Engineering Research Seminar Series 5, 2011: (i) “MU-MIMO: A Fukunaga Koontz Transform (FKT) based Design of Sub-optimal Linear Precoder” (ii) "A Region Based Adaptive (RBA) Distributed Video Coding (DVC) Paradigm”

Assoc. Prof. Dr. Varun Jeoti, Department of Electronic and Electrical Engineering, Universiti Teknologi PETRONAS

Date: 2011-02-24
Time: 14:00 to 15:30
Venue: Classroom 9-3-05


(i) We present a Multi-user multiple input multiple output (MU-MIMO) linear precoding scheme utilizing Fukunaga-koontz Transform (FKT). Our work is motivated by the precoding scheme developed by Sadek, Tarighat, and Sayed that maximizes the signal to leakage noise ratio (SLNR) and obtains the precoding matrix by both generalized eigenvalue decomposition (GEVD) and generalized singular value decomposition (GSVD). However, precoding matrix computation using GEVD gives poor results at high SNR values. Though more accurate results are obtained using GSVD, but it comes with high computational load. To reduce the computational load and maintain accurate results while also relaxing the dimensionality constraint, an FKT based precoding algorithm is proposed. The method is derived to compute precoding matrices for multiple users by removing redundant computation of FKT transform factor for each user.


(ii) The practical implementation of DVC is commonly referred to as Wyner-Ziv video coding (WZ). The WZ codec assumes the side information generated at the decoder with the help of reference frame to be an estimate of the original frame. The compression is achieved by sending only that extra information that is needed to correct the estimation error. Due to the limitation in predicting the side information accurately, the predicted frame is expected to have varying degree of success. In this work, we propose partitioning the considered frame into many coding units (regions) where each unit is encoded differently.  This partitioning allows appropriate virtual dependency channel model for each region, which in turn helps in allocating proper DVC coding parameters, such as, rate, and quantizer depth to each region. The proposed solution preserves the simplicity of the video encoder by performing the frame partitioning at the decoder during side information generation process. Experimental results show that adaptive DVC parameters selection per region enhances the overall PSNR performance when compared with frame based DVC codecs.

About the Speaker

Assoc.Prof. Dr. Varun Jeoti received his Ph.D. degree from Indian Institute of Technology Delhi India in 1992. He worked on several sponsored R&D projects in IIT Delhi and IIT Madras during 1980 to 1989 developing Surface Acoustic Wave Pulse Compression filters, underwater optical receivers etc.. He was a Visiting Faculty in Electronics department in Madras Institute of Technology for about 1 year during 1989 to 1990 and joined Delhi Institute of Technology (presently called Netaji Subhash Institute of Technology) for next 5 years till 1995. He moved to Electrical & Electronic Engineering (E&E Engg) department of Universiti Sains Malaysia in 1995 and joined E&E Engg department of Universiti Teknologi PETRONAS in 2001 where he leads the Advanced Communication Engineering cluster.

His research interests are in signal processing and wireless communication relating to wireless LAN and MAN technologies, LTE, cognitive radio, maritime communication, distributed video coding and security.