You are here

Advanced Signal Processing

Research topics:

Advanced Space-Time Processing: Beamforming, Channel Equalisation and Channel Estimation
Prof Lajos Hanzo, Prof. Shen Chen, Dr. Yosef Akhtman, Dr. Andreas Wolfgang
[1] Symmetric complex-valued RBF receiver for multiple-antenna aided wireless systems.
[2] Constant Modulus Algorithm Aided Soft Decision Directed Scheme for Blind Space-Time Equalisation of SIMO Channels.
[3] Reduced-Complexity Near-Maximum-Likelihood Detection for Decision Feedback Assisted Space-Time Equalization. [More Publications]

The ever-increasing demand for an improved throughput in wireless communication has motivated the development of adaptive antenna-
array-assisted spatial processing techniques in order to further improve the achievable spectral efficiency. A specific technique that has shown real promise in achieving substantial capacity enhancements is constituted by adaptive beamforming. The concept of beamforming is traditionally defined as a linear spatial filtering. Upon appropriately combining the signals received by the antenna array linearly, adaptive beamforming is capable of separating user signals transmitted on the same carrier frequency, provided that the signal sources are sufficiently separated in the angular domain. The name beamforming comes from the classical interpretation of the beampattern of the linear spatial filter. The beampattern is basically the discrete Fourier transform (DFT) of the linear spatial filter’s weights. The classical beamforming design, which has its root in other applications, such as radar and sonar, aims to ensure this DFT has a maximum response at the desired signal’s direction and try to minimize its response at the interfering signals’ directions. In order to create a maximum beam for the desired user and place nulls in the directions of the interfering users, it is necessary that the number of users supported is no more than the number of receive antenna elements. If this condition is not met, the system is referred to as rank-deficient or overloaded, and the classical beamforming will fail. Moreover, the classical beamforming is “zero forcing,” which is well known to suffer from a serious noise enhancement problem. A better design for the linear beamformer is the minimum mean square error (L-MMSE) solution, which trades off rejecting interference from amplifying noise. The optimal solution for the linear beamforming has been shown to be the minimum bit error rate (L-MBER) design. The L-MBER beamforming outperforms the L-MMSE one, particularly in hostile rank-deficient scenarios.

It is well known that digital communication signal detection, in general, can be viewed as a classification problem, where the receiver detector simply classifies the received multidimensional channel-impaired signal into the most likely transmitted symbol constellation point or class. For the multiple-antenna-aided beamforming receiver, if one is willing to extend the beamforming process to nonlinear, substantial performance enhancement can be achieved over the linear beamforming at the cost of an increased complexity. The idea of nonlinear beamforming has recently been developed for wireless systems with the real-valued binary phase shift-keying modulation, where a symmetric radial basis function (SRBF) network is proposed to adaptively implement the optimal nonlinear beamforming solution. This study extends nonlinear beamforming to wireless systems that employ the complex-valued quadrature phase-shift keying (QPSK) modulation scheme. For QPSK systems, the optimal Bayesian detection solution can be expressed as a complex-valued radial basis function (RBF) network. We further exploit the inherent symmetry of the optimal nonlinear beamforming solution and propose a novel complex-valued SRBF network for adaptively implementing
the Bayesian beamforming solution. It is worth pointing out that our proposed nonlinear spatial filtering approach can be interpreted as a
generalized beamforming. Instead of using the classical beampattern, which has a rather limited application, it is natural to interpret or to
visualize the a posterior probability as the generalized “beampattern” of this nonlinear spatial filter. Thus, the optimal design is to maximize
the a posterior probability for the desired user, which also implies to minimize the a posterior probabilities for the interfering users.

Advanced Signal Detection: Minimum Bit Error Ratio, Sphere Detection and Markov Chain Monte Carlo
Prof Lajos Hanzo, Prof. Shen Chen, Dr. Yosef Akhtman, Dr. Andreas Wolfgang, Dr Li Wang, Dr Shuang Tan, Dr. Shinya Sugiura
[1] Three-Stage Turbo MBER Multiuser Beamforming Receiver using Irregular Convolutional Codes.

[2] An Optimized-Hierarchy-Aided Approximate Log-MAP Detector for MIMO Systems.
[3] Three-stage irregular convolutional coded iterative center-shifting K-best sphere detection for soft-decision SDMA-OFDM.
[4] Reduced-complexity iterative Markov chain MBER detection for MIMO systems. [More Publications]

The most potent space-time detection method found in the literature is constituted by the Sphere Decoder (SD), which is capable of achieving the optimum performance of the Maximum Likelihood (ML) detector at a relatively low computational complexity. Against this background, we propose a novel SDM detection method, which we refer to as the Soft-output OPtimized HIErarchy (SOPHIE) Spatial Division Multiplexing
(SDM) detector. The proposed method may be regarded as an advanced extension of the SD methods. More specifically, our method can be employed in the above-mentioned rank-deficient and hence overloaded scenario, where the number of transmit antenna elements exceeds
that of the receive antenna elements. Furthermore our scheme is suitable for high-throughput modulation schemes such as 16- and 64-QAM. We introduce a list of optimization rules, which facilitate the achievement of the near optimum BER performance of a Log-MAP detector at a relatively low computational complexity. The trade-off between the achievable BER performance and the associated computational complexity
is controlled using two parameters. The proposed detection method exhibits two major advantages over all previously proposed techniques:
1) The bit-related soft information, which facilitates the achievement of near-optimum Log-MAP performance, is attained at the expense of a modest complexity increase over that of hard-decision ML detection.
2) Our method exhibits a particularly low polynomial complexity in both the low- and high-SNR regions. In the critical range of SNR values, which corresponds to the “waterfall” region of the BER versus SNR curve, the detection complexity versus the number of transmit antennas
remains exponential. Nevertheless, we demonstrate that the complexity can be dramatically reduced at the cost of a minor BER degradation.

Synchronization and Initial Acquisition
Prof Lajos Hanzo, Prof. Lie-Liang Yang, Dr. Seung-Hoon Hwang
[1] Iterative Code Acquisition for DS-UWB Downlink using Multiple-Component Decoders.
[2] Differential Acquisition of m-Sequences using Recursive Soft Sequential Estimation.
[3] Acquisition of m-Sequences using Soft Sequential Estimation. [More Publications]

This sequential estimation method exploits the principle of iterative soft-in–soft-out (SISO) decoding for enhancing the acquisition performance, and that of differential preprocessing for the sake of achieving an enhanced acquisition performance, when communicating over
various communication environments. Hence, the advocated acquisition arrangement is referred to as the differential recursive soft sequential estimation (DRSSE) acquisition scheme. The DRSSE acquisition scheme exhibits a low complexity, which is similar to that of an -sequence generator, while achieving an acquisition time that is linearly dependent on the number of stages in the -sequence generator. A low acquisition time is achieved with the advent of the property that the proposed DRSSE scheme is capable of determining the real-time reliabilities associated with the decision concerning a set of, say, consecutive chips. This set of consecutive chips constitutes the sufficient initial condition for enabling the local -sequence generator to produce a synchronized local despreading -sequence replica. Owing to these attractive characteristics, the DRSSE acquisition scheme constitutes a promising initial synchronization scheme for acquisition of long -sequences, when communicating over various propagation environments.

Evolutionary Algorithms: Genetic Algorithm, Ant Colony Optimisation, Particle Swarm Optimisation and Harmony Search
Prof Lajos Hanzo, Dr. Min Jiang, Dr. Chong Xu, Dr. Rong Zhang, Mr. Wang Yao
[1] Iterative Joint Channel Estimation and Multi-User Detection for Multiple-Antenna Aided OFDM Systems.
[2] Near-Optimum Multiuser Detectors Using Soft-Output Ant-Colony-Optimization for the DS-CDMA Uplink.
[3] Minimum bit error rate multiuser transmission designs using particle swarm optimisation.
[4] Iterative Multiuser Detection and Channel Decoding for DS-CDMA Using Harmony Search. [More Publications]

Multiple-Input-Multiple-Output (MIMO) Orthogonal Frequency Division Multiplexing (OFDM) systems have recently attracted substantial research interest. However, compared to Single-Input-Single-Output (SISO) systems, channel estimation in the MIMO scenario becomes more challenging, owing to the increased number of independent transmitter receiver links to be estimated. In the context of the Bell LAyered Space-Time architecture (BLAST) or Space Division Multiple Access (SDMA) multi-user MIMO OFDM systems, none of the known channel estimation techniques allows the number of users to be higher than the number of receiver antennas, which is often referred to as a “rank-deficient” scenario, owing to the constraint imposed by the rank of the MIMO channel matrix. Against this background, we propose a new Genetic Algorithm (GA) assisted iterative Joint Channel Estimation and Multi-User Detection (GA-JCEMUD) approach for multi-user MIMO SDMA-OFDM systems, which provides an effective solution to the multi-user MIMO channel estimation problem in the above-mentioned rank-deficient scenario. Furthermore, the GAs invoked in the data detection literature can only provide a hard-decision output for the Forward Error Correction (FEC) or channel decoder, which inevitably limits the system’s achievable performance. By contrast, our proposed GA is capable of providing “soft” outputs and hence it becomes capable of achieving an improved performance with the aid of FEC decoders. A range of simulation results are provided to demonstrate the superiority of the proposed scheme.

[top] [Main Page]