Neural inverse optimal control applied to design therapeutic options for patients with COVID-19

ABSTRACT

In this paper we apply an inverse optimal controller (IOC) based on a control Lyapunov function (CLF) to schedule theoretical therapies for the novel coronavirus disease (COVID-19). This controller can represent the viral dynamics of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) in the host. The virus dynamics consider the antiviral effects and immune responses as control inputs. The proposed controller is based on a Recurrent High Order Neural Network (RHONN) used as an identifier trained with Extended Kalman Filter (EKF). Simulations show that applying treatment 2 days post symptoms would not significantly alter the viral load. The proposed controller to stimulate the immune response displays a better effectiveness compared to the effectiveness displayed by the antiviral effects.