Extending H_INF identification to unstable plants

System identification complements not only the robustness requirement of maintaining control in the face of plant uncertainty but thrives on the symbiotic relationship with it. This relationship exists as an iterative process of identification and control design while the identification requirement derives it specification from the closed loop performance with the previous controller and delivers an error bound that dictates the design of the successive controller. This paper applies and extends techniques in the emerging field of H_INF identification to the problem of identifying a real plant as well as estimating the called a priori information necessary to determine worst case error bounds.