Information processing by a controlled coupling process

This Letter proposes a controlled coupling process for information processing. The net effect of conventional coupling is isolated from the dynamical system and is analyzed in depth. The stability of the process is studied. We show that the proposed process can locally minimize the smoothness and the fidelity of dynamical data. A digital filter expression of the controlled coupling process is derived and the connection is made to the Hanning filter. The utility and robustness of proposed approach is demonstrated by both the restoration of the contaminated solution of the nonlinear Schr\"{o}dinger equation and the estimation of the trend of a time series.