Calorimeter Response Deconvolution for Energy Estimation in High-Luminosity Conditions

Signal superposition, also called pile-up noise, has become an important issue for calorimeters operating under high-luminosity conditions. Current techniques model those superimposed signals as highly correlated noise, whose covariance matrix information is used to develop a minimal variance and unbiased filter for energy estimation. Hence, the filter coefficients depend on the current luminosity. This paper presents a new energy estimator whose design is independent of the luminosity. The method consists of deconvolving the front-end electronic shaping response, recovering the primary impulse signals. The deconvolution process allows the identification of each superimposed component. Based on that information, it is possible to include additional constraints in order to design orthogonal estimators for each component, providing optimal estimators suitable to operate in colliders where luminosity changes during the run. Comparisons between current approaches and the proposed method using pile-up event simulation are presented. Results point out that an improvement on amplitude estimation performance is achieved and stable signal reconstruction is provided for different luminosity ranges.