Continuous-time gray-box identification of mechanical systems using subspace-based identification methods

Abstract

We consider the problem of gray-box identification of dynamic models for mechanical systems. In particular, the problem is approached by means of continuous-time system identification using subspace-based methods based on discretetime input-output data. A method is developed, with the property that the structure of the model resulting from fundamental physical first principles is obtained and the parameter matrices have a clear physical interpretation. The proposed method is subsequently successfully validated in both simulation and using experimental data from a three-axis manipulator. In both cases the identified models exhibit good fit to the input-output data.The results indicate that the proposed method can be useful in the context of model-based control design in, for example, impedance force control for robots and manipulators, but also for modal analysis of mechanical systems.

A linear algorithm for the minimal realization problem in physical coordinates with a non-invertible output matrix

“…in vibration mechanics it it well known the so-called "inverse vibration problem", see e.g. [15], [16]. In the literature there are also methods that exploit the structure of the matrices in the continuum-time model, because of their physical meaning, to compute directly a matrix T which should transform in physical coordinates the estimated model, see [13] or [14] for recent results and further references; to obtain this it is necessary to reformulate the abstract model equations into a null-space-based problem, that brings to a quite involved solution and has in general an high computational cost or it is restricted to a small number of model structures.…”

Model-based detection of soft faults using the smoothed residual for a control system

“…As there are bound to be some differences or neglection between mechanism models and actual systems, so the mechanism model is inaccurate [27]. In addition, the identification model relied only on measurement data, which may be unreliable [28]. In addition, residual observers under the disturbance of complex noises, such as impulse interference, may cause false alarms.…”