Input-output synchronization with non-uniformly and asynchronously sampled output data

If the input and output signals of a dynamic system are recorded independently for the purpose of system identification a correct synchronization between these data sets becomes necessary. In the case of uniformly sampled inputs and non-uniformly sampled outputs this synchronization must be achieved off-line and without knowledge of the absolute sampling instants. A new model-based iterative algorithm is proposed which is structured in two stages. Using linear identification methods a dynamic model is established. The local shift between the predicted output from this model and the measured output is estimated by correlation methods and subsequently removed from the measured output by resampling. The superior performance of the new algorithm compared to pure correlation methods is demonstrated with a simulation study and with measured data from a railway track and the corresponding dynamic vehicle response.