Grant: Discovery Grants Program - Individual

Digital signal processing (DSP) techniques have found a variety of applications over the last 30 years in processing of signals in such diverse areas as biomedical signal processing, acoustics, telecommunications, image and video signal processing, and seismology. The availability of high-resolution, inexpensive, high-speed A/D converters and powerful digital signal processors in VLSI form has provided the necessary impetus for the ever-expanding applications of DSP in these areas. Multi-dimensional DSP techniques are being increasingly used in the processing of multi-dimensional signals such as pictures and moving pictures obtained from satellites, x-rays, and other sources. Videophone and HDTV are some examples of products that have gained from the research and development advances in the 2-D and 3-D DSP area. The overall objectives of the proposed research are to develop a variety of efficient and reliable algorithms for filtering and processing of signals (such as image, video, speech, and biomedical signals). In particular, we will concentrate over the next 5 years on the following projects. (a) We intend to explore the use of discrete Tchiebichef transforms (DTT) in place of DCT for image compression, since the energy compaction and decorrelation properties of DTT are very similar to that of DCT and at the same time fast multiplier-less implementations are possible for DTT. Special attention would be paid to the implementation issues for DTT as well as for many of the discrete transforms and algorithms that we have proposed in the last few years. (b) A key challenge in realizing a cognitive radio is the dynamic adaptation needed, which can be made possible by the integration of reconfigurable digital filters and filter banks in the receiver architecture. This subproject will address the design of area efficient reconfigurable filters, which will enable the design of reconfigurable filter banks with independent control over the bandwidths & center frequencies of its subbands. (c) We propose to develop an intelligent surveillance system with the ability to detect, categorize and recognize human action, and ask for human attention only when necessary. Such a system could be used for home and building surveillance, parking lot monitoring, etc. It can enhance the health services by determining the condition of a patient or elderly people in a remote area. (d) We intend to address the problem of tracking direction of arrival of multiple moving targets. This is of great interest in a variety of commercial & military applications. (e) Development of algorithms for biomedical applications. In particular, we intend to develop methods to address the processing of long-term EEG recordings The results from these projects should be beneficial to the telecommunication, medical signal processing, VLSI, image and video signal processing, and wireless sectors of the industry.