Distributed Parameter Identification Using the Method of Characteristics

A method is presented for estimating parameters in distributed systems which can be analyzed by the method of characteristics. Both the very real problems of noisy measurements and limited available measurement transducers are discussed in some depth. Convergent algorithms are developed using a Robbins-Munro stochastic approximation scheme. The extension of these methods to the identification of function multidimensional systems, and diffusion terms is considered.