In an increasingly digital world, computational sensing and signal acquisition is an indispensable technology. As sensors are rapidly becoming ubiquitous, I explore the significant challenges and immense opportunities presented in their deployment. My work systematically develops theory and methods to improve all aspects of signal acquisition and sensing. I am trying to advance modeling and algorithms and exploit the wide availability of computational power to overhaul the signal acquisition paradigm.

Acoustic array circa WWII

Acoustic array circa WWII

Digitization and computation has already significantly improved our sensing capabilities. Our understanding of signal spaces and sampling theorems—originating with the Whittaker-Shannon-Nyquist theory—enabled signal acquisition and processing using digital systems. Processing systems, riding the advances in computation due to Moore’s law, have provided remarkable functionality once the signal is acquired. However, these advances have not, until recently, affected our sensing capabilities or the sampling complexity.

Compressive sensing (CS) and computational imaging are recent developments demonstrating how computation can drastically improve our sampling systems. CS showed that computation combined with appropriate signal models can significantly reduce sampling complexity. Computational imaging, on the other hand, uses computation to provide capabilities not available in conventional imaging systems. My research intends to provide such capabilities in a wide variety of sensing systems. I prefer to use the term “computational signal acquisition” to emphasize the use of computation and to reduce the emphasis on the “compressive” nature of CS systems.

I am exploring the intersection of theory and practice. Applications are important and guide the development of the theory, ensuring it is relevant and current. On the other hand, exclusive focus on the application might miss the big picture. Theory is necessary to capture the fundamental concepts and make them portable and useful to a variety of applications without re-inventing the wheel. The advances in theory, modeling and algorithms are highlighted in a recent report of the President’s Council of Advisors on Science and Technology (PCAST) on Networking and Information Technology R&D (also a press release and a webcast):

in many areas, performance gains due to improvements in algorithms have vastly exceeded even the dramatic performance gains due to increased processor speed.

My interests are varied, including the effects of quantization in sensing, even down to 1-bit quantization, the interaction of CS theory and practice, signal and system models, as well as applications in synthetic aperture radar (SAR), array processing, video acquisition, and ultrasonic sensing. These pages are a small sampling of my interests, updated as projects are evolving and as time permits. For a more complete picture of my work, please see my publications and my presentations. Also, you can follow me on twitter for updates on my publications and other things I find interesting.

To stay current in this area, there are a number of resources. Igor Carron in his Nuit-Blanche blog and twitter feed keeps track of the recent developments, papers, code etc. Also, the Rice CS page is a good repository of work in the area, including papers, tutorials, presentations, code and relevant links.

Image Credit: For more acoustic radar images look here (where the image above is taken from) and here.