行政院國家科學委員會專題研究計畫成果報告
W-CDMA 基地台接收系統之初始擷取與多用戶偵測子系統之
研究與實作
Study and Implementation of the Acquisition and Multiuser
Detection Subsystem for W-CDMA systems
計畫編號：NSC 90-2219-E-009-011
執行期限： 90 年 8 月 1 日至 91 年 7 月 31 日
主持人：吳文榕教授 國立交通大學電信系
Email:[email protected]
I. 中文摘要 近年來，在第三代行動通訊系統中， 寬頻多碼分工系統(W-CDMA)已經被廣 泛使用在增加使用人數容量與傳送速度 上。對於多碼分工系統來說，使用多使用 者偵測(multiuser detection)技術可以增加 使用人數容量，多使用者偵測有很多做 法，在考量效能及實際實現之複雜度後， 在本計劃中，我們提出一個寬頻分碼多工 系統之全數位最小均方差(MMSE)多使用 者偵測法。利用MMSE的運算法則能找到 一個最佳之二級部分平行干擾消除(PIC) 的部分消除因子(PCF)。在模擬的結果中 顯示出，偵測的效能可以經由所導出的最 佳PCF而有改善。最後，根據第三代行動 通訊的規格，我們以FPGA的設計流程實 現所提出來的接收機架構。 關鍵詞： Abstr actMultiuser detection (MUD) is one of the techniques increasing the capacity of a DS-CDMA system. Many MUD algorithms have been proposed. Taking performance and implementation complexity into account, we propose a two-stage partial parallel interference cancellation (PIC) scheme in this project. We use the MMSE criterion to find the optimal partial cancellation factors (PCFs) for the two-stage partial PIC receiver. It is shown that the detection performance can be greatly enhanced using the derived optimal PCFs. The other advantage of the partial PIC approach is that the computational complexity is low. Then
we propose an efficient hardware
architecture and implement it using the FPGA design flow.
II. 計畫緣由與目的
Directly-sequence code-division
multiple access (DS-CDMA) [1]-[2] has
been considered as the standard
transmission technique for the next
generation mobile radio systems. In
DS-CDMA systems, the code orthogonality property cannot always hold at the receiver due to the channel impairment. As a result, the multi-access interference (MAI) arises and this limits the capacity of the CDMA system. Thus, much research has been directed to solve this problem. Multiuser detector has been considered as the most promising technique. The optimal multiuser
detector, which uses the maximum
likelihood criterion, is difficult to
implement due to its high computational complexity. Among the many suboptimal solutions, the PIC is considered as a good
candidate for practical applications.
However, the interference cancellation is not always reliable for the PIC. Thus the partial PIC scheme was proposed [20]-[22]. In the partial PIC approach, partial cancellation factors (PCFs) are introduced to control the interference cancellation level. However, the optimal PCE, which greatly influence the detection performance, cannot easily found. This project is aimed to solve the problem. We design a two-stage partial PIC receiver using the minimum mean square error (MMSE) criterion. The optimal PCFs can then theoretically derived. Using the optimal PCFs, we can significantly
improve the detection performance. We
also propose an efficient hardware
architecture and this leads to an all-digital multiuser W-CDMA receiver. Finally, we use the FPGA design flow to implement the proposed receiver.
III. 研究方法與成果
Figure 1 is the structure of a two-stage
partial PIC. Consider a K-user synchronous
DS/CDMA system the received signal is expressed as
( )
(
(
)
(
)
)
(
)
(
)
(
)
(
)
, , 1 1, 2, . . K k I k k Q k k k k k k k k r t P d t j d t a t j a t n t τ τ τ τ τ = = − + ∗ − − + ∗ − + −∑
At the first stage the ith bit complex-valued
decision metric of the kth user is given by
( 1 ) 1 ( ) ( ) ( ) 1 ( ) 2 2 k k k K k m m m k m m k Z i P b i i P b i R N N η = ≠ = +
∑
+where bk(i) is the ith bit of user k, Pk is the
signal power, Rmk is the correlation function
between user m and user k, and ηk( )i is
the noise term. At the second stage, the
regenerated signal of user k is produced by
µ µ 1 ( ) ( ) ( ), K k k j j j k r t r t C s t = ≠ = −
∑
whereCk is the PCF, and µ ( )s tj denotes the
estimate of the jth interference and can be
expressed as µ µ (1) ( ) ( ) ( ) ( ) ( ) j j j j b i s t a t d i Z i π t iT ∞ =−∞ =
∑
−Then the second-stage complex-valued
decision metric Zk(2)( )i can be obtained.
The error signal ek(i) is defined as the
difference between the user signal and
(2)
( )
k
Z i . Assume that the I and Q channel data sequence can be modeled as a sequence of independent and identically distributed random variables with equal probability. Then we form the mean square error (MSE) term as MSE=E{e ik( )e ik( ) *}. The optimal
Ck can be determined by letting
/ k 0
MSE C
∂ ∂ = . The MMSE PCF derived
above requires knowledge of user powers, cross-correlations of signature sequences, and the noise variance. If a new user is
added or dropped, the correlation matrix will change accordingly. Besides, the long scrambling codes of W-CDMA systems may make the correlation matrix to vary symbol by symbol. One way to reduce the computational complexity is to consider the signature sequence (including real and image part, i.e. a1,k(i), a2,k(i)) as a random
vector of length. Let the components of these vectors be independent and identically distributed random variables with equal probability. The correlation products can then be substituted by their expectation counterparts. The same procedures can be
repeated to calculate e i and MSE.k( )
Differentiating MSEk with respect to
Ck and setting the result to zero, we can
obtain the MMSE PCF for long codes as
[ ] [ ] [ ] 0 1 2 0 1 1 2 ( 1) 8 ( 1)( 1) ( 2)(3 3) 2 8 K m m m k k K K k m m m m k m k N N P N K K C N P K KN P N K N K K N = ≠ = = ≠ ≠ + − + − = − − + + − + − + − + ∑ ∑ ∑
The partial PIC detector which uses the above PCF is called the random MMSE (RMMSE ) PIC detector. If all users have equal power (under perfect power control), i.e., P1= P2= … = PK, then we have a simple
result as [ ] 0 2 0 ( 1)( 2) ( 1) 8 ( 1)( 1) ( 1) ( 2)(3 3) ( 1)( 2 ) 8 k k N N K N K K C N P K KN K N K N K K K N − + − + − = − − + − + − + − + − − +
Figures 3 and 4 show the performance comparison for the conventional and proposed detector. We can see that that the proposed algorithm performs much better than the matched filter.
The second part of this project is concerned about the FPGA implementation. The received consists of several functional blocks, i.e., (1) complex-valued multiplier, (2) pulse shaping filter, (3) chip matched filter, (4) interpolation filter, (5) the noncoherent DLL, and (6) the adder. We have proposed many efficient architectures to decrease the implementation complexity. Figure 2 is the detailed structure for the proposed receiver. Figure 4 is the RTL architecture for the proposed receiver.
IV. 結論
In this project, we have developed a two-stage MMSE partial PIC detector. The
PCF is derived based on the MMSE criterion. Using simulation results, we show that the proposed algorithm performs significantly better than the conventional matched filter and the two-stage full PIC receiver. We also propose a hardware architecture for the detector and implement
it using the FPGA design flow.
V. 參考文獻
[1] R. L. Peterson, R. E. Ziemer, and D. E. Borth, “Introduction to spread spectrum communication,” Prentice-Hall, 1995. [2] R. L. Pickholtz, D. L. Schilling, and L. B.
Milstein, “ Theory of Spread-Spectrum
Communications-A Tutorial,” IEEE
Trans. Commun., vol. COM-30, no. 5, pp. 855~883, May. 1982.
[3] W. Huang, I. Andonovic, and, M.
Nakagawa, “ Code tracking of
DS-CDMA systems in the presence of multiuser interference and additive
noise,” IEEE 5th International
Symposium, vol. 3, pp. 843~847, 1998. [4] S. Verdu, “ Minimum probability of
error for asynchronous Gaussian
multiple-access channels,” IEEE Trans. Inform. Theory, vol. IT-32, pp. 85~96, Jan. 1986.
[5] Shimon Moshavi, Bellcore, “Multi-User
Detection for DS-CDMA
Communication,” IEEE
Communications Magazine, pp.
124~136, Oct. 1996.
[6] Dimitris Koulakiotis and A. Hamid Aghvami CTR, “ Data Detection Techniques for DS/CDMA Mobile Systems: A review,” IEEE Personal Communication, pp.24~34, June 2000. [7] Dariush Divasalar, Marvin K. Simon,
and Dan Raphaeli, “Improved parallel interference cancellation for CDMA,” IEEE Trans. Commun., vol. 46, no. 2, pp. 258~268, Feb. 1998.
[8] Neiyer S. Correal, R. Michael Buehree, and Brain D. Worner, “A DSP-Based
DS-CDMA multiuser receiver
employing partial parallel interference cancellation,” IEEE J. Select. Areas Commun. Vol. 17, no. 4, pp. 613~630, Apr. 1999.
[9]. Tero Ojanpera, Nokia Research Center and Ramjee Prasad, Delft University of Technology, “An Overview of Air
Interface Multiple Access for
IMT-2000/UMTS,” IEEE Commun. pp. 82~95, Sept. 1998.
[10]. P. Chaudhury, W. Mohr, and S. Once, “ The 3GPP Proposal for IMT-2000,” IEEE Commun. Mag. , vol. 37, pp. 72-81, Dec. 1999.
[11].Technical Specification 3Gpp TS 25.211 V4.1.0 (2001-06)
[12]. Harri Holma and antti toskala, Nokia, Finland, “WCDMA for UMTS”
[13]. R. L. Pickholtz, D. L. Schilling, and L.
B. Milstein, “ Theory of
Spread-Spectrum Communications-A Tuorial,” IEEE Trans. Commun., vol. COM-30, no. 5, pp.855-883, May. 1982.
[14]Shimon Moshavi, Bellcore, “Multi-User
Detection for DS-CDMA
Communication,” IEEE
Communications Magazine, pp.
124~136, Oct. 1996.
[15] R. Lupas; S. Verdu, ”Linear multiuser
detectors for synchronous
code-division multiple-access
channels,” IEEE, Trans. Info. Theory, vol. 35, Issue 1, pp. 123~136, Jan., 1989
[16] Xie, Z.; Short, R.T.; Rushforth, C.K., “A family of suboptimum detectors for coherent multiuser communications,” IEEE Communications, vol., 8 Issue: 4, May 1990, pp. 683 -690
[17] Koulakiotis, D.; Aghvami, A.H., “Data detection techniques for DS/CDMA mobile systems: a review,” IEEE Personal Commun., Vol. 7 Issue: 3, June 2000, pp. 24 –34
[18] J. M. Holtzman, “DS/CDMA
successive interference cancellation,” IEEE Third International Symposium, vol. 1, pp.69~78, 1994
[19] M. K. Varanasi and B. Aazhang,
“Multistage Detection in
Asynchronous Code-Division
Multiplt-Access Communications,”
IEEE Trans. Commun., vol. 46, no. 11, Nov. 1998, pp. 1492~1432
[20] D. Divsalar and M. Simon, “Improved CDMA Performance Using Parallel Interference Cancellation,” JPL pub. 95~21, Oct. 1995 (patent pending) [21] N.S CORREAL, R. M. BUEHREE,
and B. D. WORNER, “A DSP-Based
DS/CDMA multiuser receiver
employing partial interference
cancellation,” IEEE J. Select. Areas Commun., vol. 17, no. 4, pp. 613~630, Apt. 1999.
[22] Divslar. D., Simon, M. K. and Raphaeli, D., “Improved parallel interference
cancellation for CDMA”, IEEE Trans. Commun., 1998, 46, (2), pp.258~268
Figure 1. The partial PIC structure
Figure 3. BER vs. user numbers
(asynchronous, perfect power control,
SNR=10 dB).
Figure 4. BER vs. user numbers (asynchronous, SNR=10 dB, SIR=-6 dB)
