A new method is proposed for the classification of data in a spatial context, based on the minimization of a variance-like criterion taking into account the spatial correlation structure of the data. Kriging equations satisfying classification bias conditions are then derived for interpolating the rainfall data while taking into account the classification.

During the registration procedure, redundant data are usually generated. We present two different approaches for data reduction. The first method is based on recursive volumetric division for determining locally saturated regions. The second method uses an interpolating surface obtained through Radial Basis Functions. Some numerical tests regarding data reduction and quality preservation are de...

Design of high speed low voltage data converters for UWB communication systems (May 2006) Choong Hoon Lee Chair of Advisory Committee: Aydin Karsilayan For A/D converters in ultra-wideband (UWB) communication systems, the flash A/D type is commonly used because of its fast speed and simple architecture. However, the number of comparators in a flash A/D converter exponentially increases with an ...