Analysis of Successive Interference Cancellation Scheme with M-ary QAM in DS/CDMA Systems

Ann-Louise Johansson and Arne Svensson

Chalmers University of Technology Department of Information Theory S-412 96 Gothenburg, Sweden e-mail anne.johansson@it.chalmers.se/arne.svensson@it.chalmers.se

Abstract
In this paper we analyse a successive IC scheme for higher modulation in DS/CDMA systems. The performance under Rayleigh fading is analysed for single modulation systems and systems employing either of two different methods, mixed modulation or parallel channels, for handling multiple data rates. We show that the successive IC scheme for higher modulation together with mixed modulation or parallel channels give a considerable increase in performance and exibility compared to single modulation systems employing a conventional detector.

is known as the near/far problem. One way to combat this problem is to use stringent power control [3]. Another approach would be to use more sophisticated receivers which are near/far resistant. Recently a lot of attention has been directed to the area of multi-user detectors, which has the prospect of mitigating the near/far problem and by cancelling the MAI also increasing the total system capacity. The main problem with multi-user detectors is the complexity. Several sub-optimal detectors have been analysed, but the problem remain, especially for schemes using parallel detection [4]-[6]. An alternative to parallel detection, and therefore also parallel interference cancellation, is serial or successive Interference Cancellation (IC)[7]. See also reference list in [1]. In the case of a multi-rate DS/CDMA systems, a method to handle multiple data rates would be to let different users use different forms of modulation. A user could, depending on the need, choose between using e.g. BPSK, QPSK or 16QAM modulation. Therefore, an IC scheme for M-ary QAM is analysed in this paper, based on the IC scheme for coherent BPSK modulation derived by Patel and Holtzman [8][10]. Another approach to handle multiple data rates is to let each user transmit over one or several parallel channels according to requirements. This can of course also be used in combination with higher modulation. In this paper we consider the coherent case of demodulation. We assume perfect knowledge of the phase and the time delay for each signal, and in the performance analysis we assume that we have a perfect estimate of the channel gain.

1 Introduction
In the future we will demand mobile telephone systems to be able to handle services with other requirements than speech, like e.g. facsimile, Hi-Fi audio and computer data, which is not possible today. To achieve this we need to use a multiple-access method which is exible and has the prospect of capacity increases and being able to handle variable data rates. Recently Code Division Multiple Access (CDMA) has been suggested to be a multiple-access method able to stand these requirements. The motivation for this work is then to evaluate an efcient detector for a multi-user and multi-rate Direct-Sequence Code Division Multiple Access (DS/CDMA) system. This paper will give an outline of the work, a deeper analysis will be found in [1]. A DS/CDMA system has several unique features. Some of them are spectrum sharing, rejection of multipath signal components or possibility to utilizing them for recombining [2] and frequency reuse factor of one in a cellular scenario [3]. These features are highly desirable, though a CDMA system employing a conventional matched lter detector is interference limited, which directly determines the system capacity. The presence of a number of users in the system introduces Multiple-Access Interference (MAI), since the signature sequences used are not perfectly orthogonal, which leads to an irreducible error probability. In a mobile radio scenario the transmitters move in relation to the receiver and the energies of the received signals are to be neither equal nor constant. In this situation the conventional detector fails to demodulate weak users, even when the cross-correlation between the signals is relatively low. This
Ann-Louise Johansson 1

2 Successive Interference Cancellation Scheme with M-ary QAM

For operating the IC scheme we need the knowledge of all the users signature sequences. The received signal is correlated with each different signature sequence and then the strongest user is selected using the outputs of the linear correlators. Information about the instantaneous powers is therefore not needed but, hence for a IC scheme with M-ary QAM, the knowledge about the decision boundaries, and therefore the average power, is required for correct decoding. Consequently, this scheme would only work when transmitting over a slowly fading channel. Where slowly

Arne Svensson

Analysis of Successive Interference Cancellation Scheme with M-ary QAM in DS/CDMA Systems

fading implies that the local mean of the mobile radio signal is sensibly constant over a small area or a short period of time. The fading also has to be sufciently slow so estimation of the phase is possible and coherent detection is practicable [2].

I that user 1 has the strongest signal. Accordingly, A 1, 0 and Q are decoded rst at time + T and cancelled at baseA 1, 0 1 band from the composite signals. Subsequently all the users are decoded and cancelled in non-decreasing order of their powers. At this moment we do not bother about how the
Q c 1 ( t 1 ) sin 1 I c 1 ( t 1 ) cos 1 I S 1, 0 Detector 1 Q S 1, 0 Q I S 1, 0 S 1, 0

2.1 System Model and Decoder Structure

We consider a system model for a system utilizing square lattice QAM, where the received signal, for a K user system, is the sum of all the transmitted signals embedded in AWGN with two-sided power spectral density of N 0 2 . The information-bearing signal for all the users is an innite sequence I I of A k,/ Q amplitude, rectangular pulses of duration T. A k, l l Q and A k, l are the amplitudes of the quadrature carriers for the k th users l th symbol element, which together generate M equiprobable and independent symbols. Each user k has I two signature sequence, c k / Q ( t ) , which are used for spreading the signal in the in-phase and the quadrature branch. It consists of a sequence of, antipodal, unit amplitude, rectangular pulses of duration T c . The period of all the users signature sequences are N = T T c , so there is one period per data symbol. In the asynchronous, though symbol-synchronous, case the time delay, k , and the phase, k , are i.i.d. uniform random variables over [ 0, T ) and [ 0, 2 ) , respectively. Both parameters are assumed to be known. Additionally the amplitude of the signal is affected by the channel gain, k .
c iI ( t i ) cos i Z iIIl + 1 , -- dt T c iQ ( t i ) sin i IQ 1 -- dt Z i, l T c iI ( t i ) sin i 1 -- dt Z QI i, l T cos c t

+
I ( t) +

LPF
r ( t)

Detector 2 Q ( t)

LPF
sin c t

I S K, 0 Detector K Q I S 1, 0 S 1, 0 Q S K, 0 I c 1 ( t 1 ) ( sin 1 ) Q c 1 ( t 1 ) cos 1

Select Max & Decode

Figure 2. M-ary QAM receiver with IC ranking is done. We use the previous dened decision variaI Q bles S 1, 0 and S 1, 0 to regenerate user 1s baseband signal for symbol element 0 and cancel it from the composite sigI Q nal. The resulting baseband signals, 1 ( t ) and 1 ( t ) , after the rst cancellation will consist of all K users baseband signals except user 1s symbol element 0, though instead some additional noise caused by imperfect cancellation will be included. In a situation like the one shown in Figure 3, the strongest user, user 1, has a time delay 1 which is shorter than any other users time delay as e.g. user 2s . In 0 0
1

cos c t I ( t)

S iI, l

Decision Device

LPF
r ( t) Q ( t)

( )2 ( )2
Decision Device

+
S iQl ,

( S i, l ) 2

1 T
1 + T 2 2 + T

2 2T
1 + 2T 2 + 2T

3T

LPF
sin c t

c iQ

( t i ) cos i + Z iQQ 1 ,l -- dt T +

User 1 User 2

Figure 1. M-ary QAM DS/CDMA receiver Figure 1 shows the structure of the i th users receiver. The I-branch as well as the Q-branch is correlated with both the I and Q signature sequences of the i th user to form four different Z i, l factors at each instant of T. These factors, which contain all information, are then used to form the decision variables and subsequently select the user with the strongest signal. Combining these correlator outputs we obtain the decision variables S iI, l and S iQl , see Figure 1, for the I and , Q components of the i th user. They consist of one constant, the symbol amplitude, and a noise component. User K

K + T

K + 2T

Figure 3. Asynchronous system this case it would not be enough to cancel only symbol element 0 from the composite signal to reduce the interference noise since user 2s zeroth symbol element will be partly correlated with user 1s symbol element 1. To overcome this we decode and cancel both user 1s zeroth and rst symbol element before we continue the IC scheme for the second strongest user. Without loss of generality we assume that the strongest user has the longest time delay and the second strongest user has the second longest time delay and so on. Proceeding in the same manner, cancelling symbol element
2 Arne Svensson

2.2 Interference Cancellation Scheme

A receiver with IC is shown in Figure 2. Without loss of generality we consider the decision of symbol element 0 and

Ann-Louise Johansson

Analysis of Successive Interference Cancellation Scheme with M-ary QAM in DS/CDMA Systems

0 for the second strongest user, e.g. user 2, and all the other users in non-decreasing order. After h cancellations there are h cancelled and K-(h+1) remaining symbol elements 0 in I the composite baseband signals, h/ Q ( t ) . In the next step I Q of the IC scheme, the total noise components, N h/+ 1 , of the th user consists of three decision variables for the ( h + 1 ) parts: noise caused by the remaining interfering users, noise caused by imperfect cancellations and AWGN.

10

Processing Gain = 127 Random Sequences 10 Pb 10

2 1

Conv

2.3 Performance Analysis

We use the Gaussian approximation for calculating the I Q probability of error. The noise, N h/+ 1 , is a sum of independent random variables with zero mean and we then assume that the noise is Gaussian. With this assumption the probability of error is easily obtained from the theory of single transmission of QAM signals over a AWGN channel [1]. In the case of fading, we consider a system where the K users are received through independent, frequency-nonselective slowly fading channels. For a wideband signal the model apply in sub-urban areas where the local mean of the mobile radio signal can be said to be constant over a small area, with slow variations when the receiver moves. When the symbol period, T, is less than the coherence time, the multiplicative process, caused by fading, may be regarded as constant during at least a few signal intervals. This can be considered the condition for the term slowly fading [2]. The channel has to fade sufciently slow for making it possible to estimate the phase shift, k , from the received signal of the k th user adequately. This is a requirement for coherent detection [2]. Operating an IC scheme for M-ary QAM also demands that the fading is slow, since the channel gain, k , has to be estimated for correct decoding of the symbols and the ranking of the users. In this paper we will use the same method, for analysing the single-path Rayleigh fading channel, as was used in [10] for BPSK modulation. We assume that k is estimated perfectly and we obtain the noise variance and error probability conditioned on k . The amplitudes decide the order of cancellation, thus the order will change continuously with fading. One way to decide which user is the strongest is to dene a frame of n bits for each user [10]. After we have received the whole frame we can decide which user is the strongest by comparing the average received power for all the users. The distribution of the ordered amplitudes is obtained by using order statistics [10]. Then the unconditioned error probability for each stage of cancellation is easily obtained from the conditioned probability of error. Transmission under Rayleigh fading for systems with IC employing different modulation forms, though the same total transmitted bit rate, and a single BPSK user is shown in Figure 4.

BPSK (40) QPSK (20) 16QAM (10) Single BPSK Single 16QAM 5 10 15 Eb /N0 (dB) 20

IC

10

25

Figure 4. Performance with IC under fading choose a modulation form in correspondence with required transmitted data rate [11]. We have analysed the performance of a system where the users employ different forms of modulation, e.g., a combination of BPSK, QPSK and 16QAM users. We consider a system where we have K 1 BPSK, K 2 QPSK and K 3 16-QAM users. To make a comparison between different forms of modulation we let the average SNR/bit be equal for all users. That is, the transmitted bit energy is equal and independent of the modulation form used. When calculating the total BER for the whole system we weight together each users BER using the rates as weights. A mixed system with 20 BPSK, 10 QPSK and 5
10
0

Processing Gain = 127 Random Sequences 10 Pb 10

2 1

Mix Conv

Av. BPSK Av. QPSK Av. 16QAM Mix (20/10/5) Single BPSK 5 10 15 Eb /N0 (dB)

IC

10

20

25

Figure 5. Mixed system with IC under fading 16-QAM users is shown in Figure 5 together with a mixed system employing a conventional detector. The channel gain is calculated using order statistics and are then distributed randomly among the users in the mixture independent of modulation form. The result is an average over 100 mixtures. The average performance for a BPSK, QPSK respectively 16-QAM user in the mixture is also shown in Figure 5.

3 Performance Analysis of Mixed Modulation Systems

One way to handle multi-rate systems is to let each user
Ann-Louise Johansson 3

Arne Svensson

Analysis of Successive Interference Cancellation Scheme with M-ary QAM in DS/CDMA Systems

4 Performance Analysis of Parallel Channel Systems

We have previously discussed the possibility to handle multi-rate systems by the means of employing higher modulation forms. Another alternative is to use parallel channels for transmission of information from each user [11]. We simply let a user send simultaneously over as many channels as required for a specic data rate. In this kind of system there are going to be a larger number of interfering signals, though the synchronous signals will have considerably lower cross-correlation than asynchronous signals. In Figure
10
0

sequences and in some cases it would be better to add a 16QAM user, if the user can accept a small decrease in performance, instead of a QPSK user with two parallel channels.

6 Acknowledgement
This project was sponsored by the Swedish National Board of Industrial Technical Development, NUTEK. Project 9303363-2.

7 References
Processing Gain = 127 Gold Sequences Conv

1.

10

A. Johansson and A. Svensson, Analysis of successive interference cancellation scheme with M-ary QAM in a DS/CDMA system, Internal report. J. G. Proakis, Digital Communications, 2 ed, McGraw-Hill, 1989. K. S. Gilhousen, et al, On the capacity of a cellular CDMA system, IEEE Trans. on Vehicular Technology, vol. 40, pp. 303-311, May. 1991. R. Lupas and S. Verdu, Near-far resistance of multiuser detectors in asynchronous channels, IEEE Trans. on Communications, vol. 38, pp. 497-507,April 1990. M. Varanasi and B. Aazhang, Multistage detection in asynchronous code-division multiple access communications, IEEE Transactions on Communications, vol. COM-38, pp. 509-519, April 1990. A. Duel-Hallen, Decorrelating decision-feedback multiuser detector for synchronous code-division multiple access channel, IEEE Trans. on Communications, vol. 41, pp. 285-290. Feb. 1993. Magnus Ewerbring, et al, CDMA with Interferece Cancellation: A Tequnique for High Capacity Wireless Systems, in proceedings ICC93 (Geneva, Switserland), May 1993. P. Patel and J. Holtzman, Analysis of a DS/CDMA successive interference cancellation scheme using correlations, in proceedings, Globecom (Houston, Texas), Dec. 1993. P. Patel and J. Holtzman, Analysis of successive interference cancellation in M-ary orthogonal DSCDMA system with single path rayleigh fading, in proceedings, International Zurich Seminar, (Zurich, Switzerland), March 1994. P. Patel and J. Holtzman, Analysis of a simple successive interference cancellation scheme in a DS/ CDMA, IEEE Journal on Selected Areas in Communications, vol. 12, pp. 796-807, June 1994. T. Ottosson and A. Svensson, Performance of different multi-rate schemes in DS/CDMA systems, in proceedings, Nordiskt radioseminarium (Linkping, Sweden), Oct. 1994.

2.
Pb 10
2

IC

3.
10
3

QPSK (15), P=2 QPSK (30) 16QAM (15) Single BPSK

4.
20 25

10

10 15 Eb /N0 (dB)

5. Figure 6. Comparison between parallel QPSK and asynchronous 16-QAM 6 a comparison between a system with only QPSK users employing parallel channels and a pure asynchronous system with 16-QAM users. 6.

5 Summary and Conclusions

We have demonstrated the use of higher modulation together with successive IC in single modulation systems and compared the average BER between the systems. We have analysed two different methods, mixed modulation and parallel channels, for handling multiple data rates together with IC. The conclusion is that the successive IC scheme has relatively low complexity even in the case of higher modulation and yields considerable increase in performance [1], [10]. Mixed modulation systems offers more exibility with the cost of a slight decrease in average performance compared to pure asynchronous QPSK systems. The 16-QAM users in the mixed system, who are most sensitive to noise, will have the highest BER. Therefore, if we instead have a QPSK system where the users employ parallel channels the average performance would be almost equal for all the users. (It would be equal if all the users in the system use the same number of parallel channels.) The drawback with parallel channels is that sooner or later we run out of signature

7.

8.

9.

10.

11.

Ann-Louise Johansson

Arne Svensson