Heterogeneous Data Fusion for Traffic-induced Excitation Identification of Truss Bridges

We present a time domain Bayesian inference-based regularization approach for the identification of traffic-induced excitations of truss bridges using heterogeneous data fusion. The acceleration, strain and displacement measurements are fused and rescaled by a state space model for force identification. The unknown excitation time histories are solved by a regularization approach based on Bayesian inference. A smoothing operator is used for signal de-noising. Finally, the proposed algorithm is numerically illustrated by a truss bridge. Results demonstrate the effectiveness of the proposed algorithm for traffic-induced excitation identification with a high accuracy. doi: 10.12783/SHM2015/47