Identification of impulse response from normal operating data using the delay line synthesizer principle

A computational technique is presented here for identifying the impulse response of a linear system from normal operating noisy data. No assumption, however, is made regarding the nature of the noise. The technique derives its idea from the delay line synthesizer (DLS) though in this case the DLS coefficients which discretely represent the weighting function are computed automatically employing the steepest descent method. The method has been tried out on a first-order as well as a second-order system simulated on a digital computer; the estimated impulse response is found to be very close to the actual one.