Reconstruction of 1-D signals from irregularly sampled data
A project proposal


Vamsidhar Juvvigunta	   ( vamsi@erc)  Nagamohan Kompella (mohi@erc) Balakrishna Nakshatrala (balki@erc)

 This project deals with the reconstruction of  signals from irregularly sampled data. The motivation for this project is the bigger problem of feature detection in unstructured grids. This feature detection is proposed to be used  for the detection of shocks in the flow field solved on unstructured grids. This will help in evolving strategies for localizing the region of the grid that needs adapting. Though this problem is in 2-D space, our project will deal with the reconstruction of 1-D signals from irregularly sampled data.
	                  The project will involve the implementation and comparision of three algorithms for the reconstruction, namely....
   1) 	 The projection onto convex  sets (POCS) method
   2)  The adaptive weights method (ADPW)
   3)   The Voronoi method or the Frame (WILMAR)  method (Wiley / Marvasti)	                                              
 The criteria for comparision will be convergence, error estimates, computational expense etc.
 Assumptions:
          The main assumption these algoritms make is that the signal is band limited.
Time table for  progress:
	   Sept 15th            :   Completion of literature survey and final detailed proposal.
   October  31st     :   Implementation of all three algorithms by this date.
   November 15th  :   Debugging and testing of code with real data.