Nonlinear stochastic filtering using global transformations

Abstract

A new approach for solving the nonlinear filtering problem is introduced. This approach is based on the concept of exact linearization using global transformation techniques. Using this concept, a nonlinear stochastic system can be transformed, using a suitable mapping, to an equivalent linear stochastic system for which well developed linear filtering techniques can be applied. Sufficient conditions for the existence of such a mapping are derived and the benefits of the work are illustrated via an example.