The lattice algorithm for estimating parameters of an input/output model based on the delta operator is presented. The motivation is twofold. First, the derivation would allow the lattice implementation of the least-squares method, with all of the advantages lattices have with respect to numerical stability and conditioning. Second, the order recursive property of the lattices will provide the flexibility to change the order of the filter, if the need arises. An example is provided to illustrate the ability of the delta operator-based lattices to identify the parameters of a plant in the presence of measurement noise.
展开▼
