A comparison of time and frequency domain-based approaches to laser Doppler vibrometer instrument vibration correction

When laser Doppler vibrometers are used in the presence of ambient vibration, it is essential to compensate for the additional vibration signal content. In practice, compensation is realised by independently determining the instrument vibration and subtracting it from the erroneous measurement. When these vibrations are transient in nature, time domain-based processing should be used to carry out the correction. However, recent implementation of such an approach on stationary signals showed a factor of eight increase in performance over the previously established frequency domain-based alternative. Therefore, the work described in this paper initially focuses on determining the cause of the inconsistency and proposes a revised frequency domain approach. This revised approach offers near-equivalent performance to its time domain-based equivalent, with the latter approach offering only a factor of 0.26 increase in performance. However, despite the advantages of selecting the time domain-based technique, it typically requires high oversampling factors to allow for the accurate synchronisation of the various transducer type signals. Up until now, the only method available to determine the relationship between the sampling frequency and the performance would be experimentally, which is laborious and time consuming. Therefore, the significance of this paper is the development and experimental validation of an analytical model which predicts the sampling frequency dependence of the time domain correction technique performance. Using this, a framework was developed which allows for the optimal implementation of either correction technique and specifies the required acquisition parameters.