ITERATED CONSTRAINED STATE ESTIMATOR FOR NONLINEAR DISCRETE-TIME SYSTEMS WITH UNCERTAIN PARAMETERS

In this paper, an estimator is developed to estimate the states of nonlinear stochastic discrete-time dynamical systems with uncertain parameters. The system model and the measurements are assumed to be corrupted by uncorrelated zero mean white Gaussian noise sequences. The parameters of the system are assumed to be uncertain. The proposed approach is based on the extended Kalman filter and the active set method, in which multiple projection approach is used to get the dynamics of the proposed estimator. Although the developed state estimator uses the nominal values of the system parameters, it shows to be more stable when compared with other existing techniques and gives satisfactory results. To illustrate the effectiveness and simplicity of the developed approach, an illustrative example is presented. Simulation results show that the developed technique leads to a stable state estimator even in many cases in which the extended Kaman filter and the extended Kalman filter with parameter estimation diverge.