Título: Otimização do Sistema Neural de Seleção Online de Eventos num Detector de Partículas através do Processamento Estatístico de Sinais
Autores: Souza, Edmar E. P.; Simas Filho, Eduardo F.; Farias, P. C. M. A.; Seixas, José M.
Resumo: The ATLAS is the largest particle detector of the LHC (Large Hadron Collider). Considering the different ATLAS subsystems, the calorimeter comprises more than 100,000 sensors and is responsible for measuring the energy of the incoming particles. Electron detection is very important to the experiment as these particles are directly related to interesting physical signatures. The identification of electrons heavily relies on calorimeter information, and the background noise, composed of hadronic jets, may occur with frequencies up to 10^5 times higher than the physics of interest, making the identification process a difficult task. In previous studies it has been proposed an alternative electron detection algorithm (Neural Ringer), in which the energy profile measured at the calorimeter is formatted into concentric rings and these signals are used to feed an artificial neural network classifier. This work proposes the use of statistical signal processing techniques added to the Neural Ringer chain, aiming at extracting more discriminant features from the preprocessed calorimeter information. Through the proposed technique it is possible to both, increase the efficiency of Neural Ringer discriminator and reduce the amount of stored information in 60%. Results with simulated data are used to show the benefits of the proposed method.
Código DOI: 10.21528/CBIC2013-133
Artigo em pdf: bricsccicbic2013_submission_133.pdf
Arquivo BibTex: bricsccicbic2013_submission_133.bib