Robust control and state observer design for neural mass model applications using simulated EEG signals

Andrei Popescu, Catalin Buiu

Abstract


The paper presents, on the one hand, the design of a robust control method using H? tools applied to a nonlinear neural mass model of a cortical column using EEG recordingsas signal measurements. The objective of the control problem is to suppress the neuronalactivity of the cortical column by ensuring guaranteed performance specifications as well asrobustness against model uncertainties and measurement noise. On the other hand, to monitorthe hidden, unmeasured, activity of a cortical column an Extended Kalman Filter is designedbased on the neural mass model of the macrocolumn and EEG measurements of its activity. Thecapabilities of these methods are tested, in simulation, using the neural mass model descriptionof a cortical column for an epileptic seizure. Both, the robust controller and the state observer,show promising results in simulation.

DOI: 10.61416/ceai.v25i4.8846


Keywords


neural mass model; convolution-based model; robust control problem; H? tools; estimation problem; extended Kalman filter; nonlinear observer application; EEG recordings; epileptic seizure model

Full Text: PDF