Evolutionary Computation:

My early research focused on EC and it remains close to my heart to this day. I feel that as computing power becomes more plentiful and cheaper evolutionary algorithms will become more and more frequent, including on-the-fly genetic programming for robotic controllers and onboard parameter optimization via genetic algorithms and/or CMA-ES for systems using probabilistic graphical models in intractable situations.

Swarm Robotics:

No surprise that my other interests are also in bio-mimetic areas. I firmly believe that swarm-robotic systems will flourish in the near future. Low-cost, fault-tolerant systems are nigh non-existant in modern robotics, but if robotics are to be as ubiquitous in the future are they are currently expected to be then such systems must be developed. Much as computing has moved from its originally envisioned model of large, expensive, central processing to low-cost distributed systems so robotics must follow.

My undergraduate research, as well as all of my publications so far, have been focused on the evolution of new wavelets for wavelet-based digital image compression. The basic technique is to take the coefficients of an existing wavelet as the starting point and feed them into a genetic algorithm, evolving a new set of coefficients. These coefficients are then tested by using them to compress a set of images with similar properties. The result is then compared to the same images compressed with the original wavelet, and the Mean Squared Error between the two images/image sets is used as the fitness value of that coefficient set.

My masters research so far has been performed in concert with USC's Information Sciences Institute. The ISI's Polymorphic Robotics Lab, led by Dr. Wei-Min Shen, has developed a reconfigurable robot named SUPERBOT. My research has been working to develop swarm-level reconfiguration routines for SUPERBOT, with publication expected in the year 2012.