Space Modulation on Wireless Fading Channels

Yawgeng A. Chau and Shi-Hong Yu

Department of Electrical Engineering

Yuan-Ze University, Taiwan 320

Abstract-- To increase the data rate for future wireless two antennas is developed based on maximizing the corre-
systems, a. novel scheme using multiple transmit antennas sponding likelihood function. In Section 111,we present the
for space (digital modulation is proposed in this paper. In
the scheme of space modulation, more than two antennas SSK-BPSK scheme with performance analysis. To exam-
are emp1o:yed to transmit and signal the information bits. ine the performance of the SSK scheme, numerical results
The novel space modulation scheme is termed as Space Shift are illustrated in the section IV. Conclusions are drawn in
Keying (S9K). Numerical results are presented to examine Section V.
the reception performance.
11. SSK AND OPTIMAL RECEIVER
I. INTRODUCTION In this section, binary signaling with two transmit an-

V ARJOUS techniques using antenna arrays have been

developed for the deployment of smart antennas ([l]-
[4]). The space and time transmit diversity (STTD) has
tennas is considered, and its optimal receiver is developed
using ML detection.

been employed in the W-CDMA for the third-generation A . SSK Signaling Model
mobile communications, where the transmit diversity tech- Consider the signaling on a slowly nonfrequency-selective
nique is used for the enhancement of downlink perfor- Rayleigh or Rice fading channel. With the SSK modulation
mance. To increase the data rate for future wireless sys- scheme, denote the transmitted signal vector of the antenna
tems, a novel scheme using antenna diversity for space dig- elements’ by
ital modulation is proposed in this paper. In the scheme of
space modulation, more than two antennas are employed to sm(t)= [sml(t) sm2(t)]’ m = 0 or 1 (1)
transmit and signal the information bits. For example, two
for 0 5 t 5 Tb, where s m l ( t ) is the transmitted signal
antennas are used for binary signaling as in the BPSK, or
of the Zth antenna, Tb is the transmission duration. For
binary-FSK modulation on wireless fading channels. The 0 5 t 5 Tb, the binary signals of the two antennas given in
novel space modulation scheme is termed as Space Shift (1) are characterized by
Keying (SSK) .
The principle of the SSK is based on distinct multipath s,1 ( t )= @ c o s 2w,t (2)
characterj stics of different antennas on a wireless fading
channel. Thus, the receiver can utilize the distinct received for m = 0, 1, and
signals from different antennas to discriminate the trans-
mitted information messages, which results in a simple and
cost-effective receiver structure. Therefore, the demodula-
tion and detection of the SSK signaling will be easier to where P denotes the total transmission power, and each
implement. antenna uses half of the power P.
In the context, to present the idea of SSK, binary sig- According to the above signal model of the SSK modula-
naling is considered with one receiver antenna in Rayleigh- tion, the digital information is carried and then transmitted
or Rice-type wireless fading channels. The binary informa- by different number of antennas. In the following context,
tion is transmitted by different number of antennas at the the optimal receiver for SSK demodulation and detection
transmitter site. If the information bit 0 is to be transmit- is devised.
ted, a designed signal is sent only by the first antenna. On For a slowly Rayleigh or Rice fading channel, the fading
the other hand, if information bit 1 is to be transmitted, factor 01 is assumed to be complex Gaussian distributed
the designed signal is sent by both the first and the second with mean bl and variance $ 2 1 . Thus, the received signal in
antennas. At the receiver site, maximum likelihood (ML) a complex form at the receiver site for 0 5 t 5 Tb is given
detector is employed for the estimation of information data by
bits
To increase the data rate, the SSK modulation scheme a1 d P e j w c t + n(t) ifm=O
can also be combined with the traditional modulation
methods, such as PSK, or FSK, to form the SSK-BPSK
T(t) =
{ al@ejwct + a2ejWct+ n(t) if m = 1 (4)

or SSK-FSK modulations. where al,for Z = 1,2, denotes the complex fading factor
In Section 11, the modeling of SSK and the received SSK from Zth transmit antenna to the receiver antenna, and n(t)
signal are described, and the optimal receiver of SSK with ‘The superscript means the transpose of the corresponding vector.

0-7803-7C105-81011$10.000 2001 IEEE 1668

is the additive channel noise that has a complex Gaussian In (9), the random variable y = Iz - 81' has a noncentral
distribution with zero mean. In (4),we assume that the Chi-square distribution with the noncentrality parameter
quadrature and in-phase parts of n(t)have the same power-
spectrum-density N0/2, and a1 and a2 are independent.
The received signal after down-converting in the equiva-
lent base-band form is characterized by ([3])
z={ a1dF+v ifm=O
alO+apdF+v ifm= 1 (5)
The probability of bit error is characterized by
where v has a Gaussian distribution with zero mean and
variance N O ,and the complex fading factors (YL for 1 = 1,2
are to be estimated at the receiver for the detection of
information bits. Under the condition that the information
data bit m has been transmitted, the received signal z has a where Q1(., denotes the first-order Marcum's Q function
a)

complex Gaussian distribution with mean pm and variance (151).

on the event that m = 1 has been transmit-
ok, where po = obi, = O ( b 1 b2), 00" = PC21 +No ted,Conditioned
+ the bit error probability of the SSK receiver is charac-
and 0: = P(C21+ C22) NO + terized by
B. ML Detection
In accordance with the maximum likelihood criterion for
optimal binary detection of the SSK signal, the information
data bit is estimated by minimizing the decision variable

for m = 0 or 1. Thus, the estimated data bit is given by

From the derivation of PssklO, we have

for m = 0 or 1.
Then, conditioned on the event that m = 0 has been
Psskll = Pl
0
Iz - el2 < v (15)

where 8 and 7 are defined by (10) and ( l l ) ,respectively.

transmitted, the bit error probability of the SSK receiver Under the condition that m = 1 has been transmitted, the
is given by random variable y = z - 8 has a noncentral Chi-square
distribution with the noncentrality parameter

Thus,
With some manipulations, the above probability of bit er-
ror reduces to the form

where
Pssk10 = PO
0
Iz - 81' > v) (9) Consequently, the total probability of bit error under
equally likely transmission is
1
bz
8 = f i b 1 - -(Pi21 +NO) pssk = S(Paskl0 -k Psskll)
dFQ2
and

111. SSK-BPSK SCHEME

The above SSK scheme is independent of the traditional
digital modulation methods. To increase the transmission
2Fte{z} and Re{=} mean the real and the image part of z,respec- data rate, we can combine the SSK with BPSK to form the
tively. quadric-modulation SSK-BPSK . If two transmit antennas

1669
are used. for the SSK-BPSK modulation, the received signal where
in 0 5 t 5 T b is configurated as

a 1dFejwct ifm=O
(yl @ej(uct+n) ifm=1
s(t) =
a 1 @ejwct +a 2 @@ct ifm=2
al@ej(&t+*) +a 2 O e ( j w c t + ” ) if m = 3
(19) and
After down-converting, the received signal is

i-(
a 1 d F + w ifm=O
-a1dF+w ifm=1
(20) with
aldF+a2@+2) ifm=2
-aldF-azdF+w ifm=3

that is complex Gaussian distributed with the mean bm Consequently, the final form of the probability of symbol
and the variance e$, where PO = PO, j i 1 = -PO, ,132 = 411, error is characterized by
f i 3 = - p l , 5-0” = cf = a:, and 32 2 - u3
-2 -u1.
Using the ML detection for optimal receiver, the de- 1
cision variable is characterized by (6) for m = 0,1,2,3. Pssk-bpsk = s ( P s s k - b p s k / O + Pssk-bpsk12)
Due to symmetry, the probability of symbol error condi-
tioned on m = 0 is equal to that conditioned on m = 1,
i.e. P s s k - b p s k l O = P s s k - b p s k l l . Similarly, P s s k - b p s k l 2 =
P s s k - b p s k l 3 . Using the union bound, we have
+;QI(? 6 0
,“)uo so &7
+tQi(;,--)

(32)
In the above upper-bound,
IV. NUMERICAL RESULTS
(22) To examine the performance of SSK and SSK-BPSK
modulation schemas, numerical evaluation of the error
probability is executed. For the SSK-BPSK scheme, the
(23) upper bound is employed to evaluate the probability of
symbol errors. For these numerical evaluations, two cases
and are considered with different means of fading factors. In
the first case, symmetric statistics of the two fading factors
(24) of antennas are considered, where the parameters

bl = b2=l+j
R1 = R2=3 (33)
are employed. In the second case, the following non-
symmetrical fading statistics are used
(25)
bl = l+j
and b2 = 1.5+1.5j
R1 = a 2 = 3 (34)
Conditioned on the event that m = 2 has been transmit- ln ~i~~~ 1 below, we presentthe bit rate @ER)
ted, P s s k - b p s k p is upper-bounded by versus the signal-to-noise ratio (SNR) for the SSK scheme.
Pssk-bpsk12 = p2(D2
3
’ O2 > D1’
D2 > 03)
Figure 2 gives the BER versus the SNR for the SSK-BPSK
scheme, where the performance degrades with a higher data
rate.
IC p 2 ( ~ 2 >om) (27) From the above numerical results, it is disclosed that
m=O
m#2 non-symmetric antennas will yield better performance.
1670
 V. CONCLUSIONS
In the paper, a novel scheme of space modulation is pro-
posed and corresponding receiver performance is analyzed; 10’’
The receiver performance will be enhanced if the fading
statistics of the two antennas are distinct, which increases it- Symmekric case
* Non-symmetric case
the distance between the signals in the signal space. The
developed SSK scheme can also be used with other tradi-
tional digital modulation schemes, such as QPSK.
More than two antennas can be used to form trans-
mit diversity, and to enhance the reception performance.
Furthermore, with more transmit antennas, the proposed
scheme of space modulation can also be extended to the
case of higher modulation level as in the M-ary signaling.
REFERENCES 10’‘
[l] A.F. Naguib, A. Paulraj, and T. Kailath, ‘‘Capacity improvement
with base-station antenna arrays in cellular CDMA,” IEEE Vehic.
Technol., vol. 43, pp. 691-698, Aug. 1994.
[2] L.C. Godara, “Application of antenna arrays to mobile commu- J
nications, Part I & Part 11,” IEEE Proc., vol. 85, pp. 1029-1245, in
&”

Aug. 1997. 1 2 3 4 5 6 7 8 9 10
[3] J.H. Winters, “Optimum combining in digital mobile radio with SNR
cochannel interference,” ZEEE Trans. Veh. Technol., vol. 33, pp. Figwe 2 The BER versus SNR for the SSKBPSK scheme
144155, 1984.
[4] S. J. Grant and J.K. Cavers, “Performance enhancement through
joint detection of cochannel signals using diversity array,” ZEEE
Trans. Commun., vol. COM-46, Aug. 1998.
[5] R.L. Peterson, R.E. Ziemer, D.E. Borth, Introduction to Spread
Spectrum Communications, Prentice Hall, 1995.

1 2 3 4 5 6 7 8 9 10
SNR
Figure 1 The BWversus SNRfor h e SSK schane

1671