Abstract
We study the location of local minima of the finite sample approximation to the constant modulus cost function. This paper concentrates on source separation. The main result is a connection between the number of samples and the probability of obtaining a local minimum of the finite approximation within a given sphere around the local minimum of the CM cost function. The motivations for our study are two problems: equalization of communication signals, and blind separation of a desired signal in multiuser environment. In order to maintain simplicity we focus on the case of blind beamforming which is somewhat simpler to analyze
| Original language | American English |
|---|---|
| Title of host publication | Acoustics, Speech, and Signal Processing, 2000. ICASSP '00. Proceedings. 2000 IEEE International Conference on |
| Publisher | IEEE |
| State | Published - 2000 |
Bibliographical note
Place of conference:IstanbulFingerprint
Dive into the research topics of 'On the finite sample behavior of the constant modulus cost function'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver