown in Table 1.
Table 1 Channel and Modulations
The scrambled sequence is then modulated to create a block of complex valued modulated
symbols as )
1
(q)
symb
(M
(q)
d
(0),...,
(q)
d − . In QPSK modulation a pairs of bits are mapped to
complex valued modulation symbols I+ jQ, as shown in Table 2. Similarly the distributed
arithmetic processing is applied for 16QAM, 64QAM [4].
Table 2 QPSK Modulation
3.1.3 Layer Mapping
The modulated symbols are then layer mapped to one or more layers depending upon the number
of antenna ports selected .The complex modulated input symbols are layer mapped as
[ ]T
i
x
i
x
i
x )
(
...
)
(
)
( )
1
(
)
0
( −
= 
, 1
,...,
1
,
0 layer
symb −
= M
i where  is the number of layers and layer
symb
M is the
number of modulation symbols per layer in the layer mapping module.In this paper, transmitter
diversity is adopted; the input symbols are mapped to layers according to Table 3 [4].
Channels Type of Modulation
PBCH,PCFICH,PDCCH QPSK
PDSCH QPSK, 16 QAM, 64 QAM
PMCH QPSK, 16 QAM, 64 QAM
PHICH BPSK
b(i),b(i+1) I Q
00
2
1 2
1
01 2
1 2
1
−
10 2
1
− 2
1
11 2
1
− 2
1
−

5
Table 3 Mapping to layers
3.1.4 Precoding
The precoder takes a block from the layer mapper x(0)
(i), x(1)
(i),... x(v-1)
(i), and generates a
sequence for each antenna port y(p)
(i), p is the transmit antenna port number and is {0},{0,1} or
{0,1,2,3}[4]. For transmission over a single antenna, port processing is carried out by (3).
( )
( ) = ( )
( ) … . (3)
Precoding for transmit diversity is available on two or four antenna ports. In two antenna port
precoding, an Alamouti scheme is used for precoding. This precoding procedure for two antenna
case is defined by (4)
( )
( )
( )
( )

























−
−
=














+
+
)
(
Im
)
(
Im
)
(
Re
)
(
Re
0
0
1
0
1
0
0
1
0
0
0
1
2
1
)
1
2
(
)
1
2
(
)
2
(
)
2
(
)
1
(
)
0
(
)
1
(
)
0
(
)
1
(
)
0
(
)
1
(
)
0
(
i
x
i
x
i
x
i
x
j
j
j
j
i
y
i
y
i
y
i
y
..... (4)
with 1
,...,
1
,
0 layer
symb −
= M
i with layer
symb
ap
symb 2M
M = . For transmission on four antenna ports, { }
3
,
2
,
1
,
0
∈
p ,
the output [ ]T
i
y
i
y
i
y
i
y
i
y )
(
)
(
)
(
)
(
)
( )
3
(
)
2
(
)
1
(
)
0
(
= , 1
,...,
1
,
0 ap
symb −
= M
i of the precoding operation
is defined by (5)
( )
( )
( )
( )
( )
( )
( )
( )











































































−
−
−
−
=




















































+
+
+
+
+
+
+
+
+
+
+
+
)
(
Im
)
(
Im
)
(
Im
)
(
Im
)
(
Re
)
(
Re
)
(
Re
)
(
Re
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
1
)
3
4
(
)
3
4
(
)
3
4
(
)
3
4
(
)
2
4
(
)
2
4
(
)
2
4
(
)
2
4
(
)
1
4
(
)
1
4
(
)
1
4
(
)
1
4
(
)
4
(
)
4
(
)
4
(
)
4
(
)
3
(
)
2
(
)
1
(
)
0
(
)
3
(
)
2
(
)
1
(
)
0
(
)
3
(
)
2
(
)
1
(
)
0
(
)
3
(
)
2
(
)
1
(
)
0
(
)
3
(
)
2
(
)
1
(
)
0
(
)
3
(
)
2
(
)
1
(
)
0
(
i
x
i
x
i
x
i
x
i
x
i
x
i
x
i
x
j
j
j
j
j
j
j
j
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
i
y
..... (5)
Number of
layers
Layer mapping
i=0,1,..., Mlayer
symb-1
1 X(0)
(i)=d(0)
(i)
2
X(0)
(i)=d(0)
(2i) Mlayer
symb=M(0)
symb/2
X(1)
(i)=d(0)
(2i+1)
4
X(0)
(i)=d(0)
(4i)
X(1)
(i)=d(0)
(4i+1)
X(2)
(i)=d(0)
(4i+2) Mlayer
symb=M(0)
symb/4
X(3)
(i)=d(0)
(4i+3)

6
with
( )





≠
−
=
=
0
4
mod
if
2
4
0
4
mod
if
4
)
0
(
symb
layer
symb
)
0
(
symb
layer
symb
ap
symb
M
M
M
M
M .
3.1.5 Mapping to Resource Elements
For each of the antenna ports used for transmission of the physical channel, the block of complex-
valued symbols )
1
(
),...,
0
( ap
symb
)
(
)
(
−
M
y
y p
p
shall be mapped in sequence starting with )
0
(
)
( p
y to
resource elements ( )
l
k, , where ‘k’ represents the rows of the LTE grid and ‘l’ represents the
columns of the LTE grid. The downlink channels’ precoded symbols are mapped to the LTE grid
in their respective resource element groups (REG), and control channels are mapped only in the
first OFDM symbol of each subframe and are transmitted.
3.2 CHANNEL PROCESSING STEPS OF RECEIVER
3.2.1 Demapping From Resource Elements
While data is received on the antenna ports, the block of complex-valued symbols
)
1
(
),...,
0
( ap
symb
)
(
)
(
−
M
y
y p
p
shall be demapped in sequence starting with )
0
(
)
( p
y from resource
elements ( )
l
k, [4].
3.2.2 Decoding
The decoder takes as input a block of vectors [ ]T
p
i
y
i
y ...
)
(
...
)
( )
(
= , 1
,...,
1
,
0 ap
symb −
= M
i demapped
from resources on each of the antenna ports, where )
(
)
(
i
y p
represents the signal from antenna
port p and generates a block of vectors [ ]T
i
x
i
x
i
x )
(
...
)
(
)
( )
1
(
)
0
( −
= 
, 1
,...,
1
,
0 layer
symb −
= M
i for the
delayer mapping. For reception on a single antenna port, decoding is defined by (6) [4].
( )
( ) = ( )
( ) ... (6)
Similarly for reception on two antenna ports the reverse operation of (4) is performed.
3.2.3 Delayer Mapping
The complex-valued symbols for each of the code words to be received are delayer mapped from
one or several layers. Complex-valued symbols )
1
(
),...,
0
( (q)
symb
)
(
)
(
−
M
d
d q
q
for codeword q shall
be demapped from the layers [ ]T
i
x
i
x
i
x )
(
...
)
(
)
( )
1
(
)
0
( −
= 
, 1
,...,
1
,
0 layer
symb −
= M
i where  is the
number of layers and layer
symb
M is the number of symbols per layer. For a single antenna port, a
single layer is used, and the delayer mapping is defined by (7) [4].
( )
( ) = ( )
( ) … (7)
Similarly for two antenna ports the operations performed in Table 3 are reversed, and the two
layers are delayer mapped into a single layer.

7
3.2.4 Demodulation
The complex-valued symbols )
1
(
),...,
0
( (q)
symb
)
(
)
(
−
M
d
d q
q
of code word q
, shall be demodulated
using a demodulation scheme which is reverse of transmitter, resulting in a block of bits
)
1
(
~
),...,
0
(
~ (q)
bit
)
(
)
(
−
M
b
b q
q
3.2.5 Descrambling
The demodulated symbols in a code word q , after descrambling results in the block of bits
)
1
(
),...,
0
( )
(
bit
)
(
)
(
−
q
q
q
M
b
b , where )
(
bit
q
M is the number of bits in code word q received on each
control channels in one sub frame. The descrambling is defined according to (8).
( )
( ) = ( )
( ) + ( )
( ) 2 ........ (8)
where ( )
( ) is the generated pseudo random Gold sequence. The descrambling
sequence is initialized with same values as that of the transmitter at start of each subframe.
4. PROPOSED ARCHITECTURE OF PHYSICAL DOWNLINK
CHANNELS FOR LTE
4.1 TRANSMITTER ARCHITECTURE
Figure 4 Pipelined Buffer Controlled Architecture of Transmitter for downlink
Channels in LTE
The reason for going to pipelined architecture is to reduce the critical path delay which can be
exploited to increase the clock speed or to reduce the power consumption for the same speed. In
this pipelined buffer control architecture, the different processing modules are segregated to
segments and they have buffers in between them to work independently .This reduces the critical
path as a whole when compared with buffer less segmented approach. The buffer control is used
7
3.2.4 Demodulation
1
(
),...,
0
( (q)
symb
)
(
)
(
−
M
d
d q
q
of code word q
)
1
(
~
),...,
0
(
~ (q)
bit
)
(
)
(
−
M
b
b q
q
3.2.5 Descrambling
)
1
(
),...,
0
( )
(
bit
)
(
)
(
−
q
q
q
M
b
b , where )
(
bit
q
( )
( ) = ( )
( ) + ( )
( ) 2 ........ (8)
where ( )
CHANNELS FOR LTE
Channels in LTE
7
3.2.4 Demodulation
1
(
),...,
0
( (q)
symb
)
(
)
(
−
M
d
d q
q
of code word q
)
1
(
~
),...,
0
(
~ (q)
bit
)
(
)
(
−
M
b
b q
q
3.2.5 Descrambling
)
1
(
),...,
0
( )
(
bit
)
(
)
(
−
q
q
q
M
b
b , where )
(
bit
q
( )
( ) = ( )
( ) + ( )
( ) 2 ........ (8)
where ( )
CHANNELS FOR LTE
Channels in LTE

8
to synchronize all the segments and send the information out to the grid.Figure 4 explains the
pipelined buffer controlled architecture for the transmitter. The transmitter side of the architecture
consists of Scrambling, Modulation, Layer Mapping, Precoding and Mapping to the Resource
Elements .In addition to these buffers are included in this architecture.
The scrambler generates scrambling bits at the rate of one bit for every clock cycle. For the
PHICH channel the one bit input either 0 (indicating ACK) or 1 (indicating NACK) is first
modulated using BPSK modulation and then scrambled. The modulation mapper takes 2bits /4bits/
6bits for QPSK/16QAM/64QAM modulations respectively as inputs. Hence a buffer is included
before the modulation block. The task of the buffer is to store either 2bits/4bits/6bits in two clock
cycles or four clock cycles or six clock cycles based on the type of modulation given to it by the
variable “Modcon”. After the corresponding bits have been stored, the buffer will send a buffer
control signal indicated by the variable “BC1” to the modulation block. For every possible pair of
2 bits/4 bits/6 bits i.e., “00” to “11” /”0000” to “1111”/ “000000” to “111111”, there resides a
corresponding modulated value (16 bits) stored in the modulation mapper. All the modulation
mapper has to do is, check the variable “Modcon” for the type of modulation given and simply
maps the incoming bits to the correct segment of modulated data. The output of modulation
mapper is 16 bits for every clock cycle which will be represented as a segment in the following
sections. The modulation block after performing the modulation sends a clear signal denoted as
“clr” to the buffer indicating the buffer to send the next stream of bits based on the “Modcon”. The
layer mapping module simply maps the modulated bits (i.e. segment) into 1/2/4 layers depending
on the transmitter diversity (SISO, MISO and MIMO) indicated by “Transdiv” variable. For the
single antenna (1x1) the layer mapper maps every single segment of the modulator to the first
layer/data path of all the physical downlink channels. In case of MISO (2x1 and 4x1) every single
segment is mapped into 2 layers and four layers respectively. In case of MIMO (2x2 and 4x2) two
different segments are mapped to two layers (one segment in each layer) and 4 different segments
are mapped to 4 layers respectively.
To achieve this kind of mapping in the respective layers, a buffer is included in front of the layer
mapper. For the SISO and MISO concepts, the buffer will wait for one clock cycle to store a
single segment. For the 2x2 MIMO the buffer will wait for two clock cycles to store the two
different segments and for the 4x2 MIMO the buffer will wait for four clock cycles to store four
different segments. After storing the respective segments, the buffer will send a buffer control
signal “BC2” to the layer mapper. The layer mapper after performing the mapping operations
sends a “clr” signal to the buffer signalling it to send the next set of segments. The precoding
block adds parity bits to the segments by appending zeros and conjugates of original data. The
transmitter diversity “Transdiv” configures the precoding block also. When the transmitter
diversity is SISO or MISO the 1/2/4 data paths consists of the same single segment. They are
precoded as such. When the transmitter diversity is MIMO two different cases exists, 2x2 and
4x2.The single segment in every layer is precoded into two segments for 2 layers and precoded
into 4 segments for 4 layers. This all happens in a single clock cycle. The Resource element
mapper only maps a single segment at a clock cycle to the resource elements. The buffer in front
of the RE mapper stores the (1/2/4) segments in 1/2/4 data paths and delivers a segment for every
clock cycle to the LTE grid.

9
Table 4 Total Delays experienced
4.2 RECEIVER ARCHITECTURE
The receiver side of the architecture consists of Demapping from the Resource Elements,
Decoding and Detection, Delayer Mapping and Descrambling as shown by the Figure 5.The
receiver architecture is designed with two receiving antennas. When the transmitter diversity is
SISO (1x1) or MISO (2x1 and 4x1) antenna 1 is enabled to receive. Similarly when MIMO (2x2
and 4x2) case occurs both the antennas are enabled to receive. The data demapper module demaps
the segments from the resource elements.
Figure 5 Pipelined Buffer Controlled Architecture of Receiver for downlink
Channels in LTE
For the SISO and MISO cases there is only one segment. The buffer stores the single segment in a
clock cycle. When it is MIMO the buffer waits for two clock cycles and stores the two segments
and sends the buffer control signal (BC1) , signalling the decoding module to accept the segments
in the next clock cycle. The decoding module removes the parity bits i.e. the extra padded zeros
and conjugate bits in the next clock cycle and sends the clear signal represented as “clr” to the
buffer and the detector modules.
Transmitter
Diversity
Representation Modulation Delays
(in CLK)
SISO (1X1)/
MISO 2X1)/
MISO(4X1)
001/011 / 101 BPSK
QPSK
16QAM
64QAM
14/14/14
6/7/9
8/9/11
10/11/13
MIMO2X2 100 BPSK
QPSK
16QAM
64QAM
14 CLK
10 CLK
12 CLK
14 CLK
MIMO 4X2
110
BPSK
QPSK
16QAM
64QAM
14 CLK
16 CLK
27 CLK
37 CLK
9
Channels in LTE
Transmitter
Diversity
(in CLK)
SISO (1X1)/
MISO 2X1)/
MISO(4X1)
001/011 / 101 BPSK
QPSK
16QAM
64QAM
14/14/14
6/7/9
8/9/11
10/11/13
MIMO2X2 100 BPSK
QPSK
16QAM
64QAM
14 CLK
10 CLK
12 CLK
14 CLK
MIMO 4X2
110
BPSK
QPSK
16QAM
64QAM
14 CLK
16 CLK
27 CLK
37 CLK
9
Channels in LTE
Transmitter
Diversity
(in CLK)
SISO (1X1)/
MISO 2X1)/
MISO(4X1)
001/011 / 101 BPSK
QPSK
16QAM
64QAM
14/14/14
6/7/9
8/9/11
10/11/13
MIMO2X2 100 BPSK
QPSK
16QAM
64QAM
14 CLK
10 CLK
12 CLK
14 CLK
MIMO 4X2
110
BPSK
QPSK
16QAM
64QAM
14 CLK
16 CLK
27 CLK
37 CLK

10
The detector consists of precomputed modulated data stored in it and a comparator. Initially after
the reception of “clr” signal from the decoder, the received data and the precomputed data are
subtracted and the results are compared with each other to find the minimum distance value using
the Alamouti scheme. Once the comparisons are done and the output is generated the detector
sends its clear signal to the delayer mapping module. Thus the detected or demodulated output is 2
bits/4 bits/6 bits for QPSK/16QAM/64QAM modulations respectively for each layer. This process
goes for every clock cycle. The delayer mapping module delayers the two layers of MIMO into a
single layer thus comprising 4 bits/8 bits/12 bits depending on the Modcon (modulation) input
variable given and sends the clear signal. This all occurs in a single clock cycle. Since SISO comes
in a single layer, there is no need of delayer mapping. The descrambling is performed by XORing
the input bits with the Gold sequence used in the transmitter and produces 1 bit (ACK/NACK) for
PHICH and 2 bits/4 bits/6 bits depending on the “Modcon” for the remaining channels in a single
clock cycle. Thus the data and control messages are retrieved in the receiver.
5. RESULTS AND DISCUSSIONS
5.1 Simulation output for SISO transmitter
The SISO transmitter for PBCH channel is shown in Figure 6. The output of bit scrambler, the first
module of transmitter “scr_pbch” is a single bit for every clock cycle and it’s own clock
canc_pbch is shown below. The buffer indicated by “buffscr2pbch” in the figure 4 wait for two
clock cycles and got the pair of bits from scrambler. The modulation will be performed in the next
clock cycle. The modulated segment “mod_pbch” is layer mapped into a single layer, precoded
and data mapped in the forth coming clock cycles with a delay of one clock cycle between each
process. The precoded data shown as “prelayer1pbch” for transmit antenna 1 is mapped into the
corresponding resource element for the PBCH channel in the LTE grid in the next clock cycle.
Therefore the SISO transmitter for PBCH takes a total delay of 6 clock cycles. This explanation
suits well for the PDSCH, PMCH, PDCCH and PCFICH channels.
Figure 7 and 8 describe the SISO transmitter for PDSCH [16QAM] and PDSCH [64QAM]
respectively. The buffer represented as “buffscr246pdsch” will wait for 4 clock cycles and 6 clock
cycles before sending 4 bits and 6 bits to the modulation mapper respectively. The delays required
for remaining modules are similar to PBCH (SISO). Therefore the PDSCH channel takes a total
delay of 8 clock cycles for 16QAM and 10 clock cycles for 64QAM modulations.
Figure 6 Simulation result for SISO transmitter (1X1) PBCH

11
Figure 7 Simulation result for SISO transmitter (1X1) PDSCH [16QAM]
Figure 8 Simulation output for SISO transmitter-PDSCH [64QAM]
The SISO transmitter for PHICH is shown in Figure 9. The first step is a modulation instead of
scrambling, in which a single bit input whether 0 (ACK) or 1(NACK) is BPSK modulated. Then
the modulated bits are scrambled to produce a output of 12 segments (192 bits) represented as
“out” in the figure 9.The layer mapping and precoding are done in the next occurring clock
cycles. The data mapping module takes 12 clock cycles to map the 12 PHICH segments to the
transmitter antenna 1.Therefore total delay is 14 clock cycles.
Figure 9 Simulation output for SISO transmitter-PHICH
5.2 Simulation output for MISO (2x1) transmitter
The MISO (2x1) transmitter for PDCCH channel is shown in Figure 10. It is similar to SISO
except that the modulation mapper’s output “modpdcch” is layer mapped into two layers
represented as “layer1pbch” and “layer2pdcch”. The single segment in each of the two layers is
precoded into two segments represented as “prelayer1pdcch” and “prelayer2pdcch”, The data
mapper module for the transmit antenna 1 and 2 maps the two segments into their LTE grids in
two clock cycles. Therefore the total delay is 7 clock cycles. This explanation suits well for the
PDSCH [QPSK], PMCH, PBCH and PCFICH channel. PDSCH [16QAM] experiences a delay of
9 clock cycles the extra two clock cycles was being spent in waiting for 4 bits input to the
modulation mapper. Also PDSCH [64QAM] experiences a delay of 11 clock cycles.

12
The MISO (2x1) transmitter for PHICH is shown in Figure 11. The output of the modulation
mapper and scrambler module “out” consists of 12 segments which is layer mapped into 2 layers
“layer1phich” and “layer2phich” as 6 segments each in the next clock cycle. The precoder module
precodes these 6 segments with parity bits, giving an output of 12 segments in each data path in
the next clock cycle. The data mapping module for the first two transmit antennas map the 12
segments into the resource block allocated for PHICH channel in 12 clock cycles. Therefore the
total delay is 14 clock cycles. The first two antennas represented as transmit_0 and transmit_1 with
the same data being transmitted for every clock cycle is shown in the Figure 11.
5.3 Simulation output for MISO (4x1) transmitter
The MISO (4x1) transmitter for PCFICH is shown in Figure 12. The processing steps and the
delays experienced by the MISO (4x1) transmitter is similar to MISO (2x1) except that the layer
mapping block maps the modulation output into 4 layers/data paths and the precoder precodes
each segment in every layer into 4 segments. The data mapping module maps the 4 segments in 4
clock cycles. Therefore a total delay of 9 clock cycles is experienced. This explanation suits well
for the PDSCH [QPPSK], PBCH and PDCCH channels. PDSCH [16QAM] experiences a total
delay of 11 clock cycles. PDSCH [64QAM] experiences a total delay of 13 clock cycles.
The MISO (4x1) transmitter for PMCH [QPSK] is shown in Figure 13. The PMCH channel is only
transmitted in the fourth antenna i.e.transmit_3. As it is transmitted in a single antenna, the
channel processing steps of PMCH channel follows SISO strategy. Therefore the delays
experienced by each module are similar to SISO transmitter. The total delay is thus 6 clock cycles.
PMCH [16QAM] and PMCH [64QAM] are very much similar to the SISO transmitter PDSCH
channel described earlier, therefore the total delay is 8 clock cycles and 10 clock cycles
respectively. The MISO (4x1) transmitter for PHICH is shown in Figure 14. After the modulation
and scrambling the 12 segments generated are layer mapped into 4 layers as such: 3 segments for
each layer in the next clock cycle. The precoding module adds extra zeros and conjugates thus
generating 12 segments in each of the 4 layers in the next clock cycle. The data mapping modules
takes 12 clock cycles to map the precoded data into the resource blocks of 4 transmit antennas.
Therefore a total delay of 14 clock cycles exists in PHICH channel.
.
Figure 10 Simulation output for MISO (2x1) Figure transmitter-PDCCH

13
Figure 11 Simulation output for MISO (2x1) transmitter-PHICH
Figure 12 Simulation output for MISO (4x1) transmitter-PCFICH
Figure 13 Simulation output for MISO (4x1) transmitter-PMCH [QPSK]
Figure 14 Simulation output for MISO (4x1) transmitter-PHICH

14
5.4 Simulation output for MIMO (2x2) transmitter
The MIMO (2x2) transmitter for PDSCH [QPSK] is shown in Figure 15. In MIMO (2x2), the
channel processing steps and their corresponding delays are similar to MISO (2x1) upto the
modulation module. The buffer represented as “pdschbuffer246” before the layer mapping module
get two different segments from the modulation mapper in the sixth clock cycle. The layer
mapping is done in the next clock cycle. The precoding is thus done in the 8th
clock cycle
generating two segments for each data path. The data mapping module takes another two clock
cycles. Thus a total delay of 10 clock cycles exists. This explanation suits well for the PBCH,
PDCCH and PCFICH channels. PDSCH [16QAM] experiences 12 clock cycles delay and for
PDSCH [64QAM] experiences 14 clock cycles delay. The MIMO (2X2) transmitter for PHICH is
similar to MISO (2x1). The transmitter outputs transmit_0 and transmit_1 carry different segments
in every clock cycle.
5.5 Simulation output for MIMO (4x2) transmitter
The MIMO (4X2) transmitter for PBCH is shown in Figure 16. In MIMO (4x2), the channel
processing steps and their corresponding delays are similar to MISO (4x1) up to the modulation
module. In the case of MIMO (4x2), four different segments (modulation output) have to be layer
mapped into the four layers/datapaths corresponding to the four transmit antennas. Hence the
buffer represented as “pbchbuffer” before the layer mapping module get four different segments
from the modulation mapper in the tenth clock cycle. The layer mapping is done in the next clock
cycle. The precoding is thus done in the 8th
clock cycle generating four segments for each data
path. The data mapping module takes another four clock cycles. Thus a total delay of 16 clock
cycles exists. This explanation suits well for the PDSCH, PMCH, PDCCH and PDCCH channels.
In PDSCH [16QAM] the modulation is performed in the fifth clock cycle as usual. The buffer
before the layer mapping block will wait for 16 clock cycles to receive 4 different segments from
the modulator and the layer mapping is performed in the 22nd
clock cycle. The precoding is done
in the next clock cycle. The data mapping module needs another 4 clock cycles. Thus a total delay
of 27 clock cycles is experienced. In PDSCH [64QAM] the modulation is performed in the
seventh clock cycle. The buffer before the layer mapper waits for 24 clock cycles to receive 4
different segments. The remaining delays in the successive modules are as such. Hence a total
delay of 37 clock cycles exists. The MIMO (4X2) transmitter for PMCH is similar to MISO (4x1).
The total delay is 6 clock cycles for QPSK, 7 clock cycles for 16QAM and 9 clock cycles for 64
QAM modulations.
Figure 15 Simulation output for MIMO (2x2) transmitter-PDSCH [QPSK]

15
Figure 16 Simulation output for MIMO (4x2) transmitter-PBCH
5.6 Simulation output for SISO (1x1), MISO (2x1) and MISO (4x1) Receiver
The simulation results of SISO and MISO receiver for all the physical downlink channels are the
same. The simulation output for SISO and MISO receiver for PBCH channel is shown in Figure
17. The output of the data demapping module of antenna 1 is “prelayer1pbch”. It is given to the
decoding module in the next clock cycle. After decoding, the detector output which is 2 bits for
QPSK modulation is generated in the next clock cycle. The delayer mapping and descrambling are
performed in the next occurring clock cycles. The output of the descrambling module is 2 bits at a
clock cycle. Therefore the total delay is 5 clock cycles. This is similar to PDCCH, PCFICH,
PDSCH [QPSK] channels.
The simulation output for SISO and MISO receiver for PDSCH [16QAM] shown in Figure 18. In
this case also the total delay is 5 clock cycles and the descrambler output is 4 bits at a single clock
cycle. In the SISO and MISO receiver for PDSCH [64QAM] the total delay is 5 clock cycles and
the descrambler output is 6 bits at a single clock cycle. The simulation output for MISO (4X1
only) receiver for PMCH [QPSK] is also similar to PDSCH. Here also the delay is 5 clock cycles
with descrambler output 2 bits at a clock cycle. PMCH [16QAM] and [64QAM] also experiences
5 clock cycles delay with descrambler outputs 4bits and 6 bits respectively. The simulation output
for SISO and MISO receiver for PHICH shown in Figure 19. The 12 segments are received by the
data demapping module in 12 clock cycles. After that the decoding, delayer mapping,
descrambling, detection and decision modules are performed in the next forthcoming clock cycles.
The output of the decision module is either 0 (ACK) or 1 (NACK).Thus a total delay of 17 clock
cycles occur.
5.7 Simulation output for MIMO (2x2) and MIMO (4x2)
The simulation output for MIMO (2x2 and 4x2) receiver for PDSCH [QPSK] channel shown in
Figure 20. The data demapping modules of antenna 1 and 2 receives the two segments in clock
cycles from the 2 antennas. Therefore the buffer has to wait for 2 clock cycles and in third clock
cycle the decoder output is generated. The detector output is two bits in two data paths in the
fourth clock cycle. In the fifth clock cycle the delayer mapper delayers the two layers into a single
layer thus comprising 4 bits. The descrambler produces 2 bits in the sixth clock cycle. Therefore a
total delay of 6 clock cycles occurs. This is similar to PDCCH, PCFICH, PBCH channels. PDSCH
[16QAM] and PDSCH [64QAM] have a total delay of 6 clock cycles with 4 bits at a clock cycle in
the descrambler output and 6 bits at a clock cycle in the descrambler output respectively. PMCH
channel for MIMO (4x2) is similar to MISO (4x1). PHICH channel for MIMO is also similar to
SISO and MISO cases.

16
Figure 17 Simulation output for SISO and MISO receiver-PBCH
Figure 18 Simulation output for SISO and MISO receiver-PDSCH [16QAM]
Figure 19 Simulation output for SISO and MISO receiver-PHICH
Figure 20 Simulation output for MIMO (2x2 and 4x2) receiver-PDSCH [QPSK]

17
5.8 Implementation Results
The RTL schematic view of the transmitter shows the input variables to each of the scrambling
module, buffers, modulation, layer mapping module, precoding and data mapping modules. This is
shown in Figure 21 for designer’s reference. The final outputs i.e. transmit_0, transmit_1,
transmit_2 and transmit_3 are also shown in the Figure 22. RTL schematic view of the receiver is
shown in Figure 22. The final outputs i.e. the data and control bits are also shown for designer’s
reference.
Figure 21 RTL design of Transmitter
Figure 22 RTL design of Receiver
Power estimation of LTE downlink channels for transmitter and receiver is shown in Figure 23a
and b. The total on chip power is 951 mW for buffer controlled architecture of transmitter. This
transmitter consumes 265mW for IO, 235mW for core dynamic (clock, logic) and 452mW for
device static. In core dynamic, clock signal consumes 107mW, logic signal takes 127mW. The
total on chip power is 1036 mW for this receiver. This receiver consumes 20mW for IO 563mW
for core dynamic (clock, logic) and 452mW for device static. In core dynamic, clock signal
consumes 124mW, logic signal takes 438mW. Resource estimation of transmitter and receiver
are shown in Figure 24a and b. From the graphical representation it is clear that out of the total
resources about 1% is used for registers, 4% for Look up tables, 7% for the slices, 14% for the
IO, 18% for BUFG. Similarly out of the total resources used for receiver, about 1% is used for
registers, 23% for Look up tables, 36% for the slices, 5% for the IO, 5% for BUFG. The FPGA

18
framed by plan Ahead tool for transmitter and receiver are shown in Figure 25a and b. The
routing between the inputs and output with available components in the chip is shown
Figure 23a and 23b Power estimation of Transmitter and Receiver
Figure 24a and 24b Resource utilization of the Transmitter and Receiver
Figure 25a and 25b FPGA editor of the Transmitter and Receiver

19
6. CONCLUSIONS
The synthesis and implementation of pipelined buffer controlled architecture of the Single Input
Single Output (SISO), Multiple Input Single Output (MISO-2x1, 4x1), and Multiple Input
Multiple Output (MIMO-2x2, 4x2) transmitter and receiver for Physical Downlink Channels of
LTE are thoroughly discussed in this paper. The buffers and the delays experienced by each of the
modules have been explained in this architecture. The transmitter and receiver architecture for all
the channels are simulated and the results are discussed. The programming was done in Verilog
HDL. The simulation was done in Modelsim, synthesized using Plan Ahead 13.4 and
implemented in Virtex 5 kit. Simulation results and implementation results (RTL design, power
estimation, resource estimation and FPGA editor) for transmitter and receiver of LTE downlink
channels are discussed. This architecture shows improvement in terms of reduced power
consumption, reduced maximum delay with high speed. The pipelined buffer controlled
architecture shows a minimum period of 59.334ns (nano seconds) and a maximum frequency of
16.854 MHz It maintains continuity between the modules. Also the formation of junk data
between the modules is avoided by the use of “BC” and “clr” signals.
ACKNOWLEDGEMENT
The authors wish to express their sincere thanks to All India Council for Technical Education,
New Delhi for the grant to do the project titled Design of Testbed for the Development of
Optimized Architectures of MIMO Signal Processing (No: 8023/RID/RPS/039/11/12) .They are
also thankful to the Management and Principal of Mepco Schlenk Engineering College, Sivakasi
for their constant support and encouragement to carry out this part of the project work
successfully.
REFERENCES
[1] S. Syed Ameer Abbas, A. Kanimozhi , S. J. Thiruvengadam ,“Realization of the SISO Architecture for
Downlink Data Channels of 3GPP-LTE using PlanAhead Tool and Virtex-5 Device”, IFRSA
International Journal Of Electronics Circuits And Systems |Vol 1|issue 2|July 2012.
[2] S. J. Thiruvengadam, Louay M. A. Jalloul, “Performance Analyis of the 3GPP-LTE Physical Control
Channels,” EURASIP Journal on Wireless Communications and Networking, vol. 2010, Article ID
914934, 10 pages, Nov. 2010.
[3] S .Syed Ameer Abbas, Geethu K .S, S.J Thiruvengadam, “Implementation of SISO Architecture for
LTE Downlink Control Channels in Virtex 5”, International Journal of Engineering and Innovative
Technology (IJEIT) Volume 1, Issue 5, May 2012.
[4] 3GPP TS 36.211,” Evolved Universal Terrestrial Radio Access (E- UTRA); Physical Channels and
Modulation (Release 8)”.
[5] Abbas, S. S. A, Praba. R. L, Thiruvengadam. S. J, “PCFICH Channel Design for LTE using FPGA” in
Proceedings of Recent Trends in Information Technology(ICRTIT) 2011 International Conference, pp
53-63, Chennai, Tamilnadu.
[6] 3GPP TS36.212, “Evolved Universal Terrestrial Radio Access (EUTRA); Multiplexing and Channel
Coding (Release 8)”.

20
Authors
Mr.S.Syed Ameer Abbas has completed his M.E degree in Applied Electronics and now
pursuing his Ph.D. in the area of VLSI Signal processing at Anna University, Chennai. At
present he is working as Assistant professor in ECE department, Mepco Schlenk
Engineering College, Sivakasi Tamil Nadu, India. He has published more than 35 papers in
International Journals, International and National Conferences. He has been as Principal
Investigator and Co-ordinator for two AICTE projects.
Beril Sahaya Mary. M received the B.E degree in the department of Electronics and
Communication Engineering. Now she is currently doing her M.E degree in
Communication Systems in Mepco Schlenk Engineering College, Sivakasi, Tamilnadu,
India.
Rahumath Nisha .J received the B.E degree in the department of Electronics and
Communication Engineering. Now she is currently doing her M.E degree in
Communication Systems in Mepco Schlenk Engineering College, Sivakasi, Tamilnadu,
India.
Dr.S.J.Thiruvengadam has completed his Ph.D.in the area of Signal Processing and
Post Doctoral Studies in MIMO Wireless Communications at Stanford University, USA.
At present he is working as Professor, Electronics and Communication Department,
Thiagarajar College of Engineering, Madurai, Tamil Nadu, India. He has published
papers in more than 12 International journals and 60 International Conferences. He has
been as Principal Investigator for more than 10 Government Projects.

Implementation of Pipelined Architecture for Physical Downlink Channels of 3GPPLTE

More Related Content

Similar to Implementation of Pipelined Architecture for Physical Downlink Channels of 3GPPLTE (20)

More from josephjonse (20)

Recently uploaded (20)

Implementation of Pipelined Architecture for Physical Downlink Channels of 3GPPLTE