Abstract
This paper is concerned with theoretical and computational aspects of a method for identifying the boundary shape of distributed parameter systems under noisy observations. First, the system model is given by a partial differential equation of parabolic type derived by a known distributed input. Noisy observations are made by sensors allocated on known subregions of the system. Secondly, based on the concept of the maximum likelihood estimate, an estimation algorithm is presented in a form of the recursive computation. Convergence properties of the estimated process are discussed by using the stochastic stability theory. Finally, the validity of the proposed estimation algorithm is shown through results of digital simulation experiments.
