## Output MAI Distributions of Linear MMSE Multiuser Receivers in DS-CDMA Systems (2001)

### Cached

### Download Links

- [degas.eecs.berkeley.edu]
- [www.eecs.berkeley.edu]
- [degas.eecs.berkeley.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | IEEE TRANS. INFORM. THEORY |

Citations: | 49 - 8 self |

### BibTeX

@ARTICLE{Zhang01outputmai,

author = {Junshan Zhang and Edwin K. P. Chong and David N. C. Tse},

title = {Output MAI Distributions of Linear MMSE Multiuser Receivers in DS-CDMA Systems},

journal = {IEEE TRANS. INFORM. THEORY},

year = {2001},

volume = {47},

number = {3},

pages = {1128--1144}

}

### OpenURL

### Abstract

Multiple-access interference (MAI) in a code-division multiple-access (CDMA) system plays an important role in performance analysis and characterization of fundamental system limits. In this paper, we study the behavior of the output MAI of the minimum mean-square error (MMSE) receiver employed in the uplink of a direct-sequence (DS)-CDMA system. We focus on imperfect power-controlled systems with random spreading, and establish that in a synchronous system 1) the output MAI of the MMSE receiver is asymptotically Gaussian, and 2) for almost every realization of the signatures and received powers, the conditional distribution of the output MAI converges weakly to the same Gaussian distribution as in the unconditional case. We also extend our study to asynchronous systems and establish the Gaussian nature of the output interference. These results indicate that in a large system the output interference is approximately Gaussian, and the performance of the MMSE receiver is robust to the randomness of the signatures and received powers. The Gaussianity justifies the use of single-user Gaussian codes for CDMA systems with linear MMSE receivers, and implies that from the viewpoints of detection and channel capacity, signal-to-interference ratio (SIR) is the key parameter that governs the performance of the MMSE receiver in a CDMA system.

### Citations

8659 | Elements of information theory - Cover, Thomas - 1991 |

4715 | Topics in Matrix Analysis - Horn, Johnson - 1991 |

3048 | Convergence of Probability Measures - Billingsley - 1968 |

1028 | Multiuser Detection
- Verdú
- 1998
(Show Context)
Citation Context ...E) receiver. Recently, there has been a substantial literature devoted to the study of linear multiuser receivers because they are practically appealing (see, e.g., [14]–[16], [18], [22], [26], [27]=-=, [29]-=-, [31]). We focus on the case where the MMSE receiver is employed because the MMSE receiver is optimal in the class of linear multiuser receivers in the sense of minimizing the mean-square error [15].... |

629 | Stochastic Processes - Ross - 1983 |

527 |
Principles of Spread Spectrum Communications
- Viterbi
- 1995
(Show Context)
Citation Context ...common approach for power control in practice is to drive the received powers for all the users having the same QoS requirements to be a fixed predetermined value, namely, power balancing (see, e.g., =-=[32]-=-). Correspondingly, the mean received powers are about the same for all the users having the same QoS requirements. Then it can be shown that the condition (3.A1) is satisfied. We have the following r... |

419 | Statistical Inference - Casella, Berger - 2001 |

402 | Varaiya P, “Capacity of fading channels with channel side information
- Goldsmith
- 1997
(Show Context)
Citation Context ...eceived powers. (This is a reasonable model because fading channel gains are typically assumed to be stationary and ergodic, and all the users in one class can be assumed to have i.i.d. channel gains =-=[6]-=-, [7].) For convenience, let denote the collection of the signatures and received powers when the processing gain is , and the sequence . Let denote the conditional distribution of given . Following [... |

324 | Real Analysis - Royden - 1988 |

310 | Linear Multiuser Detectors for Synchronous Code Division Multiple Access Channels
- Lupas, Verdu
- 1989
(Show Context)
Citation Context ... the minimum mean-square error (MMSE) receiver. Recently, there has been a substantial literature devoted to the study of linear multiuser receivers because they are practically appealing (see, e.g., =-=[14]��-=-�[16], [18], [22], [26], [27], [29], [31]). We focus on the case where the MMSE receiver is employed because the MMSE receiver is optimal in the class of linear multiuser receivers in the sense of min... |

267 | Linear multiuser receivers: Effective interference, effective bandwidth and user capacity
- Tse, Hanly
- 1999
(Show Context)
Citation Context ...e error (MMSE) receiver. Recently, there has been a substantial literature devoted to the study of linear multiuser receivers because they are practically appealing (see, e.g., [14]–[16], [18], [22]=-=, [26]-=-, [27], [29], [31]). We focus on the case where the MMSE receiver is employed because the MMSE receiver is optimal in the class of linear multiuser receivers in the sense of minimizing the mean-square... |

258 | Adaptive Multiuser Detection
- Honig, Madhow, et al.
- 1995
(Show Context)
Citation Context ... in fact not crucial for the construction of the MMSE receiver. We will elaborate further on this in Section III. Worth noting is that the MMSE receiver does require knowledge of the timing of user 1 =-=[9]-=-. Since the MMSE receiver has no knowledge of the instantaneous powers of the interferers, the power of the overall interference is a function of and hence random, and can be shown to be as follows: A... |

220 | Spectral Efficiency of CDMA with Random Spreading
- Verdu, Shamai
- 1999
(Show Context)
Citation Context ... It is well known that CDMA systems have the desirable features of dynamic channel sharing, wide range of operating environments, graceful degradation, and ease of cellular planning (see, e.g., [11], =-=[31]-=-). CDMA systems also offer the potential to support a heterogeneous mix of transmitting sources with a broad range of bursty traffic characteristics and quality-of-service (QoS) requirements. In this ... |

157 | On the epmirical distribution eigenvalues of a class of large dimensional random matrices
- Silverstein, Bai
- 1995
(Show Context)
Citation Context ...47, NO. 3, MARCH 2001 (nonrandom) distribution function with probability one, and the Stieltjes transform of is the solution to the following functional equation: for all It is further pointed out in =-=[24]-=- that the limiting distribution function is unique. In this section, we prove Theorem 3.1 via the dependent central limit theorem in [17], which is a central limit theorem for martingale difference ar... |

144 | Probability of error in mmse multiuser detection - Poor, Verdú - 1997 |

143 |
Proakis, Digital Communications
- G
- 1995
(Show Context)
Citation Context ...reasoning in [23], we assume that the signatures are known to the receivers once they are picked. We consider user 1 without loss of generality. A discrete-time composite waveform channel (see, e.g., =-=[17]-=-) seen by user 1 is shown in Figure 2, where all the other users' signals contribute to the MAI. By definition, every linear receiver generates an output of the form y (N) 1 j c t Y (N) (see, e.g., [1... |

123 | An Introduction to Optimization - Chong, Zak - 1996 |

92 | Stochastic power control for cellular radio systems
- Ulukus, Yates
- 1996
(Show Context)
Citation Context ...e present numerical examples and draw our conclusions in Section 6. 2 System Model Figure 1 depicts a simplified block diagram of the uplink of a CDMA system equipped with the MMSE receiver. Based on =-=[23, 25, 28]-=-, we have the following discrete-time model for the uplink of a synchronous CDMA system. For any symbol interval, the baseband received signal at the front end of the receiver is Y (N) = K X i=1 p P i... |

79 | Optimal sequences, power control, and user capacity of synchronous CDMA systems with lin169 MMSE multiuser receivers - Viswanath, Anantharam, et al. - 1999 |

60 |
Dependent central limit theorems and invariance principles
- McLeish
- 1974
(Show Context)
Citation Context ...wing functional equation: for all It is further pointed out in [24] that the limiting distribution function is unique. In this section, we prove Theorem 3.1 via the dependent central limit theorem in =-=[17]-=-, which is a central limit theorem for martingale difference arrays (cf. [13]). We begin with some technical lemmas. A. Technical Lemmas Define We have the following lemmas, the proofs of which have b... |

55 | Power control and capacity of spread-spectrum wireless networks
- Hanly, Tse
- 1999
(Show Context)
Citation Context ...en the signatures and powers, which is particularly useful in the scenario where repetition of the same random signatures is adopted. (Repetition of the same random signatures has been pointed out in =-=[7]-=- to be more suitable for the implementation of MMSE receivers that try to exploit adaptively the structure of the interference.) Suppose there are K users in the system when the processing gain is N .... |

54 | Linear multiuser receivers in random environments - Tse, Zeitouni - 2000 |

38 |
User-separating demodulation for code-division multiple-access systems
- Rupf, Tarkoy, et al.
- 1994
(Show Context)
Citation Context ...-square error (MMSE) receiver. Recently, there has been a substantial literature devoted to the study of linear multiuser receivers because they are practically appealing (see, e.g., [14]–[16], [18]=-=, [22]-=-, [26], [27], [29], [31]). We focus on the case where the MMSE receiver is employed because the MMSE receiver is optimal in the class of linear multiuser receivers in the sense of minimizing the mean-... |

37 |
Throughput analysis for code division multiple accessing of the spread spectrum channel
- Hui
- 1984
(Show Context)
Citation Context ...stems. It is well known that CDMA systems have the desirable features of dynamic channel sharing, wide range of operating environments, graceful degradation, and ease of cellular planning (see, e.g., =-=[11]-=-, [31]). CDMA systems also offer the potential to support a heterogeneous mix of transmitting sources with a broad range of bursty traffic characteristics and quality-of-service (QoS) requirements. In... |

22 | Effective interference and effective bandwidth of linear multiuser receivers in asynchronous systems
- Kiran, Tse
- 2000
(Show Context)
Citation Context ...us. More specifically, the offset (also called delay) of the interferer relative to user 1, denoted as , is a nonnegative integer in terms of the number of chips, . Following the line of reasoning in =-=[12]-=-, we assume that the offsets are known to the receiver although they are random. As pointed out in [12], [15], [30], a typical interferer has two different but consecutive symbols interfering with the... |

19 | The multi-access fading channel: Shannon and delay limited capacities,” in 33th Annu
- Hanly, Tse
- 1995
(Show Context)
Citation Context ...ed powers. (This is a reasonable model because fading channel gains are typically assumed to be stationary and ergodic, and all the users in one class can be assumed to have i.i.d. channel gains [6], =-=[7]-=-.) For convenience, let denote the collection of the signatures and received powers when the processing gain is , and the sequence . Let denote the conditional distribution of given . Following [25], ... |

14 |
CDMA systems in fading channels: admissibility, network capacity, and power control
- Zhang, Chong
- 2000
(Show Context)
Citation Context ...e ’s. Our results are asymptotic in nature, with both and going to infinity. Throughout this paper, the ratio of to is denoted by and taken to be fixed as ,asis standard (see, e.g., [13], [26], [31]=-=, [34]-=-). We consider user 1 without loss of generality. Roughly speaking, our main results for synchronous systems can be summarized as follows. Unconditional Distribution: Assuming that the empirical distr... |

14 | On the average near-far resistance for MMSE detection of direct sequence CDMA signals with random spreading - Madhow, Honig - 1999 |

9 |
Two different philosophies in CDMA—A comparison
- Vembu, Viterbi
- 1996
(Show Context)
Citation Context ...N (s i1 ; : : : ; s iN ) T , where the s in 's are i.i.d. with Pfs in = 1g = Pfs in = \Gamma1g = 1 2 , n = 1; : : : ; N and i = 1; : : : ; K. This model is applicable to several scenarios (see, e.g., =-=[7, 14, 23, 24, 26, 29]-=-). First, consider the situation where users employ long pseudo-random spreading sequences with periods considerably larger than the number of chips per symbol interval (such as in IS-95 systems), whi... |

7 |
An asymptotic analysis of DS/SSMA communication systems with random polyphase signature sequences
- Lok, Lehnert
- 1996
(Show Context)
Citation Context ... of the ’s, not the ’s. Our results are asymptotic in nature, with both and going to infinity. Throughout this paper, the ratio of to is denoted by and taken to be fixed as ,asis standard (see, e.=-=g., [13]-=-, [26], [31], [34]). We consider user 1 without loss of generality. Roughly speaking, our main results for synchronous systems can be summarized as follows. Unconditional Distribution: Assuming that t... |

7 |
Shitz), "Spectral efficiency of CDMA with random spreading
- Verdu, Shamai
- 1999
(Show Context)
Citation Context ...stems. It is well-known that CDMA systems have the desirable features of dynamic channel sharing, wide range of operating environments, graceful degradation, and ease of cellular planning (see, e.g., =-=[10, 29]-=-). CDMA systems also offer the potential to support a heterogeneous mix of transmitting sources with a broad range of bursty traffic characteristics and quality-of-service (QoS) requirements. In this ... |

6 |
MMSE interference suppression for directed-sequence spread-spectrum CDMA
- Madhow, Honig
- 1994
(Show Context)
Citation Context ... [29], [31]). We focus on the case where the MMSE receiver is employed because the MMSE receiver is optimal in the class of linear multiuser receivers in the sense of minimizing the mean-square error =-=[15].-=- Assuming the signatures are deterministic, Poor and Verdú [18] have established the Gaussian nature of the MAI-plus-noise at the output of the MMSE receiver in several asymptotic scenarios (the outp... |

6 | the average near-far resistance for MMSE detection of directed-sequence CDMA signals with random spreading - “On - 1999 |

6 |
Shitz), "Multiuser detection with random spreading and errorcorrection codes: fundamental limits", presented at Allerton Conference
- Verd'u, Shamai
- 1997
(Show Context)
Citation Context ...le. In this paper, we study the behavior of MAI at the output of the minimum-meansquare -error (MMSE) receiver. The MMSE receiver is one type of linear multiuser receivers that have been advocated in =-=[28, 29]-=- and the references therein to be adopted as front ends in a CDMA system because they are practically appealing. There recently has been a substantial literature devoted to the study of linear multius... |

5 |
On Conditional Weak Convergence
- Sweeting
- 1989
(Show Context)
Citation Context ...], [7].) For convenience, let denote the collection of the signatures and received powers when the processing gain is , and the sequence . Let denote the conditional distribution of given . Following =-=[25]-=-, we say that converges almost surely to a probability measure if for any bounded continuous function , with probability one. We note that convergence almost surely of conditional distribution is an i... |

4 |
Two different philosophies
- Vembu, Viterbi
- 1996
(Show Context)
Citation Context ...odel for binary random signatures is as follows: , where the ’s are independent and identically distributed (i.i.d.) with and This model is applicable to several scenarios (see, e.g., [8], [16], [26=-=]–[28]-=-, [31]). First, consider systems where users employ (1)s1130 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 47, NO. 3, MARCH 2001 Fig. 1. A simplified block diagram of the uplink of a CDMA system with ... |

3 | Design and performance analysis of power-controlled CDMA wireless networks with linear receivers and antenna arrays - Zhang - 2000 |

2 |
control and capacity of spread-spectrum wireless networks
- “Power
- 1987
(Show Context)
Citation Context ...ry-valued. The model for binary random signatures is as follows: , where the ’s are independent and identically distributed (i.i.d.) with and This model is applicable to several scenarios (see, e.g.=-=, [8], -=-[16], [26]–[28], [31]). First, consider systems where users employ (1)s1130 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 47, NO. 3, MARCH 2001 Fig. 1. A simplified block diagram of the uplink of a ... |