Many methods of damage identification in bridge structures have focused on the use of numerical models, modal parameters or non-destructive damage tests as a means of condition assessment. These techniques can often be very effective but can also suffer from specific pitfalls such as, numerical model calibration issues for non-linear and inelastic behaviour, modal parameter sensitivity to environmental and operational conditions and bridge usage restrictions for non-destructive testing. This paper covers alternative approaches to damage identification of bridge structures using empirical parameters applied to measured vibration response data obtained from two field experiments of progressively damaged bridges subjected to ambient and vehicle-induced excitation, respectively. Numerous non-modal vibration-based damage features are detailed and selected for the assessment of either the ambient or vehicle-induced excitation data based on their inherent properties. The results of the application to two real bridges, one under ambient vibration and the other of forced vibration, demonstrate the robustness of the proposed damage features for damage identification using measurements of ambient and vehicle excitations. Moreover, this investigation has demonstrated that the novel empirical vibration parameters assessed are suitable for damage detection, localisation and quantification. AbbreviationsCAV cumulative absolute velocityCAD cumulative absolute displacementDVI distributed vibration intensityMCVI mean cumulative vibration intensityIVI instantaneous vibration intensityAIVI Amalgamated instantaneous vibration intensityEMD empirical mode decompositionICEEMDAN improved complete ensemble empirical mode decomposition with adaptive noiseHHT Hilbert–Huang TransformIMF intrinsic mode functionsMCD minimum covariance determinateMSD mahalanobis squared distanceMTS Mahalanobis Taguchi system.