A procedure for generating locally identifiable reparameterisations of unidentifiable non-linear systems by the similarity transformation approach

A method is presented for the generation of locally identifiable reparameterisations of non-linear systems which have been shown to be unidentifiable via application of the similarity transformation approach. The existence of the reparameterised system in terms of the maximum permissible number of locally identifiable parameters is provided and is crucially dependent upon the ability to find the rank deficiency of an appropriate (and possibly infinite) jacobian matrix. The reparameterisation procedure is described in detail, and is illustrated with application to two known non-trivial examples of unidentifiable non-linear systems. (C) 1998 Elsevier Science Inc. All rights reserved.