Methods of synthesis of the optimal non-divergent filtering algorithm without use of the distribution density
Methods of synthesis of the optimal non-divergent filtering algorithm without use of the distribution density
The non-divergent filtering algorithm is constructed for the filter \(x_{n+1} = Ax_n + K_{n+1}e_{n+1}\) without any knowledge of the noise distribution. This result is obtained by determining \(K\) as minimizing the function \(FP(K)+GV(K)\) where \(P(K)\) is a correlation matrix of the filtration error, \(V(K)\) measures the rate of change of \(P(K)\) and \(F\), \(G\) are weight matrices.
Finite difference methods for boundary value problems involving PDEs, stable filter, non-divergent filtering algorithm, Computational methods for problems pertaining to systems and control theory, Kalman filter, Filtering in stochastic control theory, Inference from stochastic processes and prediction
Finite difference methods for boundary value problems involving PDEs, stable filter, non-divergent filtering algorithm, Computational methods for problems pertaining to systems and control theory, Kalman filter, Filtering in stochastic control theory, Inference from stochastic processes and prediction