Particle Filter Method (PFM) based on Bayesian theory gives a reliable estimate of hidden parameters from the measured signal. A suitable mathematical model of physical system has to be considered to obtain forward solution in the identification method. The present paper compares the efficiency of different models adopted in PFM for identification of moving mass on a bridge. A field test under moving truck has been conducted on an existing pre-stressed concrete bridge to collect response data at different locations. The field test data has been utilized in PFM using three models- (i) moving mass-beam (ii) sprung mass-beam coupled and (iii) Finite Element Model with moving load. The study reveals that sprung mass-beam coupled model including roughness induced dynamic force gives better estimate than the other two models considered in the study.

Keywords: Hidden parameters, Identification, Mathematical model, Moving mass, Flexural oscillation.