Parameter estimation and filtering for chaotic time series by steepest descent

This paper presents a new method for simultaneous filtering of and parameter estimation for chaotic time series corrupted by additive white Gaussian noise. In particular it is assumed that the underlying model of the nonlinear time series is known, but the corresponding parameters are not. The new method treats the problem of parameter estimation and filtering for chaotic time series as a nonlinear minimization process and solves it using a steepest descent algorithm. A chaotic time series generated by computer is applied to verify the performance of the new method. And compared with the existing method, the numerical simulation experiments show that the new method is better than the existing method