Regions of parameter identifiability

A procedure for computing explicit regions in the parameter space in which the Gauss-Newton method will converge to a unique solution is developed. The computed convergence regions are applicable in any control systems applications using the Gauss-Newton method. The uniqueness result is of value by itself because it is independent of the Gauss-Newton procedure and gives explicit regions of parameter identifiability. The main feature is the computational aspect which makes use of interval analysis.