JORDAN Mario Alberto
congresos y reuniones científicas
A differential Geometric Approach for Parametric Nonlinear Estimation
JORDAN, MARIO; BAMBILL, H.
San Diego, USA
Congreso; IEEE American Control Conference (ACC); 1999
In this article a new approach for nonlinear parameter estimation is presented. It is based on differential geometric properties of trajectories and objective functions in an extended space composed by parameters and time. For the analysis concepts of straight and folded slots of scalar surfaces are introduced. A multilinear model in this space is applied for capturing possible local minima of the objective function. It is shown that certain input-dependent conditions for the forced trajectory must be satisfied for global identifiability of the model parameters. Finally tensor properties of the multilinear model are combined with a search algorithm to conform the parameter adaptive law. Examples of application show advantages over other standard adaptive laws.