Block algorithms for state estimation and functional observers

This paper presents a new block algorithm for the Sylvester-observer equation, which seems to be computationally more efficient than the Van Dooren's algorithm and is suitable for high performance computing using LAPACK. It also presents a new algorithm for the design of functional observer in which the vector Kx(t), required in the implementation of the feedback law u(t)=Kx(t), can be directly estimated without estimating the vector x(t). Numerical examples are provided to illustrate the algorithms and results on comparison with the well-known Van Dooren's algorithm for reduced-order state estimation are illustrated using graphs.