 |
General: 663 |
P3: Memory Efficient Signal Reconstruction from Phaseless Coefficients of a Linear Mapping
Author: Naveed Haghani , Advisor: Radu Balan (MATH)
Problem Statement Presentation
Project Proposal
Abstract
The project proposed in this paper will implement and test an iterative, recursive least squares algorithm described in [5] to perform phaseless reconstruction from the magnitudes of the coefficients of a linear transformation. Testing will be done on synthetically generated input data created using random number generation. A random input vector will be generated and passed through a transformation algorithm. The transformed signal will then be passed to the iterative, recursive least squares algorithm to reconstruct the original signal. Following that, post processing will be done on the results. The implementation will be programmed in MATLAB. The implementation will be designed to prioritize memory efficiency. Memory efficiency, in this regard, applies primarily to the storage of the resulting linear system involved in reconstruction. The linear system will be on the order of 10,000x10,000. Avoiding the costly storage of this system and deriving its contents when needed will be the primary focus during implementation of the algorithm. Following the algorithm’s completion, the program’s performance will be studied with regards to time efficiency, accuracy, and scalability with problem size.
MidYear Progress Report and
Presentation
Final Presentation ,
Final Report
