Robust Relay Network Beamforming: A Probabilistic Constrained Optimization Approach

Visiting researcher: Dr Pei-Jung Chung (UoE)
Host researcher: Prof. Ming Chen (SEU)
Duration: 3 months
Status: Completed

Biography of visiting researcher

Pei-Jung Chung received Dr.-Ing. degree in 2002 from Ruhr-University Bochum, Germany with distinction. From 1998 to 2002 she was with Signal Theory Group, Department of Electrical Engineering and Information Sciences, Ruhr-University Bochum, Germany. From 2002 to 2004 she held a post-doctoral position at Carnegie Mellon University and University of Michigan, Ann Arbor, USA, respectively. From 2004 to 2006 she was Assistant Professor with National Chiao Tung University, Hsin Chu, Taiwan. In 2006 she joined the Institute for Digital Communications, School of Engineering, the University of Edinburgh, UK as a Lecturer.
She has authored over 50 papers in leading international journals and conferences, including five IEEE Trans Signal Processing papers. She has served in technical program committees of major conferences including EUSIPCO 2007 -2010. In 2010, she becomes associate member of IEEE Signal Processing Society Sensor Array Multichannel (SAM) Technical Committee. Her current research interests include array processing, statistical signal processing, wireless MIMO communications and distributed processing in sensor networks.

Background and Motivation

Beamforming is a versatile and powerful approach to receive, transmit or relay signals of interests in a spatially selective way in the presence of noise and interference. In the last decade, there has become a major research area in wireless communications, where multiantenna systems have emerged as one of the enabling technologies to accommodate rapidly growing data rates. A key assumption in most beamforming techniques is the availability of perfect channel state information. However, imperfections in channel estimates are inevitable in practical cellular networks due to finite feedback channel capacity, quantization errors and time-varying propagation environments. The channel errors often lead to severe performance degradation . Hence, robustness becomes a critical issue in practical wireless communication systems.
Current robust beamforming designs can be categorized into the worst-case analysis and the stochastic approach. The worst-case design considers channel errors as deterministic and bounded. The optimal beamformer is obtained by maximizing the worst-case performance. Since extreme errors occurs at a low probability, this approach often leads to an overall conservative performance. On the other hand, the stochastic approach focuses on average performance by incorporating channel statistics into design criteria. Without paying attention to the extreme case, it may have severe consequences in the extreme cases. The recently proposed probabilistic constraint optimization based approach provides a much more flexible solution by taking extreme conditions into account proportionally. It maximizes the average system performance and ensures robustness against channel uncertainties by keeping the probability of worst-case performance at a low level. This method offers the highest robustness and outperforms several popular transmit beamformers in large channel errors.
In the proposed research, we will extend this promising probabilistic constraint based beamforming technique to cooperative transmission in networks that shares the antenna sources and relays the signals in order to maximize spectral and power efficiency. Compared to conventional smart antenna technology, the introduction of relay nodes causes asynchronization and frequency selective channels. How to incorporate these effects along with channel imperfections into the optimal beamforming design is a challenging and important issue for the next generation wireless communications.

Research Activities

The aims of the proposed work are to develop relay network beamforming techniques that optimize average system performance and keep the requirements on quality of service of network users satisfied. More specifically, we shall
  1. Identify proper system models for relay networks that take channel uncertainty into account.
  2. Develop design criteria by considering various performance metrics such as capacity, signal interference noise ratio (SINR), bit error rate (BER) and define quality of service (QoS) requirement of network users.
  3. Formulate network relay beamforming as probabilistic constrained optimization problem and convert it into convex optimization problem so that it can be efficiently solved by modern optimization software tools.


To be updated
Powered by Drupal, an open source content management system