![]() |
![]() |
I am currently looking for graduate students to work on the following projects. Most of these projects are continuations (or are related to) previous research work in our group, and the new students would take advantage of the existing knowledge in the group (Please refer to other parts of this website for a list of previous theses and publications).
i) Iterative Decoding in Analog VLSI: This project involves ongoing research in our group on theory and implementation of iterative decoding algorithms in analog VLSI. The group has designed, implemented and tested the first ever full-CMOS (standard CMOS technology with standard biasing techniques) analog iterative decoder, and is looking for master's and Ph.D. students that are interested in analog electronics as well as communications to continue this exciting project. The students would be co-supervised with faculty members in Electronics.
ii) Robust Transmission of Image and Video over Internet and Wireless Links: Transmission of image and video while guaranteeing a certain quality of service is in great demand in many applications. The problem is challenging as digital video and images require a large amount of bandwidth and in a band-limited regime, no choice other than using compression techniques exists. These techniques however are generally very sensitive to transmission errors and the loss of a packet in the compressed stream could be catastrophic in reconstruction of the original image or video. In particular, error propagation is a serious problem for compression techniques. To guarantee that the most important part of data is received before the occurrence of the first transmission error, and to ensure that this part of data can be used to reconstruct the image (with a lower resolution or quality), progressive modes have been included in several image and video standards such as JPEG, JPEG-2000, MPEG-2, MPEG-4 and H.264. In this work, we are concerned about the robust transmission of progressively coded images and video over error-prone and loss-prone channels such as wireless links and Internet. Our focus would be on the design of error control schemes to protect compressed images and video and to obtain the highest quality of the reconstructed signal under the constraints on bandwidth (transmission bit rate) and the channel characteristics. In our design, we use advanced coding schemes such as turbo codes and low-density parity-check codes and employ techniques such as unequal error protection of the transmitted packets, combined with proper anti-loss packet marking schemes. We are concerned about not only the performance of our scheme but also its complexity. We are therefore interested in designing coding schemes which not only guarantee the highest quality of reconstructed image and video but also are simple to implement in software or hardware.
iii) Iterative Coded-Modulation Schemes: This project involves efficient design of coded-modulation schemes based on turbo codes and LDPC codes for fading channels and includes techniques such as multilevel coding (MLC) and bit-interleaved coded-modulation (BICM).
iv) Space-Time Processing and MIMO Systems: This project involves the application of iterative coding, modulation and equalization techniques in MIMO processing for efficient high-rate transmission of information over wireless channels.
v) Low-density Lattice Codes: This project involves analysis and design of lattice codes constructed based on LDPC codes. The coding gain of lattices increases with dimension but a practical limitation in using high-dimentional lattices is their exponentially increasing decoding complexity. This project which continues the work of a Ph.D. student in the group, investigates the construction of lattices with high coding gain and yet low iterative decoding complexity. This project is only open to Ph.D. students.
vi) Network Coding and Security: In the paradigm of network coding, the nodes in a network are allowed to encode the information received from the input links in a variety of ways. A very simple encoding strategy is the usual function of switching. The main goal of network coding is to achieve the full capacity of the network by proper design of encoding and decoding strategies at the network nodes. This full capacity can be well beyond the standard capacity of networks whose nodes are constrained to only perform switching. This research deals with design of coding strategies for practical networks that increase the information rate while having a reasonably low complexity. We are also interested in incorporating network coding and information security by designing secure network codes under different types of wiretaps and security attacks.