PARAMETRIC-FEEL ALGORITHM: DEVELOPING A PARAMETRIC VECTORFITTING MODEL FOR EVENT LOCALIZATION IN CALIBRATED STRUCTURES

Document Type

Conference Proceeding

Publication Date

11-27-2023

Department

Department of Mechanical Engineering-Engineering Mechanics

Abstract

For smart structures, especially in the context of human activity, the force exerted and the location it happened is of significant relevance. This paper revisits and improves the performance in localizing and characterizing an input force with precalibrated structures through vibration measurement. The Force Estimation and Event Localization (FEEL) Algorithm have been discussed as a means of calculating the force of an impact and pinpointing its location. Unlike other time-of-flight approaches, FEEL does not require time synchronization, instead using transfer functions between possible impact locations and sensor locations to estimate force and localize impact. However, this approach is limited to locations where transfer functions are available. To overcome this limitation, a rowing hammer test was used to determine Frequency Response Functions (FRFs) at various points on a beam with a uniform rectangular cross-section. The Vector-Fitting algorithm was then used to improve the FRF approximation by moving poles to more advantageous locations, enhancing convergence, and lowering noise. Using the curve fitting approach, residues and FRFs were interpolated for additional locations. The extended FEEL algorithm was then used to localize impacts and estimate forces at these additional locations. This method can be used in applications such as tracking customer movement in retail establishments, detecting falls, tracking rehabilitation progress, and estimating building occupancy.

Publication Title

Proceedings of ASME 2023 Conference on Smart Materials, Adaptive Structures and Intelligent Systems, SMASIS 2023

ISBN

9780791887523

Link to Full Text

Share

COinS