Вы находитесь на странице: 1из 6

1

Capacity Enhancement of MIMO system using


water-filling Model
Kuldeep Kumar Manwinder Singh
Rayat Institutes Of Engg. and IT, Rayat Institutes Of Engg. and IT,
Ropar, Punjab Ropar. Punjab
Email: kuldeep_heer@yahoo.com Email: Singh.manwinder@gmail.com

Abstract
Here we discuss the a general theoretical framework for
power budget allocation in the wireless cellular network
,where multiple access points or small base stations send
independent coded information to multiple mobile
terminals through orthogonal Code division multiplexing
channels. Here in we study, the outage probability of a
MIMO system with the change in the power budget also
we would analyze the mean capacity of a system and the
effect with the water filling algorithm. For independent
continuous fading channels, we show that as we increase
the power budget in the water filling algorithm the mean
capacity of the system increases also the outage capacity
of a system reduces for the water filling algorithm as the
power budget in increased. Hence efficiently Power
budget allocation at the transmitter increases the system
capacity and reduces the outage probability.
Key Words water filling, outage probability, Multi Input
Multi Output (MIMO), system capacity .Signal to Noise
Ratio (SNR), Power Budget
1. INTRODUCTION
It has been shown that Multiple-Input Multiple-Output
(MIMO) systems is used to support higher data rate in
the MIMO system we keep the transmit power budget and
performance requirement same as compared to a Single-
Input Single-Output (SISO) systems. If we compare the
MIMO system with a SIMO system the MIMO system
require lesser transmit power than the SISO system. If we
consider the energy consumption in total which will
includes the Transmit energy and the circuit energy then
we need to find out which system is more energy efficient
as there is a huge complexity involved in the circuit of a
Multiple-Input Multiple-Output (MIMO) systems .The
MIMO system has multiple transmit and receive antennas
and Orthogonal Frequency-Division Multiplexing
(OFDM), is used it is sought as a one of the solution for
increasing the capacity and data rate of a system in a
environment where the communication take place in a
frequency-selective fading environments and there can be
probably more chances for data corruption. It has been
found out from the research and the result that Multiple-
Input and Multiple-Output (MIMO) [2] can be effectively
used to increase the capacity of the system by a factor of
the minimum number of transmitter and receiver
antennas attached in the MIMO system as compared to a
Single-Input Single-Output (SISO) system that has flat
fading or frequency selective fading environment or
narrowband channels, OFDM can also increase diversity
gain and minimize the inter-symbol interference on a
time-varying multi-path fading channel. When we know
the parameters of the channel both at the transmitter end
and at the receiver end, we can further increase the
capacity of MIMO OFDM systems by assigning power at
the transmitter according to the water-filling algorithm to
the channels . At transmitters, we send the transmitted
signals of all the carriers that are Eigen beamformed
independently to orthogonal modes of all the channels in
a MIMO OFDM systems. In a MIMO OFDM we use a
improved power allocation scheme called water filling as
compared to the classical water-filling power allocation
scheme .In a MIMO system the results show that there is
increase in the MIMO capacity with water filling. In order
to apply the water filling algorithm we need to know the
channel parameters . The capacity can further be
increased by assigning the power to the transmitted
signals to the orthogonal eigen values according to the
water filling rule.We transmit the different signals at the
transmitter which are eigen beamforced independently to
a orthogonal mode in space time and is sent
simultaneously by multiple carriers .A novel eigen modes
are used to directly start the adaptive power allocation to
the data bits allocation as such this is usually done with a
single carrier transmitter .Here we have two type of water
Kuldeep Kumar* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 1, 092 - 097
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 92
I
J
A
E
S
T

2
filling models one is the classical water filling model and
other is the improved water filling model. Of all the
power allocation schemes the greedy power allocation
scheme is the optimal one to transmit the bits. The article
is organized as follows. In section 2, discusses the MIMO
capacity and apply the water filling algorithm to increase
the power. Section 3, we describe the water filling In
section 4, we discuss the outage probability of a MIMO
system and section 5, we conclude our discussion with the
results.
2. MIMO CHANNEL CAPACITY
Lets us have a MIMO system with M ,N transmitter and
receiver antennas respectively . The space time codes are
used to encode the data along with the space time
direction with T data symbols. The model of
communication system is given as

e s w r + = *
(1)
Where r is the received signal and s is the sent signal .* is
the convolution and e is the error function .
w
is the
channel matrix .When M=N=1 then is a SISO .We will
discuss about a MIMO system where M=N
1 =
here it is
understood that we can increase the capacity of a MIMO
system without changing the bandwidth and transmit
power rather we can put more antennas at transmitter and
receiver side to increase the capacity.One of the important
part of a MIMO system is the channel matrix denoted by
) , ( t H t
.Consider the impulse response between the i
th
(i
= 1, 2, ..., n
r
) Receive antenna and j
th
(j = 1, 2 ..., n
t
)
transmit antenna,The MIMO channel matrix is given by
H(
t
, t)
( ) ( ) ( )
( ) ( ) ( )
( ) ( ) ( )

=
t h t h t h
t h t h t h
t h t h t h
t H
t r r r
t
t
n n n n
n
n
, ....... , ,
.
.
.
.......
......
.....
.
.
.
.
.
.
, ....... , ,
, ....... , ,
) , (
, 2 , 1 ,
, 2 2 , 2 1 , 2
, 1 2 , 1 1 , 1
t t t
t t t
t t t
t
(2)
In frequency domain, the channel is made of a complex
matrix having the independent and identically distributed
entries with unit variance and zero mean. The generalized
capacity formula for a MIMO system is given by

Hz s b HH I er C
t
nr
/ / det log

|
|
.
|

\
|
|
|
.
|

\
|
+ =
n
p
(3)
In this equation, deter means the determinant, I is a
identity matrix and means conjugate transpose of the
matrix.

Hz s b C
n
k
k
t
/ / 1 log
1
2
2 2
=

|
|
.
|

\
|
+ > v
n
p
(4)
The capacity formula for where the transmit and receive
antennas are the same i.e (n
r
= n
t
= n) is given as [5].
| | Hz s b C
n
/ / . 1 log
2
2 2
v p + =
(5)
Here we describes some of the very important properties
of channel matrix H for a MIMO system . We describe in
this section some statistical properties of the channel
matrix H. The channel matrix HH using eigen value
decomposition can be rewritten as follows
HH = E^E (6)
Where E is eigenvector matrix of orthonormal columns
and ^ is a diagonal matrix having the Eigen values on the
main diagonal. Using this notation above , the capacity of
MIMO channels can be written as below :
{ }

|
|
.
|

\
|
. + =

2
det log E E I E C
t
nr h
n
p
(7)
In general channel matrix rank is given by is given by
} {
t r
n n k H rank , min ) ( s =
(8)
Using the equation (6), with the fact known that the
determinant of a unitary matrix is 1, the capacity
expression can be written as:

)
`

|
|
.
|

\
|
+ =

=
t
t
k
i
H
n
E C
p
1 log
1
2
(9)

)
`

|
|
.
|

\
|
+ =

=
t
t
k
i
H
n
E o
p
1 log
1
2
(10)
Kuldeep Kumar* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 1, 092 - 097
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 93
I
J
A
E
S
T

3
Here
A
is the diagonal matrix and
i

are the eigen


values and
i
o
are the squared singular values of the
diagonal matrix

.
. When the channel matrix is known
at the transmitter, the maximum capacity of a MIMO
channel can be achieved by using the water-filling
algorithm [3] on the transmit covariance matrix. The
capacity is then given by:

)
`

|
|
.
|

\
|
+ =

=
t
t
i
k
i
H
n
E C
p
c 1 log
1
2
(11)
Where i is the i
th
channel and
i
c
is a scalar that shows
the portion of the available transmit power that is
available in the outgoing i
th
channel. So, by using the
water-filling algorithm we can meet the total power
constraint.
3. Water filling Power allocation algorithm
The adaptive modulation can be implemented in two
aspects , the adaptive power allocation technique is such
that the transmitted power maximize the number of the
transported bits. Here in this study we study the SNR and
the Capacity of a 4x4 MIMO system. We would find out
the mean capacity of the system with and without the
water filling algorithms .Here the power is distributed in
the orthogonal eigenmodes in order to maximize transmit
bits to get the maximum capacity of the MIMO
system.The transmit power required to send the c bits for
an additive white Gaussian noise (AWGN) channel with
M-QAM modulation, is given by the following formula.
) 1 2 (
4 3
) (
2
1
2

|
.
|

\
|
=

e e
P
Q c P
o
(12)
Q is called the as complementary error function. Then,
for given total transmit power of the MIMO system is
given by , the number of bits transported by the AWGN
channel, can be derived by the Eq. 12, as shown in the
following formula

|
|
.
|

\
|
|
|
.
|

\
|
|
.
|

\
|
+ =

2
1
2
4
3
1 log
e
P
Q
P
f loor c
o
(13)
As we have to maximize the total number of bits to be
transported , The results shows that improved water-
filling power allocation scheme is given on the base of
classical water filling schemes. As per the scheme, the
adaptive power and bit allocation are conducted in two
steps. Firstly, the initial power allocation is given by
classical water-filling scheme, and then is, the first step is
is to allocate the power for all the orthogonal eigen
modes according to the classical water-filling scheme to
all the eigen values. Then, after finding the transported
bits at channel eigenmodes,we allocate the residual power
by reallocated among these eigenmodes to transport
additional bits.Here we see that the Mimo system with
water filling has a greater mean capacity as compared to a
MIMO system without water filling .Also as per the
results in Fig.1,2, 3 the mean capacity of a MIMO system
increases with the increase in the power budget at the
input of the transmitter. Following figures are the result
that we get for the capacity of a 4X4 MIMO system by
varying the power budget.

Fig. 1.Mean Capacity at Pt=0.2 with water filling
algorithm
The figure 1,2,3,4 shows the mean capacity of the system
when the power budget P
t
0.2,0.3,0,5 and 1 with and
without the water filling model ..Figure 4 shows the mean
capacity of a 4x4 MIMO system for all the water filling .
By observing the different figures we see that as the
power budget is increased the capacity of the system also
increases
-10 -5 0 5 10 15 20 25 30
0
5
10
15
20
25
30
35
Mean Capacity vs SNR
SNR (dB) --->
M
e
a
n

C
a
p
a
c
i
t
y

b
p
s
/
H
z




-
-
-
>


4x4 MIMO at Pt=0.2
4x4WF MIMO at Pt=0.2
Kuldeep Kumar* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 1, 092 - 097
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 94
I
J
A
E
S
T

4

Fig. 2.Mean Capacity at Pt=0.3 with water filling
algorithm

Fig. 3.Mean Capacity at Pt=0.5 with water filling
algorithm

Fig. 4.Mean Capacity at different P
t
with water filling
algorithm
4. OUTAGE PROBABLITY
The outage probability is defined as that an outage will
occur within a given time period. It can also be defined as
temporary suspension of the operation The signal outage
probability is calculated if one knows the probability
distribution of the fading either Rayleigh or Racian .The
outage will take place if the signal drops below the signal
of the noise power level.
Calculation of the outage probability involves that we find
the probability that the signal-to-interference ratio that
drops below a certain threshold. As all signals that are
fading have fluctuating signal powers, It mean that in
order to find the outage we must integrate the probability
density functions of all signals involved in fluctuating.
The Outage formula is given by
) / ; ( 1 ) ( 1 ) ( p z f Laplace z SNR P out P = > =
(14)
Where SNR is Signal to noise ratio,
z
is the minimum
SNR required ,
f
is the probability distribution function
of the power,
p
is the mean power of the wanted signal.
In order to find the outage probability of a MIMO
network we take a 4x4 MIMO and would try to analyze
the variation of the outage probability with the variation
of power budget. It has been shown in the results that with
the increase in th power budget the outage probability
decreases. Also it is found from the graphs that the outage
of the waterfilling model is more in a MIMO than the
MIMO system.
The outage probability in a MIMO system is given by ,

+ =

HH
N
SNR
I P outage P
t
r r
det( log ) (
2
(15)
Where H is the channel matrix,I is the information
transmitted,
t
N
is the number of the transmit antennas
and
* 1
) * ( H H H H

=

Following are the graphs


showing the outage capacity (probability) variation of a
SISO,SIMO, MISO, MIMO system with SNR

-10 -5 0 5 10 15 20 25 30
0
5
10
15
20
25
30
35
Mean Capacity vs SNR
SNR (dB) --->
M
e
a
n

C
a
p
a
c
i
t
y

b
p
s
/
H
z




-
-
-
>


4x4 MIMO at Pt=0.3
4x4WF MIMO at Pt=0.3
-10 -5 0 5 10 15 20 25 30
0
5
10
15
20
25
30
35
Mean Capacity vs SNR
SNR (dB) --->
M
e
a
n

C
a
p
a
c
i
t
y

b
p
s
/
H
z




-
-
-
>


4x4 MIMO at Pt=0.5
4x4WF MIMO at Pt=0.5
-10 -5 0 5 10 15 20 25 30
0
5
10
15
20
25
30
35
40
SNR (dB)---->
M
e
a
n

C
a
p
a
c
i
t
y

b
p
s
/
H
z

-
-
-
-
>


4x4WF MIMO at Pt=0.2
4x4WF MIMO at Pt=0.3
4x4WF MIMO at Pt=0.4
4x4WF MIMO at Pt=0.5
4x4WF MIMO at Pt=1.0
Kuldeep Kumar* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 1, 092 - 097
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 95
I
J
A
E
S
T

5

Fig. 5.Outage probability at Pt=1 for different systems

Fig. 6.Outage probability at Pt=0.5 for different systems

Fig. 7.Outage probability of MIMO system at Pt=0.2

Fig. 8.Outage probability of MIMO system at Pt=0.2


Fig.9.Outage probability of MIMO WF system
5. CONCLUSION
This paper we have developed an understanding and
described the Mean capacity allocation in a wireless
cellular network based on the water filling power
allocation in order to enhance the capacity of a MIMO
systems with different channel assumptions. Here each
transmitter decide the distribution of power to the several
independent fading channels. We studied the change in
the Mean Capacity of the system with the power budget
of a system. Also, an improved water-filling scheme is
proposed for increase in the system capacity. Results
indicates that the water-filling scheme has better capacity
than without water filling at greater value of power
budget. We also discussed the variation of the outage
probability of the system .
2 4 6 8 10 12 14 16 18 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
SNR
O
u
t
a
g
e

C
a
p
a
c
i
t
y


1x1 SISO at Pt=1.0
1x4 SIMO at Pt=1.00
4x1 MISO at Pt=1.0
4x4 MIMO at Pt=1.0
4x4WF MIMO at Pt=1.0
2 4 6 8 10 12 14 16 18 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
SNR
O
u
t
a
g
e

C
a
p
a
c
i
t
y


1x1 SISO at Pt=0.5
1x4 SIMO at Pt=0.5
4x1 MISO at Pt=0.5
4x4 MIMO at Pt=0.5
4x4WF MIMO at Pt=0.5
2 4 6 8 10 12 14 16 18 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Outage probability vs SNR for 4 bps/Hz
SNR in dB --->
O
u
t
a
g
e

P
r
o
b
a
b
i
l
i
t
y




-
-
-
>


4x4 MIMO at Pt=0.2
4x4WF MIMO at Pt=0.2
2 4 6 8 10 12 14 16 18 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Outage probability vs SNR for 4 bps/Hz
SNR in dB --->
O
u
t
a
g
e

P
r
o
b
a
b
i
l
i
t
y




-
-
-
>


4x4 MIMO at Pt=0.4
4x4WF MIMO at Pt=0.4
2 4 6 8 10 12 14 16 18 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
SNR
O
u
t
a
g
e

C
a
p
a
c
i
t
y


4x4WF MIMO at Pt=0.2
4x4WF MIMO at Pt=0.3
4x4WF MIMO at Pt=0.4
4x4WF MIMO at Pt=0.5
4x4WF MIMO at Pt=1.0
Kuldeep Kumar* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 1, 092 - 097
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 96
I
J
A
E
S
T

6
6. REFERENCES
[1] J. H. Reed, Software radio: a modern approach to
radio engineering: Prentice Hall, 2001.
[2] A. Paulraj, R. Nabar, and D. Gore, Introduction to
Space-Time Wireless Communications, preprint,
Cambridge University Press, Cambridge, UK, 2003.
[3] S. Ariyavisitakul and C. Li Fung, "Signal and
interference statistics of a CDMA system with feedback
power control," IEEE Transactions on Communications,
vol. 41, pp. 1626-1634, 1993.
[4] J. S. Evans and D. Everitt, "Effective bandwidth-based
admission control for multiservice CDMA cellular
networks" IEEE Transactions on Vehicular Technology,
vol. 48, pp. 36-46, 1999.
[5] T. S. Rappaport, Wireless Communications: Principles
and Practice, 2nd Edition ed.: Pearson Education, 1996.
[6] B. Hashem and E. S. Sousa, "Reverse link capacity
and interference statistics of a fixed-step power-
controlled DS/CDMA system under slow multipath
fading," IEEE Transactions on Communications, vol. 47,
pp. 1905-1912, 1999.
[7] B. Hassibi and B. M. Hochwald. How much training
is needed in multiple-antenna wireless links IEEE
Transactions on Information Theory, 48(4):951 963,
April 2003.
[8] A. Paulraj, R. Nabar, and D. Gore, Introduction to
Space-Time Wireless Communications preprint,
Cambridge University Press, Cambridge, UK, 2003.
[9] S. Zhang, Y. Chen, and S. Xu, Improving Energy
Efficiency through Bandwidth, Power, and Adaptive
Modulation, in IEEE 72nd VTC 2010Fall, 2010
[10] Fabien, Hoshyar, Reza and Tafazolli, (2009) A
closed-form approximation of the outage probability for
distributed MIMO systems In: IEEE 10th Workshop on
Signal Processing Advances in Wireless
Communications, 2009. SPAWC '09, 21-24 June 2009
[11] M. Kaynia, G. E. Oien, A. J. Goldsmith, and N.
Jindal, "Outage probability in wireless networks", Proc.
IEEE Information Theory Winter School (ITWS), Loen,
Norway, March 2009.



Kuldeep Kumar* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 1, 092 - 097
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 97
I
J
A
E
S
T

Вам также может понравиться