Robust failure detection of dynamic system-an operator approximation approach

Robust failure detection methods based on redundancy relation techniques are extended to the general separable Hilbert space using an operator approximation approach. The approach is direct and simple and provides an intuitive geometric interpretation. The results can be applied to the design of the parity relations for infinite-dimensional systems. The computation procedure required is the Schmit expansion of an operator whose counterpart in the finite-dimensional case is singular value decomposition. An example arising from the sensor failure detection design for space structures is given to illustrate the methodology suggested.<
>