Wai Ho Mow

A VLSI architecture of a K-best lattice decoding algorithm for MIMO channels

Lattice decoding algorithms have been proposed for implementing the maximum likelihood detector (MLD), which is the optimal receiver for multiple-input multiple-output (MIMO) channels. However the computational complexity of direct implementation of the lattice decoding algorithm is high and the throughput is variable. In this work, a K-best algorithm is proposed to implement the lattice decoding. It is computational inexpensive and has fixed throughput. It can be easily implemented in a pipelined fashion and has similar performance as the optimal lattice decoding algorithm if high value of K is used. In this paper, we describe a pipelined VLSI architecture for the implementation of the K-best algorithm. The architecture was designed and synthesized using a 0.35 /spl mu/m library. For a 4-transmit and 4-receive antennas system using 16-QAM, a decoding throughput of 10 Mbit/s can be achieved.

Complex Lattice Reduction Algorithm for Low-Complexity Full-Diversity MIMO Detection

Recently, lattice-reduction-aided detectors have been proposed for multiinput multioutput (MIMO) systems to achieve performance with full diversity like the maximum likelihood receiver. However, these lattice-reduction-aided detectors are based on the traditional Lenstra-Lenstra-Lovasz (LLL) reduction algorithm that was originally introduced for reducing real lattice bases, in spite of the fact that the channel matrices are inherently complex-valued. In this paper, we introduce the complex LLL algorithm for direct application to reducing the basis of a complex lattice which is naturally defined by a complex-valued channel matrix. We derive an upper bound on proximity factors, which not only show the full diversity of complex LLL reduction-aided detectors, but also characterize the performance gap relative to the lattice decoder. Our analysis reveals that the complex LLL algorithm can reduce the complexity by nearly 50% compared to the traditional LLL algorithm, and this is confirmed by simulation. Interestingly, our simulation results suggest that the complex LLL algorithm has practically the same bit-error-rate performance as the traditional LLL algorithm, in spite of its lower complexity.

On the Performance of the MIMO Zero-Forcing Receiver in the Presence of Channel Estimation Error

By employing spatial multiplexing, multiple-input multiple-output (MIMO) wireless antenna systems provide increases in capacity without the need for additional spectrum or power. Zero-forcing (ZF) detection is a simple and effective technique for retrieving multiple transmitted data streams at the receiver. However the detection requires knowledge of the channel state information (CSI) and in practice accurate CSI may not be available. In this letter, we investigate the effect of channel estimation error on the performance of MIMO ZF receivers in uncorrelated Rayleigh flat fading channels. By modeling the estimation error as independent complex Gaussian random variables, tight approximations for both the post-processing SNR distribution and bit error rate (BER) for MIMO ZF receivers with M-QAM and M-PSK modulated signals are derived in closed-form. Numerical results demonstrate the tightness of our analysis

The evolution path of 4G networks: FDD or TDD?

Frequency-division duplexing and time-division duplexing are two common duplexing methods used in various wireless systems. However, there are advantages and technical issues associated with them. In this article we discuss in detail the features, and the design and implementation challenges of FDD and TDD systems for 4G wireless systems. In particular, we present a number of advantages and flexibilities an TDD system can bring to 4G systems that an FDD system cannot offer, and identify the major challenges, including cross-slot interference, in applying TDD in practice. Due to the fact that cross-slot interference is one of the critical challenges to employing TDD in cellular networks, we also provide a quantitative analysis on its impact on co-channel and adjacent channel interfering cells

Universal lattice decoding: principle and recent advances

The idea of formulating the detection of a lattice-type modulation, such as M-PAM and M-QAM, transmitted over a linear channel as the so-called universal lattice decoding problem dates back to at least the early 1990s. The applications of such lattice decoders have proliferated in the last few years because of the growing importance of some linear channel models such as multiple-antenna fading channels and multi-user CDMA channels. The principle of universal lattice decoding can trace its roots back to the theory and algorithms developed for solving the shortest/closest lattice vector problem for integer programming and cryptoanalysis applications. In this semi-tutorial paper, such a principle as well as some related recent advances will be reviewed and extended. It will be shown that the lattice basis reduction algorithm of Lenstra, Lenstra and Lovasz (LLL) can significantly improve the performance of suboptimal lattice decoders such as the zero-forcing and VBLAST detectors. In addition, new implementation of the optimal lattice decoder that is particularly efficient at moderate signal-to-noise ratios will also be presented. Copyright

Maximum likelihood sequence estimation from the lattice viewpoint

Considers the problem of data detection in multilevel lattice-type modulation systems in the presence of intersymbol interference and additive white Gaussian noise. The conventional maximum likelihood sequence estimator using the Viterbi algorithm has a time complexity of O(m/sup /spl nu/+1/) operations per symbol and a space complexity of O(/spl delta/m/sup /spl nu//) storage elements, where m is the size of input alphabet, /spl nu/ is the length of channel memory, and /spl delta/ is the truncation depth. By revising the truncation scheme and viewing the channel as a linear transform, the authors identify the problem of maximum likelihood sequence estimation with that of finding the nearest lattice point. From this lattice viewpoint, the lattice sequence estimator for PAM systems is developed, which has the following desired properties: 1) its expected time-complexity grows as /spl delta//sup 2/ as SNR/spl rarr//spl infin/; 2) its space complexity grow as /spl delta/; and 3) its error performance is effectively optimal for sufficiently large m. A tight upper bound on the symbol error probability of the new estimator is derived, and is confirmed by the simulation results of an example channel. It turns out that the estimator is effectively optimal for m/spl ges/4 and the loss in signal-to-noise ratio is less than 0.5 dB even for m=2. Finally, limitations of the proposed estimator are also discussed. >

Robust joint interference detection and decoding for OFDM-based cognitive radio systems with unknown interference

Cognitive radio technology facilitates spectrum reuse and alleviates spectrum crunch. One fundamental problem in cognitive radio is to avoid the interference caused by other communication systems sharing the same frequency band. However, spectrum sensing cannot guarantee accurate detection of the interference in many practical situations. Hence, it is crucial to design robust receivers to combat the in-band interference. In this paper, we first present a simple pilot aided interference detection method. To combat the residual interference that cannot be detected by the interference detector, we further propose a robust joint interference detection and decoding scheme. By exploiting the code structure in interference detection, the proposed scheme can successfully detect most of the interfered symbols without requiring the knowledge of the interference distribution. Our simulation results show that, even without any prior knowledge of the interference distribution, the proposed joint interference detection and decoding scheme is able to achieve a performance close to that of the maximum likelihood decoder with the full knowledge of the interference distribution

Adaptive Resource Allocation and Capacity Comparison of Downlink Multiuser MIMO-MC-CDMA and MIMO-OFDMA

In this paper, we examine and compare the potential maximum sum capacity of downlink multiple-input-multiple-output orthogonal frequency division multiple access (MIMO-OFDMA) and multiple-input-multiple-output multicarrier code division multiple access (MIMO-MC-CDMA) in a single-cell multiuser environment with channel side information at the transmitter, with and without a fairness constraint. The resource allocation is formulated as a cross-layer optimization framework and optimal power allocation and user selection algorithms are proposed for both scenarios. We find that for delay-sensitive applications, where fairness is imposed, the performance gain of OFDMA over MC-CDMA is quite large at moderate path loss exponents and number of antennas. However, for delay-insensitive applications, the benefits of OFDMA over MC-CDMA are significantly reduced when the path loss exponent or the number of antennas is large

Design of spreading codes for quasi-synchronous CDMA with intercell interference

The recently proposed loosely synchronized (LS) spreading code can in principle realize an intracell-interference-free quasi-synchronous code-division multiple-access (QS-CDMA) system by creating a wide enough interference-free window (IFW). However, the problem of minimizing intercell interference (ICI) in a cellular QS-CDMA system remains an open issue. Addressing the problem from a sequence design viewpoint, the key challenge is how to generalize the known construction of a single LS code to the design of many families of generalized LS (GLS) codes so that a desirable code family can be selected for the realization of a low-ICI cellular QS-CDMA system. Our main contribution is a systematic construction of new families of GLS codes with favorable intercode cross-correlation properties within a certain window, while maintaining the desirable IFW property. Many such code families can be obtained from our construction by choosing different Hadamard matrices and different uncorrelated complementary pairs. Their effectiveness with respect to some meaningful evaluation criteria are compared. In particular, by a simplified system bit-error rate analysis, it is demonstrated that a new GLS code family significantly outperforms the conventional scrambled LS codes.

A new unified construction of perfect root-of-unity sequences

Perfect root-of-unity sequences (PRUS) have found applications in numerous spread spectrum systems such as pulse compression radars, DS/SSMA, FH/SSMA, etc. A unified PRUS construction, which includes all known constructions as special cases, was previously devised by this author. In this paper, a class of root-of-unity sequence sets with certain interesting correlation properties is identified as a very important source of PRUS. By deriving a general construction of such sequence sets, a new, even more general unified PRUS construction is obtained. In addition, a new lower bound on the total number of PRUS with given alphabet size and sequence length is derived by determining the exact number of PRUS obtainable from the construction. The bound is tight for all cases, in which the exact values are known. Furthermore, it is proved that there is no new PRUS obtainable through the application of various perfectness-invariant transformations and the direct product construction. This gives strong evidences to support the conjecture that the new unified construction in fact describes all PRUS that exist.