Reconstrução de energia em calorímetros operando em alta luminosidade usando estimadores de máxima verossimilhança

Abstract

This paper presents signal processing techniques that performs signal detection and energy estimation using calorimetry high energies. The CERN, one of the most important physics particles research center, has the LHC, that contains the ATLAS. The TileCal, important device of the ATLAS calorimeter, is the component that involves a lot of parallel channels working, involving high event rates. The reconstruction of the signal energy that interact with this calorimeter is performed through estimation of the amplitude of signal generated by this calorimter. So, accurate noise modeling is important to develop efficient estimation techniques. With high brightness in TileCal, the noise model modifies, which leads a performance drop of estimation techniques used previously. Modelling this new noise as a lognormal distribution allows the development of a new estimation technique using the MLE (Maximum Like lihood Estimation), improving parameter sestimation and leading to a more accurately reconstruction of the signal energy. A new method to analise the estimation quality is presented, wich is very effective and useful in high brightness enviroment conditions. The comparison between the method used by CERN and the new method developed revealed that the proposed solution is superior and is suitable to use in this kind of ambient that TileCal will be working from 2018.