Full text
| |
| Author(s): |
Guilherme Dean Pelegrina
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Limeira, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Faculdade de Ciências Aplicadas |
| Defense date: | 2017-02-08 |
| Examining board members: |
Leonardo Tomazeli Duarte;
Romis Ribeiro de Faissol Attux;
Guilherme Palermo Coelho
|
| Advisor: | Leonardo Tomazeli Duarte |
| Abstract | |
Several problems in signal processing are formulated as blind source separation problems. Classically, these problems are solved through the optimization of a separation criterion related to the source signals. However, in many practical situations, there is more than one information about the sources and, consequently, more than one separation criterion can be built to solve the problem. Therefore, this work proposes the application of a multi-objective approach, whose resolution is achieved by simultaneous optimization of more than one criterion, to solve blind source separation problems. With the purpose of demonstrating the applicability of this approach, numerical experiments were performed in order to compare the solutions obtained by the multi-objective approach with the solutions optimizing each criterion individually. The results suggest that the multi-objective approach provides solutions that, analyzed by the decision makers involved in the problem, are better than those achieved when only one criterion is taken into account in the model (AU) | |
| FAPESP's process: | 14/27108-9 - Multiobjective optimization method for inverse problems and blind source separation |
| Grantee: | Guilherme Dean Pelegrina |
| Support Opportunities: | Scholarships in Brazil - Master |