Research Interests

Since coming to the University of Washington in 2008, my research has been driven by one overwhelming factor: waste. Waste in healthcare and waste in human consumption. I have positioned myself in a unique role, supporting healthcare and resource management applications via the integration of sophisticated signal processing techniques into ubiquitous computing technologies. Projects that I am currently working on are: 1) mobile phone based spirometry and cough detection, (2) real-time water consumption sensing, and (3) New touch sensing technologies for interactive projection (especially thermal and depth imaging). 

Healthcare has become an explosive topic over the past two decades. With the more recent debates surrounding healthcare, it is reaching a point of incredible scrutiny and reform. In particular, access and cost are becoming primary concerns. We need systems that cater to all individuals, while decreasing cost. Washington Post correspondent T. R. Reid explains that the new burning question in healthcare is not who gets covered, but what gets covered. And, the what depends largely on cost. This perspective advocates that it is not enough to create new procedures that save lives: breakthroughs in medical research usually come with a considerable pricetag - to make access more widespread, we must make it more affordable. Assistive technologies that run from home or mobile applications are essential. The signal processing behind these applications is not trivial. It requires realistic modeling of the physics behind the processes and sophisticated, stochastic models for what we cannot model deterministically. I am working to support these types of healthcare initiatives by creating robust signal processing algorithms for medical care that use low-cost sensors attached to (or already embedded in) the mobile phone. 

I also work on signal processing that supports sustainability research using a technique known as Infrastructure Mediated Sensing (IMS). IMS infers activity in a space or residence by sensing at a single point along a home's infrastructure. For example, using a single pressure sensor on a home's plumbing system to monitor water activity throughout the entire home. Addressing sustainability and activity recognition via single point sensing promises to revolutionize eco-feedback interfaces and other modern ubiquitous applications. Aside from water, I support IMS applications that integrate with the electrical and natural gas home infrastructure. Single point network sensing, however, is the broader impact of this research. I am working on stochastic modeling of human resource usage that parallels many other human driven activities and, therefore, has an unknown potential outside of consumption. The tools that I am currently developing could be used by many others in fields of psychology, engineering, and human-computer interaction.