In this study, a damage detection approach using traffic-induced vibration data of the bridge is
developed, employing only direct analyses by means of introducing soft computing methods. Different from
identifying the structural damages using inverse analyses, the possible damage patterns of the bridge are assumed
according to theoretical and empirical considerations at first. Then, the running vehicle-induced dynamic
responses of the bridge under a certain damage pattern are calculated employing a developed vehiclebridge
interaction analysis procedure. When the calculated response is identical to the recorded one, this damage
pattern will be the solution. However, owing to the large number of damage patterns, it is difficult to
identify the exact solution. Therefore in this approach, the optimization method of Genetic Algorithm is applied
to identify the damage pattern. The feasibility investigation of the proposed approach is investigated using
numerical models in this paper.