Discrete Interval Binary Signal Optimization Using PSO

Título: Discrete Interval Binary Signal Optimization Using PSO

Autores: Fernandez Santos, Sergio;Parpinelli, Rafael

Discrete Interval Binary Signals (DIBS) is a well-known multi-frequency two-level amplitude signal used in system identification and commonly occurs harmonic distortion in the frequency spectrum as a drawback. Hence, an optimization must be done to improve the relation between signal and undesired harmonic distortion to improve the Crest Factor (CR). This paper proposes the use of Particle Swarm Optimization (PSO) to optimize DIBS. The employed PSO is the canonical codified in the continuous domain. Eight different experiments with different parameters of DIBS are proposed to validate this PSO application. PSO performance is compared with the Van Den Bos algorithm widely used in the literature. Results show better performance for PSO compared to the Van Den Bos algorithm passing the Wilcoxon rank sum test with a very small p-value. Another remarkable fact is that PSO shows good convergence to all problem instances, giving robustness to the solution.

System identification;PSO,Signal Optimization;Swarm Intelligence

Páginas: 12

Código DOI: 10.21528/CBIC2017-34

Artigo em pdf: cbic-paper-34.pdf

Arquivo BibTeX: cbic-paper-34.bib