An Antenna Diversity Scheme for Digital Front-End with OFDM Technology

1 Introduction
    In modern wireless broadband communication and digital broadcasting, orthogonal frequency-division multiplexing (OFDM)-based modulation schemes are usually used. However, an OFDM system has very poor reception when there is noise, interference, or moving objects. To solve this problem, many technologies have been proposed and used in real applications [1], [2]. Among these technologies, beamforming-based diversity technology is the most promising. It uses multiple antennas at the receiver side and spatial filtering to optimize reception in noisy and mobile environments. Figs. 1, 2, and 3 show  three representative solutions: pre-fast Fourier transform (pre-FFT), polyphase filter-bank, and post-FFT [1], [4], [5], [7], [8].

    In Fig. 1, X1(n ), X2(n ), ... , XM (n ) are the received signals in one of M antennas, and W1(n ), W2(n ), ... , WM (n ) are the weights applied to these antenna signals. All the weighted signals are summed to one channel, such as one antenna’s output and then forwarded for further FFT processing, which is required in any OFDM-based receiver. These weights are designed to meet an optimization criterion, and the most common criterion is the maximum ratio combination (MRC). With MRC, the optimized weight vector, W (n ) = [W1(n ), W2(n ), ..., WM (n )], can be obtained by

    where X (n ) = [X 1(n ), X 2(n ), ..., X M (n )] and y (n ) =W H (n )X (n ). Pre-FFT can give 5-10% gain in bit-error reduction and is not complex in implementation. However, this processing is done only in the time domain, and the same weights are used over the entire frequency band (all-carriers). Because of the shortcomings of simple pre-FFT, a polyphase filter bank scheme is proposed (Fig. 2).

    In a polyphase scheme, each antenna signal is divided into a number of frequency band signals, and different weights apply to different bands. This scheme can further improve performance at the cost of increased computational complexity that arises as a result of multichannel antenna signal decomposition (analysis) and multiplication of weights in each channel. Different weights in different frequency bins and bands can be used in the frequency domain, and this is called post-FFT (Fig. 3).

    A polyphase scheme is a special case of post-FFT. Post-FFT performs the best but has the greatest computational complexity. Pre-FFT has the least computational complexity but improves performance the least. Another approach is to divide M antennas into L  groups. Each group uses pre-FFT and obtains L  outputs. With these L  outputs, post-FFT is then used to obtain the desired outputs. Compared with the full post-FFT scheme, this alternative avoids M-L FFT computations at the cost of reduced performance. A diversity scheme that provides a better compromise between performance, complexity, and power consumption is highly desired.

2 The Proposed Scheme
    Here, we propose a new scheme in which the number of variables computed is reduced from N ×M (in post-FFT) to N+M. The scheme shown in Fig. 4 performs just as well as the post-FFT scheme.

    M antenna signals are weighted by the corresponding weight as the processing made in the pre-FFT algorithm in Fig. 1. These M weighted signals are added together to pass through an N-tap finite impulse response (FIR) eigenfilter, after which they are input for FFT processing. The number of unknown variables in the proposed scheme is N +M, which is significantly less than the N ×M needed in post-FFT. Furthermore, only one FFT operation is needed in the proposed scheme instead of M FFT operations in the post-FFT scheme. Here we will prove that the post-FFT scheme in Fig. 3 and the proposed scheme in Fig. 4 give the same frequency-domain outputs, Z (1), Z (2),...,Z (N ). Using the optimization criterion in post-FFT, the variable matrix, WN ×M, in Fig. 3 can be rewritten as

    where WN,i  is the i th column of the N ×M matrix and denotes the weights after FFT processing of the i th antenna in Fig. 3. WC is an N-dimensional vector, and Ci  is a scalar (corresponding to each antenna). This suggests that Figs. 3 and 5 give the same outputs.
Because each block in Fig. 5 is linear processing, changing the order of the blocks results in the same performance. Hence, the scheme in Fig. 4 is obtained.

    Now we determine the N +M weights in proposed scheme.

    WE, X h, and X t  denote the weight vector of the eigenfiltering, the sequence of the head-guided interval, and the tail of the symbols, respectively. If the norm is a constant (unity, for instance), the weight vector is obtained in such a way that the error is minimized. That is,

According to beamforming matrix theory [3], the solution to the optimization problem is the eigenvector (minor component) corresponding to the smallest eigenvalue of the correlation matrix R, which is

    This is also the reason that the filtering in Fig. 4 is called eigenfiltering. In practical implementation, the minor component in the following adaptive algorithm can be updated:

    where γ is a constant. This algorithm is as complex as LMS algorithm, and the required multiplications are only on the order of N. An algorithm, such as MRC used in pre-FFT, can be used to determine the weights applied to each antenna before eigenfiltering.

3 Comparison and Discussion
    Compared with existing solutions, the proposed scheme has the following features:

• The parameters to be computed in a real-time implementation are reduced from M ×N to N +M, and performance remains the same as that of post-FFT.
• Only one FFT operation is needed instead of multiple FFT computations or polyphase filtering analyses. Table 1 shows computational complexity for various standards.
• The related parameters can be adaptively updated from the received samples of the each antenna. The adaptive algorithm used is as simple as the LMS algorithm, which has the computational complexity on the order of the number of unknown parameters.

    With these features, the proposed scheme is a new tool for improving reception quality in broadband wireless communications and digital broadcasting. It will have very many practical uses in devices.

4 Conclusions and Future Work
    In this paper, we have proposed a new and efficient diversity scheme for OFDM-based receivers. We have also shown the accuracy and effectiveness of the proposed scheme. When embedding this algorithm into real silicon, it is often desirable that one platform support all existing standards (and their various modes) with high power efficiency, low cost, and short time to market [9]. Traditional integrated circuit technologies, such as application-specific integrated circuits (ASICs) and digital signal processors (DSPs), are not highly flexible and power efficient. An ASIC solution gives high performance with low power consumption and price, but it cannot support multiple standards nor is it sufficiently flexible. A DSP is highly flexible but performs poorly and consumes much power. It may include high-speed arithmetic operations, such as multiply-accumulate, but the algorithms require extensive programming and more parallelism than can be offered by a general DSP.

    One alternative is to use a general DSP or reduced instruction set computing (RISC) processor plus hardware accelerators that are designed and optimized to implement FFT algorithm as well as eigenfiltering and its adaptive algorithm. In other words, an eigenfiltering unit performs a basic filter computation, an FFT unit performs FFT computations, and an adaptive unit implements the operations defined by (5) in section 2. These optimized accelerators may meet performance goals, but the accelerators are very narrow in their applicability, and this significantly reduces the flexibility of the processor-based solution.

    An alternative that offers flexibility and parallelism is devices with FPGAs. These devices may combine processors with a programmable array of low-level logic devices, but they are expensive, and performance is limited at high temperatures.

    In a future paper, we will introduce elemental computing array (ECA), which is designed to achieve a better compromise between performance, flexibility, cost, and power efficiency. With ECA, the FGPA scheme can have performance, power, and cost similar to ASICs. ECA is far more flexible than a DSP with accelerators. It is also much cheaper and consumes far less power than FPGA devices.

References
[1] D. H. Pham, J. Gao, T. Tabata, H. Asato, S. Hori, T. Wada, “Implementation of joint pre-FFT adaptive array antenna and post-FFT space diversity combining for mobile ISDB-T receiver,” IEICE Trans., vol. E91-B, no.1, pp.127-137, 2008.
[2] Fa-Long Luo, Multimedia Mobile Broadcasting Standards: Technology and Practice Springer, 2008.
[3] Fa-Long Luo, R. Unbehauen, Applied Neural Networks for Signal Processing. Cambridge: Cambridge University Press, 1999.
[4] I. Nobuo, T. Kenichi, “HDTV Mobile Reception in Automobiles,” Proc. IEEE, vol.94, no.1, pp.274-280, 2006.
[5] T. Tabata, H. Asato, Dang Hai Pham, M. Fujimoto, N. Kikuma, S. Hori, T. Wada. “Experimental study of adaptive array antenna system for ISDB-T high-speed mobile reception,” Proc. IEEE Int. Symp. Antenna and Propag., pp.1697-1700, Hawaii, 2007.
[6] Fa-Long Luo and R. Unbehauen, “A minor subspace analysis algorithm,” in IEEE Trans. Neural Netw., vol. 8, no.5, 1997, pp1149-1155.
[7] M. Kornfeld, G. May, “DVB-H and IP datacast-broadcast to handheld devices,” IEEE Trans. Broadcasting, vol. 53, no.1, pp. 161-170, 2007.
[8] M. Chari, F. Ling, A. Mantravadi, R. Krishnamoorthi, R. Vijayan, G. Walker, R. Chandhok, “FLO physical layer: An overview,” IEEE Trans. Broadcasting, vol. 53, no. 1 pp. 145-160, Mar. 2007.
[9] Fa-Long Luo, Digital Front-End  in Wireless Communications and Broadcasting, Circuits and Signal Processing. Cambridge: Cambridge University Press, 2011.

Biographies

Fa-Long Luo (falong.luo@elementcxi.com) is chief scientist at Element CXI. He has been the editor-in-chief of the International Journal of Digital Multimedia Broadcasting since 2007. Fa-Long Luo is now the chairman of the IEEE Industry DSP Standing Committee and technical board member of the IEEE Signal Processing Society. He has 28 years of research and industry experience in multimedia, communications, and broadcasting with real-time applications and standards. Fa-Log Luo has received worldwide recognition. He has authored and edited 4 books, more than 100 technical papers and 18 patents.

Ward Williams (ward.williams@elementcxi.com) is vice president of marketing for Element CXI. Ward has broad technology marketing and business development experience in senior management positions. He has worked for companies such as Xilinx, Chips and Systems, S3/SonciBLUE, and Hewlett-Packard. He frequently speaks at international technology shows or forums such as IEEE ICASSP and Global Mobile Congress. His research interests include digital front-end, software-defined radio, single-chip solutions, SoC-based hardware, and software IPs.

Bruce Gladstone (bruce.gladstone@elementcxi.com) has worked in the semiconductor and EDA industries for over 20 years. After graduating from the University of Illinois, he went to Silicon Valley to work in IC design, applications, and technical marketing positions. Bruce has worked for companies such as AMD, Western Digital, Synopsys, and other companies in California and Japan. As a senior director at Element CXI, he is focused on putting new-generation programmable semiconductors into the next-generation communications products.

[Abstract] In this paper, we propose a new antenna diversity scheme for OFDM-based wireless communication and digital broadcasting applications. Compared with existing schemes, such as post-fast Fourier transform (FFT), pre-FFT, and polyphase-based filter-bank, the proposed scheme performs optimally and has very low computational complexity. It offers a better compromise between performance, power consumption, and complexity in real-time implementation of the receivers of broadband communication and digital broadcasting systems.

[Keywords] OFDM; digital front-end; MIMO; cross-layer processing; diversity; antenna