Several techniques are currently in use for system identification, each having its own level of performance. In this paper we present a coded interrogation approach for statistical system identification; it is based around a modified Golay sequence applied to an adaptive finite impulse response (AFIR) filter system. We show that this technique circumvents some of the shortfalls associated with other comparative approaches. The theory of the proposed technique is developed, and the operating principle of a prototype model to verify it is described. Analysis of the results obtained from experiments, performed on a simple first-order linear time invariant (LTI) system, gave an average deviation of ±2.94% from the ideal response.