Choice of pseudorandom binary signals for system identification

A simple technique, exploiting a property of pseudorandom binary signals, reduces to zero the effect of linear disturbances in the estimation of a system weighting function by cross-correlation. Errors due to quadratic disturbances are then constant and may be minimised by choice of particular pseudorandom binary signals which are tabulated. The principal remaining error is also constant, and an unbiased estimate of the system weighting function is easily obtained.