Dynamic grey clustering adaptive H_∞ filtering algorithm

To solve the problem that parameters of H_∞ filter increase with time, a new adaptive H_∞ filter algorithm is proposed. Based on real-time estimation of noise matrix and grey clustering of state variable, filter or gain matrix is modulated so as to get an estimation of the state vector. Simulation showed that, using the new algorithm, compared with the H_∞ or Kalman filter, expected precision could be achieved and parameters could be made stable in 10 s. The improved algorithm can avoid spreading of computation and has excellent robustness.