Abstract
COVID-19 pandemic response with non-pharmaceutical interventions is an intrinsic control problem. Governments balance social distancing policies to avoid overload on health system without major economic impact. A control strategy requires reliable predictions to be efficient on long-term. SARS-CoV-2 mutability, vaccination coverage and time-varying restrictive measures change virus evolution dynamics frequently. State and parameter estimations are an option do deal with these uncertainties. In this paper, a SIR-based model is proposed considering data available and feedback corrections over time. State and parameter estimations were done on state estimators with augmented states. Three observers were implemented: Constrained Extended Kalman Filter (CEKF), CEKF and Smoother (CEKF&S) and Moving Horizon Estimator (MHE). The parameters estimated therein are based on vaccine efficacy studies regarding transmissibility, severeness of disease and lethality. Social distancing is a measured disturbance calculated with Google mobility data. Six federative units from Brazil are used to evaluate proposed strategy: Amazonas, Mato Grosso do Sul, Rio Grande do Norte, Rio Grande do Sul, Rio de Janeiro and São Paulo. State and parameter estimations were realized from October 1 st 2020 to July 1 st 2021 during which Zeta and Gamma variants emerged. Results showed an efficient detection of circulating variants from proposed parameter estimation. In addition, it asserted dynamics related to virus mutations. Zeta mutations increase lethality between 19 and 45%, and increased transmissibility between 20 and 38%. Gamma mutations, on the other hand, increased lethality between 62 and 110% while increasing transmissibility between 52 and 107%. Furthermore, parameter estimation indicated existence and temporal change of subnotification on hospitalized and deceased individuals. Overall, dynamics estimated were within expectations and are applicable to control theory.
Iterative High-Accuracy Parameter Estimation of Uncooperative OFDM-LFM Radar Signals Based on FrFT and Fractional Autocorrelation Interpolation
