Research Overview
Advanced signal processing research within Next Generation Wireless addresses the reception and recovery of information in the presence of interference, uncertain radio channels and hardware complexity constraints. The original research programme is organised into four related areas: space–time processing, advanced detection, synchronisation and initial acquisition, and evolutionary optimisation for communication systems.
The topics below retain the historical research contributions and publication links from the Southampton Wireless Group programme.
Advanced Space-Time Processing: Beamforming, Channel Equalisation and Channel Estimation
Researchers: Prof Lajos Hanzo, Prof Shen Chen, Dr Yosef Akhtman and Dr Andreas Wolfgang
Adaptive antenna arrays combine signals observed at multiple receiver elements to separate desired transmissions from interference. In conventional linear beamforming, complex antenna weights shape the directional response of the array. The array beampattern is closely related to the discrete Fourier transform of these weights, and the classical approach aims to strengthen signals arriving from the desired direction while placing nulls towards interferers.
Perfect spatial nulling becomes difficult in overloaded or rank-deficient systems, particularly when the number of simultaneously supported users exceeds the number of receive antennas. Zero-forcing designs may also amplify receiver noise. Linear minimum mean-square-error (L-MMSE) beamforming balances interference rejection against noise enhancement, whereas linear minimum bit-error-rate (L-MBER) beamforming optimises the decision error criterion directly, potentially improving performance under severe multiuser interference.
Nonlinear beamforming offers a further alternative by treating signal detection as a classification problem. For complex-valued QPSK signals, a Bayesian detector can be represented using a complex radial basis function (RBF) network. By exploiting the symmetry of the symbol constellation, the symmetric RBF (SRBF) receiver can implement the Bayesian decision structure more efficiently. In this interpretation, the detector uses posterior class probabilities, rather than only a conventional linear beampattern, to distinguish the desired user from interferers.
This research area also encompasses blind and decision-directed space–time equalisation and reduced-complexity near-maximum-likelihood detection for multiple-antenna receivers.
Selected Publications
- Symmetric complex-valued RBF receiver for multiple-antenna aided wireless systems
- Constant Modulus Algorithm Aided Soft Decision Directed Scheme for Blind Space-Time Equalisation of SIMO Channels
- Reduced-Complexity Near-Maximum-Likelihood Detection for Decision Feedback Assisted Space-Time Equalization
Advanced Signal Detection: Minimum Bit Error Ratio, Sphere Detection and Markov Chain Monte Carlo
Researchers: Prof Lajos Hanzo, Prof Shen Chen, Dr Yosef Akhtman, Dr Andreas Wolfgang, Dr Li Wang, Dr Shuang Tan and Dr Shinya Sugiura
Maximum-likelihood (ML) detection provides an important benchmark for multi-antenna receivers, but direct enumeration of possible transmitted symbol vectors becomes prohibitively complex as the number of antennas and modulation order increase. Sphere decoding reduces the search space and can approach, or attain, ML performance in suitable conditions. Soft-output detection is especially important when the detector exchanges reliability information with a forward-error-correction decoder.
The original research presents a Soft-output OPtimized HIErarchy (SOPHIE) detector for spatial division multiplexing. It is designed to produce approximate Log-MAP soft information while controlling computational cost through search and pruning parameters. The approach addresses high-order constellations, including 16-QAM and 64-QAM, as well as overloaded scenarios in which the transmitter has more antennas than the receiver.
Its complexity depends strongly on the signal-to-noise ratio (SNR). The search is comparatively inexpensive at low and high SNR, but can become substantially more demanding around the error-rate waterfall region. The research therefore examines the trade-off between near-optimal bit-error performance, the quality of soft information and computational complexity.
Complementary approaches include minimum-bit-error-rate multiuser beamforming, iterative K-best sphere detection, and Markov chain Monte Carlo techniques. These techniques aim to provide robust decisions and useful soft outputs without relying on an exhaustive search of all possible transmit vectors.
Selected Publications
- Three-Stage Turbo MBER Multiuser Beamforming Receiver using Irregular Convolutional Codes
- An Optimized-Hierarchy-Aided Approximate Log-MAP Detector for MIMO Systems
- Three-stage irregular convolutional coded iterative center-shifting K-best sphere detection for soft-decision SDMA-OFDM
- Reduced-complexity iterative Markov chain MBER detection for MIMO systems
Synchronisation and Initial Acquisition
Researchers: Prof Lajos Hanzo, Prof Lie-Liang Yang and Dr Seung-Hoon Hwang
Before a spread-spectrum receiver can reliably detect data, it must acquire the spreading sequence and align its locally generated replica with the received signal. Acquisition of long maximal-length sequences (m-sequences) can be time-consuming, particularly in challenging propagation environments.
The source research investigates soft sequential estimation, including a differential recursive soft sequential estimation (DRSSE) acquisition scheme. It combines differential preprocessing with iterative soft-input–soft-output (SISO) estimation to infer the initial state of a local m-sequence generator from the received chips.
Rather than treating each tentative chip decision as equally reliable, the receiver evaluates the reliability of a short sequence of decisions. Once sufficient initial-state information is obtained, the local sequence generator can produce a synchronised spreading-code replica. The work reports acquisition-time scaling with the number of generator stages, while retaining computational complexity comparable to that of a sequence generator.
The same research line includes iterative code acquisition for direct-sequence ultra-wideband (DS-UWB) downlinks and related soft sequential acquisition methods.
Selected Publications
Evolutionary Algorithms: Genetic Algorithm, Ant Colony Optimisation, Particle Swarm Optimisation and Harmony Search
Researchers: Prof Lajos Hanzo, Dr Min Jiang, Dr Chong Xu, Dr Rong Zhang and Mr Wang Yao
MIMO-OFDM receivers must estimate the channels linking multiple transmitting and receiving antennas while detecting overlapping user signals. The joint estimation and detection problem becomes particularly difficult in rank-deficient space-division multiple-access (SDMA) systems, where the number of users exceeds the number of receive antennas.
The original work introduces genetic-algorithm-aided iterative joint channel estimation and multiuser detection (GA-JCEMUD) for such systems. The genetic algorithm searches candidate channel and symbol configurations jointly rather than estimating the channel and detecting data in isolation. Importantly, it can generate soft reliability information for the channel decoder, avoiding the limitations of hard-output evolutionary detection.
The broader programme studies other biologically inspired and population-based optimisation tools. Ant colony optimisation is used for soft-output multiuser detection in DS-CDMA; particle swarm optimisation is applied to minimum-bit-error-rate transmission design; and harmony search supports iterative multiuser detection jointly with channel decoding.
These methods explore practical alternatives to exhaustive search, with simulation-based comparisons of detection quality, decoder compatibility and computational effort.
Selected Publications
- Iterative Joint Channel Estimation and Multi-User Detection for Multiple-Antenna Aided OFDM Systems
- Near-Optimum Multiuser Detectors Using Soft-Output Ant-Colony-Optimization for the DS-CDMA Uplink
- Minimum bit error rate multiuser transmission designs using particle swarm optimisation
- Iterative Multiuser Detection and Channel Decoding for DS-CDMA Using Harmony Search