Crash Pulse Prediction Via Inverse Filtering

This paper reports a study on the use of response inverse filtering (RIF) methodology for crash pulse prediction. RIF is based on the finite impulse response (FIR) and inverse filtering (IF) methods. The FIR coefficients obtained by the digital convolution theory and the least squared error approach serve to transfer response from the input (impacting or excitation) side to the output (non-impacting or receiving) side.

The FIR method, a process of low pass filtering (e.g. truck body mount), is commonly used in predicting the non-impacting side (e.g. truck body or cab) response with the input excitation in the impacting side (e.g. truck frame). The accuracy in the validation and prediction via FIR transfer function depends on the frequency contents of the input and output accelerometer data from which the transfer function is developed. The prediction accuracy is low if the output data contain higher frequency components than the input. Taking advantage of this FIR forward prediction accuracy, the method of inverse filtering is thus utilized to develop the RIF for the backward prediction.

The new RIF transfer function is created by the IF matrix operations applied to the FIR forward transfer function. The accuracy of RIF in predicting the high frequency output (such as frame impulse) with the low frequency input (such as body excitation) has been shown to be high. In crash development of a vehicle, this technique is useful in crash pulse prediction. One of the applications in predicting the truck frame pulse based on a desired body pulse is described in details.