The “identification of systems” is a generalized form of curve fitting pertaining to systems for which a mathematical model is known, and for which input/output data is empirically available, but for which actual values of parameters in the model are unknown and are sought.
A technique for identification (that is, determination of parameters) in second-order, dynamic systems is presented and applied to a typical system; namely, a two-axle rubber-tired vehicle.
The scheme is based on establishing a set of system model equations and their use with system response data to define one or more residue functions in a manner analogous to the defining of an error function in curve-fitting by the collocation method. A performance index is defined by treating the residue as a measure of least squares fit, and the parameters are then determined by finding the set of values which simultaneously makes all the first partial derivatives of this index with respect to the parameters vanish.
The technique is applied to the determination of nine system parameters in a nonlinear, two-degree-of-freedom model of a vehicle/tire suspension system. It is found to yield the correct values to within 5%.