Received 22 January 2014, Revised 13 May 2015, Online Published 29 May 2015

1.
INTRODUCTION
The emerging trends in the wireless communication system have opened the way of 4 th generation (4G) communication systems. The
seamless advantages of the 4G system can meet the customer demands. The 4G systems can support up to 1 Gbits/s with high security. One
of the main challenges of the 4G system is its accessibility of heterogeneous network. For this, an efficient transceiver system is required.
Software-defined radio (SDR) is an efficient radio, which is adapt to any situation.
Since SDR contains software functions, it can change the parameters at any time according to the situations. The SDR can be used for
military applications, 4G, and next generation wireless communication for fast and accurate communication. SDR is the technique used for
replacing hardware components into a software program, which reduces the complexities in the implementation and testing 1. There are lots
of advantages for the replacement of hardware to software, low complexity, high efficiency, improved functionality, less area, and power
consumption, low manufacturing cost, etc. The software allows to change the parameters of the receiver at run time itself according to the
application2. The reconfigurable architectures like FPGA, DSP, and general purpose processors are used as the design platforms.
OFDM is a modulation technique, which can be used in high data rate communication systems. It is a multi-carrier modulation system
with high bandwidth efficiency and is robust to frequency selective fading 3. One of the major drawbacks of the OFDM techniques is the
highest peak-to-average power reduction (PAPR) ratio of the transmitted signal. It causes nonlinear distortion while passing through the
power amplifier, hence reduces the BER performance. There are many algorithms used to reduce the peak power of the OFDM signal, such
as clipping, filtering tone rejection, peak windowing, companding, selective mapping (SLM) 4,5, etc. Partial transmit sequence (PTS) etc 6.
The above algorithms give the moderate performance.
The peak windowing and the filtering methods are the simplest methods among the above mentioned algorithms. In clipping and filtering
method, some data loss happens. The clipping process cuts off the high peaks in the signal. It affects the information signal. In the peak
windowing method, the window cannot accommodate when the consecutive peaks come. In our proposing method, the adaptive window
can accommodate at any place by changing the window parameter. This will reduce the losses in data and will improve the performance of
the system. The filter will remove the out-of-band radiation of the signal7-9.
All other methods have the system complexity and computational complexity 10. The Viterbi algorithm is the fastest decoding algorithm
using Trellis structure used for communication system application.
2. SDR ARCHITECTURE
Software-defined radios are flexible in nature. Even at run time these can adapt to the changes in the system 11. The SDR is a basic platform
for the cognitive radio in the system can automatically adaptto any changes in the environmental conditions. The general architecture of
SDR is shown12 in the Fig. 1. The physical layer operations depend on the
Received 22 January 2014, revised 13 May 2015, online published 29 May 2015
233
Def. SCI. J., Vol. 65, No. 3, may 2015
234
environmental conditions as well as the application. The SDR
can change its parameters at any time without disturbing the other
functions in the system. In SDR all functions are designed as
software coding instead of hardware component. Therefore the
system can update itself according to the needs at any time with
low cost and complexity. The main functions of SDR are described.
2.1 Smart Antennae
The baseband controllers control the antenna operations. At the
transmitter side, the antenna makes an interface between the
network (NET) layer and the MAC layer on the receiver side. The
antenna can achieve the signal by space division multiplexing.
2.2 RF Modules
The RF module is used to assist the smart antenna. The RF
module is having the following functions, Digital frequency
synthesizer, pre-amplifier, power output, RF conversion.
2.3 Digital up/down Converters
A/D is used to convert the digital signal into an analog signal.
The signal from the antenna is in the RF range. Then the RF signal
is converted into IF signal. The high speed converters are designed
according to the sampling theorem. The up converters do the
reverse operation of this.
2.4 Software Packages
SDR uses a distributed type of architecture and processes the
information transmitted using common object request broker
architecture (CORBA), protocol with control packages and system
interface packages. There are two types of software downloaded
used in SDR, static technology and semi-static technology2.
2.5 Digital Signal Processing
The digital signal processing (DSP) functions can be
implemented on an FPGA. All the physical layer functions like
modulation/demodulation,
encoding/decoding,
and
multiplexing/de-multiplexing will be performing in this block.
3. OFDM
The proposed architecture of OFDM is shown in the Fig. 2.
The conventional system SLM techniques and filtering method
have been used. Here filter with windowing methods have been
used. This will reduces the ripples in the frequency response of the
signal.
In the SLM scheme, the information bits have been divided into
number of sequences in a particular order and one has to find the
IFFT of each sequence separately. Then one can select one
sequence with a minimum PAPR ratio 13-16. In general, for each FFT
requires N/2 log2 N multiplications and N log2 N addition. The
number of FFTs will increase, and then computational complexity
also increases. In the proposed method, the FFT is calculated only
one time for the input sequence and it is passed through the filter.
The filter can reduce the out-of-band radiations in the transmitted
signal.
In windowing techniques Kaiser window will show the best results.
One of the disadvantages of the peak windowing method is that,
when consecutive peaks come in the signal, then the window
cannot apply continuously. In the proposed method, the widow is
adaptive in nature. The adaptivity of the window allows changing
the shape of the window. This can easily accommodate when the
high peak appears. The peak detector in the system can find the
positions of every peak. According to the peak value position one
applies the Kaiser window function17-18.
The filter passes the signal through the window. The window
frequency response is given by the Eqns (9-11). The filter with
window technique filters off the unwanted noise added in the
information signals at the transmitter and helps to keep the side
information of the signal. Most of the existing methods fail in the

transmission of side information.
4. SYSTEM MODEL OF OFDM
The input of the OFMD system from the source is represented as
s(t). The main processing steps in the OFDM system are described
below
1. Mapping with BPSK/QPSK/16-QAM,
2. Serial to parallel conversion,
3. IFFT,
4. Find the peak locations,
5. Apply the filter,
6. Apply the Kaiser window, with varying values,
7. Add cyclic prefix for reducing the inter carrier
interference and passed through high power amplifier,
8. Transmitted through the AWGN channel,
9. Remove the CP,
10. De-mapping with BPSK/QPSK/16-QAM,
11. Decoding with Viterbi algorithm.
The detailed steps are explained. Generally the OFDM system has
multiple sub-carriers. The number of multiple sub-carriers can be
represented as M. Then the IFFT of the transmitted signal is given
by the equation
12(/)0
1()()MjmMmxkXmeM== 0 k M 1 (1)
Figure 2. Proposed architecture of OFDM transmitter, filter

with Kaiser window.
Figure1. SDR general architecture.
Joseph & Kumar: FPGA-BasedRealisationofSDR withOFDM Tranceiver
235
where, 1j= , X(m) = [X(0), X(1),X(M-1)] is the input signal
for IFFT and k is the discrete time index.The peak-to-average
power ratio of the transmitted OFDM signal is given by,
22max()()peakaveragexkPPAPRPExk== (2)
where, peakPis the peak power of the transmitted signal and averagePis
the average power of the transmitted signal. In OFDM, the
information signal is first mapped by using QPSK or BPSK
modulation. The expression for both modulations can be expressed
as19-20.
4.1 Case (i): BPSK Mapping

In BPSK modulation, the signal is mapped with either 0 or 1.
2()cos(2(1)),0,1bncbEItftnnT=+= (3)
For binary 0, 02()cos(2),0bcbEItftnT==
For binary 1, 12()cos(2),1bcbEItftnT==
where, fc is the frequency of the Carrier wave. For M carrier signal the frequencies are fc1 fc2 fM-1. Tb is the Time
period.
4.2 Case (ii) QPSK Mapping
In QPSK modulation the signal can be represented using four symbol, can be represented as
2()cos(2/4(21),1,2,3,4qncqEItftnnT=+= (4)
The signal can be represented as
122()cos(2)2()cos(2)cscstftTtftT==
4.3 Case (iii): QAM Mapping
The mapping of the system can also do using QAM modulation. The QAM was also designed using a function
and it can be selected whenever it needs. Where, In (t) is the modulated form of the information signal S(t) from the
source. After the modulation, the signal is passed through the IFFT.
The response of FIR filter is given by
()()()Mkynhkxnk== (5)
The frequency domain equivalent is equal to
()()()YkHkXk= (6)
where, Y(k) is the N point DFT and 2mkM=
For linear phase FIR filter,
10
()()jHHe=, where,()n= (7)
()
The conditions for the linear phase FIR filter is

()(1)hnhMn= , n = 0 , 1 , 2, .. M 1 (8)
The Eqns (5) (8) represent the FIR filter design. The output response will show some ripples in the pass band and
stop band. These ripples can be reduced using the truncation method. For this, one can use windowing techniques.
Kaiser window can be used in this method. The response of the Kaiser window is given in the Eqn (9). The Kaiser
window function is given by
01/21,01()/20,LILwlLotherwise<= (9)
where, I0(.) is the zeroth order modified Bessel function of 1st kind, and is the shaping parameter and is given by
0.1102(8.7),200.5842(21)0.07886(21),21500,21HHHHHH>=+<(10)
and L is approximately equal to
8/2.285LH= (11)
After applying windowing function and filtering, the signal can be represented as () ifYk. Therefore the total filter
response Eqn (6) become
()()()()()()()()()()()()iiiifiiiitiiiiYkXkJkYkHdkXkJkYkHkXkJkWk=== (12)
where, ()iYkis the filtered signal, ()iXk is the input signal from IFFT ()iJk is the impulse response of the filter, () iHdk
is the response of the filter, () iWkis the window function and () itYk is the transmitted OFDM signal. The transmitted
signal is sampled by a factor U before transmitting the signal through the channel.
After the windowing function, the signal is added to the cyclic prefix for avoiding the interference in the OFDM
signal. Some zeros are added with starting and ending of each symbol. The symbol can be represented as y(k). At
the receiver side the zeros (CP) will be removed.
Consider the receiver side, the block diagram is shown in Fig. (3). Here, v=n/U.
OFDM has multi-path channel, at the receiver the over sampling response is given in Eqn (13).The Impulse response
of receiving signal is given by, g(n/U), n=0..... UM-1.
Figure 3. OFDM receiver architecture. Def. SCI. J., Vol. 65, No. 3, may 2015 236
10(/)(/)(/)(/)()(/)(/)(/)()UMkrnUynUgnUonUnkUMynUgnUykUhU ==+= (13)

where, (/)(/)(/)(/)ynUxnUhnUwnU=
where, (n-k)UM represents the (n-k) modulo UN, U is the over sampling factor, r(n/U) is the received signal, g(n/U) is
the impulse response of the channel, h(n/U) is the filter response and w(n/U) is the window function. Then one has
to find the FFT of the received signal and is given by
12(/)0()(/)UMjkmUMmRkrnUe==0 k UM 1 (14)
where r(n) is a product of y(n) , g(n) and o(n) is the noise. The y(n) and g(n) can be by calculating the FFT
()()()()NjknNkNjknNkynYkegnGke====, where, 0 n N-1 (15)

After finding the FFT, the signal is passed through a decoder and demodulation. On the receiver side, Viterbi
algorithm is used to decode the received signal.
5. ViterbiDecoder
On the receiver side of the OFDM Viterbi decoder is used. This algorithm finds the most likelihood state
transitions in the Trellis structure. In OFDM, the Viterbi is used for decoding the sequence at the receiver and detects
the sequence of symbols. The Viterbi algorithm mostly applied when the problem is formulated by a Markov chain 2123.
Markov process can be represented as
111(/,)(/)kkkkkPP++= (16)
is representing different states.
The Viterbi decoder consisting of the branch metric unit (BM), add and compare unit, and one memory unit.
The BM unit is used to calculate metrics in all branches from input data. The ACS unit will calculate the path metric
of current states and two previous states. Then compare the two metrics and select the minimum metric. The
memory unit is used to store all metric values. The output unit is used to transfer all the bits from right-side 23.
5.1 Algorithm
The steps involved in the Viterbi algorithm are include:
1. Initialise the parameters,
2. Branch metric calculation. (using Hamming distance),
3. Load the branch metric,
4. ACS,
5. Check whether the states end, if yes then go to step 6 otherwise go to step 3,
6. Check whether the trellis states end, if yes then go to step 7 otherwise go to step 2,
7. Collect the decoded bit,
8. End.
The process is a Markov chain if and only if the form a countable sequence set. The transitions in the chain can
be represent either as a state diagram or using Trellis diagram. In Trellis the distance can be represented as
12(/)012(/)0
YS= (17)
Viterbi algorithm increases the speed of the decoder and easily traces out the original signal. The final output of the
Viterbi decoder is represented as R(k).
6. RESULTS
The Fig. 4 shows the simulation results. Here the modulations are taken as BPSK, QPSK. The FFT size can be
varied according to the application. The FFT is scalable, can be varying any time. The simulation shows the
comparison of the existing systems such as filtering and SLM method for a different variations. Also shows the
variation of bit error rate (BER) with various point FFT. Table 1.
Figure 4 shows the BER curve for BPSK with N = 128 and the number of carrier signals = 60. The original signal
and the simulated signal are shown in the Fig. 4. The simulated signal is almost related to the original theoretical
signal. Therefore very less error rate happens at the receiver side. In Fig. 5 the BER performances of the QPSK and
16_QAM are presented. The OFDM performance is improved by using the window technique.
The Fig. 6 shows the BER curve using different channels. The AWGN channel gives better performance than
Rayleigh for all modulations.
The Fig. 7 shows the BER performance of different modulations. This Figure shows the QPSK modulation and
BPSK modulation with AWGN channel for various N values. The Fig. 8 shows the CCDF curve for peak power for
BPSK modulation.It shows the PAPR curve of OFDM signal with N = 128
20Pppp
Table 1. Parameters used for simulation

Parameter
FFT size
Size
16, 64, 128
No of sub carrier
Modulation
Window
60
BPSK, QPSK, 16QAM
Kaiser window

Received 22 January 2014, Revised 13 May 2015, Online Published 29 May 2015

Uploaded by

Copyright:

Available Formats

Received 22 January 2014, Revised 13 May 2015, Online Published 29 May 2015

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Received 22 January 2014, Revised 13 May 2015, Online Published 29 May 2015

Uploaded by

Copyright:

Available Formats

1.

Def. SCI. J., Vol. 65, No. 3, may 2015

information of the signal. Most of the existing methods fail in the

Figure 2. Proposed architecture of OFDM transmitter, filter

Joseph & Kumar: FPGA-BasedRealisationofSDR withOFDM Tranceiver

4.1 Case (i): BPSK Mapping

()()jHHe=, where,()n= (7)

The conditions for the linear phase FIR filter is

10(/)(/)(/)(/)()(/)(/)(/)()UMkrnUynUgnUonUnkUMynUgnUykUhU ==+= (13)

()()()()NjknNkNjknNkynYkegnGke====, where, 0 n N-1 (15)

Table 1. Parameters used for simulation

You might also like