Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

IEEE COMMUNICATIONS LETTERS, VOL. 12, NO.

4, APRIL 2008 225

Information-Guided Channel-Hopping for


High Data Rate Wireless Communication
Yuli Yang and Bingli Jiao

Abstract— In this letter, we introduce information-guided the independent and identical complex Gaussian distribution
channel-hopping, a new scheme for high data rate communi- [3][4].
cation over Rayleigh fading channel using multiple transmit
antennas. This scheme is proposed based on the fact that the inde-
pendence character of multi-channel can be used as an additional A. Transmission scheme
information transmitting channel, and the maximum likelihood
decoding can be achieved in a simple way by decoupling of the At the transmitter, the IGCH scheme prescribes the simul-
signals conveyed by different methods. The analysis results prove taneous transmission of two data streams: one is mapped onto
that the capacity behavior of this scheme is better than that of the the index of the transmit antenna selected for transmission
space-time block coding for more than two transmit antennas.
(i.e., in a system with M transmit antennas, log2 M bits can
Index Terms— Capacity, multi-antenna system, information- be used to choose which transmit antenna is transmitting
guided channel-hopping (IGCH), space-time block coding at each time slot), whereas the other stream is encoded
(STBC).
conventionally and radiated by the selected transmit antenna.
We name the former channel symbol and the latter radiated
I. I NTRODUCTION symbol, expressed by x and xch , respectively, where xch ∈
{h1 , h2 , · · · , hM } with hm (m = 1, 2, · · · , M ) for the
I N [1] and [2], space-time block coding (STBC) is proposed
as a remarkable transmission scheme for the multi-antenna
system. It has an elegant mathematical solution for providing
index of the mth channel. If xch = hm , x will be transmitted
over the mth channel, and therefore, the received signal will
full diversity over the coherent, flat-fading channels. However, be expressed by
when the number of transmit antennas is more than two, STBC y = hm x + nm , (1)
incurs a loss in capacity[2][3], that is, its capacity is smaller
where x ∈ {s1 , s2 , · · · , sN } is the radiated symbol corre-
than the true channel capacity computed by [4] and [5]. The
sponding to an N -point modulation constellation with symbol
main reason behind this is that the equivalent code rate of 2
power of σX , and nm is the received additive white complex
STBC can not achieve the maximum possible transmission 2
Gaussian noise with zero-mean and variance of σN .
rate for more than two transmit antennas.
In addressing the issue of high code rate, we proposed a new
transmission method for multi-antenna system, which allows B. The decoding algorithm
additional information is conveyed by the selection among In the IGCH system, the task of decoding is to find not only
multiple independent channels and accordingly is named the radiated symbol but also the channel symbol mapped by
Information-Guided Channel-Hopping (IGCH) scheme. In the the selection of channel. Naturally, the maximum likelihood
following sections, we will describe this transmission scheme (ML) decoding algorithm can be used to estimate the radiated
and its corresponding decoding algorithm in detail. The anal- symbol and channel symbol simultaneously as following:
ysis results of the capacity behavior will show that IGCH can choose x̂ = sn and x̂ch = hm if and only if
be an alternative scheme in the multi-antenna systems with
more than two transmit antennas. d2 (y, hm sn ) ≤ d2 (y, hm sn ) , ∀m = m , n = n , (2)

where d2 (u, v) is the squared Euclidean distance between


II. I NFORMATION -G UIDED C HANNEL -H OPPING S YSTEM
signals u and v. However, when the number of transmit
Consider an M × 1 system, i.e., a system with M transmit- antennas is fixed, the computational complexity of this de-
and 1 receive- antennas. Let hm (m = 1, 2, · · · , M ) be the flat- coding algorithm (measured by the number of multiplication
fading channel coefficient from the mth transmit antenna to the operations) increases exponentially with the number of bits
receiver. We always assume that the channel state information mapped onto the modulation constellation of the radiated
is not known at the transmitter but completely known at the symbol, i.e., the number of multiplication operations in the
receiver, and that all flat-fading channels in the system satisfy decoding is N M .
In order to reduce the computational complexity, the decod-
Manuscript received November 25, 2007. The associate editor coordinating
the review of this letter and approving it for publication was P. Demestichas. ing of IGCH scheme will be divided into two steps: firstly, the
Y. Yang is with the Research & Innovation Center, Alcatel-Lucent Inc., candidate decisions are made in terms of
Shanghai 201206, China. She was with the Department of Electronics, Peking  
University, Beijing 100871, China (e-mail: yang yuli@hotmail.com). 2 h∗m y
B. Jiao is with the Department of Electronics, Peking University, Beijing x̂m = arg min d s, 2 , m = 1, 2, · · · , M,
100871, China (e-mail: jiaobl@pku.edu.cn). s∈{s1 ,s2 ,··· ,sN } |hm |
Digital Object Identifier 10.1109/LCOMM.2008.071986. (3)
1089-7798/08$25.00 
c 2008 IEEE
226 IEEE COMMUNICATIONS LETTERS, VOL. 12, NO. 4, APRIL 2008

6
where ∗ denotes the conjugate operator; and then, choose x̂ =
x̂m̂ and x̂ch = hm̂ if and only if 2x1 STBC
5 2x1 IGCH

1% Outage Capacity (bits/sec/Hz)


m̂ = arg min d2 (y, hm x̂m ). (4) 4x1 STBC
m∈{1,2,··· ,M } 4x1 IGCH
4 8x1 STBC
When the number of transmit antennas is fixed, the computa- 8x1 IGCH
tional complexity of the two-step decoding algorithm is fixed, 3
i.e., the number of multiplication operations is 2M .
2

III. C APACITY A NALYSIS


1
In an M × 1 IGCH system, there are two independent
input signal spaces (i.e., the radiated symbol space X and 0
the channel symbol space Xch ) and one output signal space 0 3 6 9 12 15 18 21
SNR (Eb/N0: dB)
Y . So, the mutual information between the input and output
signal spaces can be expressed by [6] Fig. 1. 1% outage capacity comparisons between IGCH and STBC in M ×1
(M = 2, 4, 8) channel.
I (X , Xch ; Y ) = I (X ; Y |Xch ) + I (Xch ; Y ) . (5)
6
The randomicity of channel symbol makes each channel 2x1 STBC
have the same probability of 1/M to convey the radiated 5 2x1 IGCH
10% Outage Capacity (bits/sec/Hz)
symbols. With a given realization of the flat-fading channels, 4x1 STBC
the capacity of the first term in the right-hand side of Eq.(5) 4 4x1 IGCH
is [4],[5],[6] 8x1 STBC
8x1 IGCH
M 3
1   
C1 = max I (X ; Y |Xch ) = log2 1 + ρ|hm |2 , (6)
p(x) M m=1
2
2 2
where ρ = σX /σN is the average received signal to noise
1
power ratio.
In Eq.(6), to achieve the maximization of mutual informa-
tion as capacity, the radiated symbol, x, is constrained to 0
0 3 6 9 12 15 18 21
be the random variable from independently and identically SNR (Eb/N0: dB)
distributed complex Gaussian ensemble with the probability
Fig. 2. 10% outage capacity comparisons between IGCH and STBC in
distribution function (PDF) of [3],[4],[5] M × 1 (M = 2, 4, 8) channel.
 
1 |x|2
p (x) = 2 exp − σ 2 , (7)
πσX X Eq.(5) is calculated by

where | · | denotes the modulus operator. Thus, when the mth 
  
channel is selected to convey the radiated symbol, the received C2 = I (Xch ; Y ) 
p(x)= πσ12 exp − |xσ2|2
X X
signal will satisfy a complex Gaussian distribution with PDF
M


of   1 p (y|xch = hm )
= p (y|xch = hm ) log2 dy ,
1 |x|2 M m=1 y p(y)
p (y|xch = hm ) = 2
exp − 2
, (8)
πσm σm (10)
where σm2
= |hm |2 σX2
+ σN2
(m = 1, 2, · · · , M ) denotes the where p (y|xch = hm ) and p(y) are presented in Eq.(8) and
variance of the signal received through the mth channel. Since Eq.(9) respectively.
every channel is selected with the same probability of 1/M , In conclusion, with a given realization of the flat-fading
the average PDF of the received signal is channels, the capacity of the IGCH scheme is
M CIGCH = C1 + C2 , (11)
1 
p (y) = p (y|xch = hm )
M m=1 where C1 and C2 are given by Eq.(6) and Eq.(10), respectively.
M   (9) Next, we will report the numerical results of capacity
1  1 |x|2 behaviors for the comparison between IGCH and STBC. For
= 2
exp − 2 .
M m=1 πσm σm the M × 1 channel case, STBC capacity is [3]
 M

Therefore, with a given realization of the flat-fading chan- ρ  2
CSTBC = R · log2 1 + |hm | , (12)
nels, the capacity of the second term in the right-hand side of M m=1
YANG and JIAO: INFORMATION-GUIDED CHANNEL-HOPPING FOR HIGH DATA RATE WIRELESS COMMUNICATION 227

7
rate, so its capacity can not achieve the optimal performance,
2x1 STBC while the equivalent code rate of IGCH is increased due to
6
2x1 IGCH more additional information conveyed by the selection among
Ergodic Capacity (bits/sec/Hz)

4x1 STBC more transmit antennas. Therefore, IGCH capacity is larger


5
4x1 IGCH than STBC capacity in these cases.
8x1 STBC
4 8x1 IGCH
IV. C ONCLUSION
3
Due to the capacity loss in STBC system for the cases of
more than two transmit antennas, the optimal design with
2
respect to capacity is still an open problem. Motivated by
1
this, a new transmission method named Information-Guided
Channel-Hopping (IGCH) scheme is presented in this letter.
0 In this scheme, the selection of the channel among M in-
0 3 6 9 12 15 18 21
SNR (Eb/N0: dB) dependent fading channels conveys log2 M bits of additional
information besides log2 N bits of basic information mapped
Fig. 3. Ergodic Capacity Comparisons between IGCH and STBC schemes onto the conventional N -point modulation constellation. By
in M × 1 (M = 2, 4, 8) channel.
the capacity performance comparison between IGCH and
STBC, we show that IGCH capacity is larger than STBC
where R is the equivalent code rate of STBC. For complex capacity for the cases of more than two transmit antennas,
signals, R = 1 only occurs in the two transmit antennas case. because the equivalent code rate is increased in the IGCH
If there are more than two transmit antennas, R = 1/2, and system and decreased in the STBC system with the number
only in the special cases of three and four transmit antennas, of transmit antennas.
R = 3/4.
To begin with, in Figs.1 and 2, we provide the 1% and ACKNOWLEDGMENT
10% outage capacity comparisons between IGCH and STBC The authors would like to thank the anonymous reviewers
in 2 × 1, 4 × 1 and 8 × 1 antenna cases. Furthermore, in Fig.3, for their valuable comments to improve the presentation of
we provide the ergodic capacity comparisons between IGCH this letter. Thanks also to Dr. Xiaofu Wu and Dr. Meng Ma
and STBC in 2 × 1, 4 × 1 and 8 × 1 antenna cases. Here, for their help with the edit.
the outage capacity, Cout,q , is defined as the information rate
that is guaranteed for (100 − q)% of the channel realizations,
R EFERENCES
i.e., Pr (C ≤ Cout,q ) = (100 − q)%, and the ergodic capacity
is defined as C̄ = E {C}, where E denotes the expectation [1] S. M. Alamouti, “A simple transmit diversity technique for wireless
communications,” IEEE J. Select. Areas Commun., vol. 16, no. 8, pp.
operator and the expectation is with respect to the channel 1451-1458, Oct. 1998.
realizations [7]. [2] V. Tarokh, H. Jafarkhani, and A. R. Calderbank, “Space-time block codes
As shown in Figs.1, 2, and 3, IGCH capacity is smaller from orthogonal designs,” IEEE Trans. Inform. Theory, vol. 45, no. 5,
pp. 1456-1467, July 1999.
than STBC capacity in the 2 × 1 channel case, and however, [3] S. Sandhu and A. Paulraj, “Space-time block codes: a capacity perspec-
is larger than STBC capacity in the M × 1(M > 2) antenna tive,” IEEE Commun. Lett., vol. 4, no. 12, pp. 384-386, Dec. 2000.
cases. [4] G. J. Foschini and M. J. Gans, “On limits of wireless communication in
Because the logarithm is a concave function over the a fading environment when using multiple antennas,” Wireless Personal
Commun.: Kluwer Academic Press, no. 6, pp. 311-335, Mar. 1998.
considered parameters, C1 in Eq.(6) is smaller than the STBC [5] I. E. Telatar, “Capacity of multi-antenna Gaussian channels,” Eur. Trans.
capacity in Eq.(12). In addition, C2 in Eq.(10) is a small Telecomm. ETT, vol. 10, no. 6, pp. 585-596, Nov. 1999.
increment to C1 , so IGCH capacity is not larger than STBC [6] R. M. Fano, Transmission of Information: A statistical Theory of Com-
munications. New York: John Wiley and Sons, 1961, Chapter 2.
capacity in the 2 × 1 antenna case. However, for the M × [7] A. J. Paulraj, D. A. Gore, R. U. Nabar, and H. Bolcskei, “An overview
1(M > 2) antenna cases, the equivalent code rate of STBC is of MIMO communications–a key to gigabit wireless,” Proc. IEEE, vol.
only 3/4 or even half of the maximum possible transmission 92, no. 2, pp. 198-218, Feb. 2004.

You might also like