The Neal-Smith (Neal, T. P., and Smith, R. E., An In-Flight Investigation to Develop Control System Design Criteria for Fighter Airplanes, Air Force Flight Dynamics Lab., AFFDL-TR-70-74, Vols. 1 and 2,1970) criteria for assessing handling qualities of highly augmented aircraft are rewritten in terms of HOC-norms to arrive at a mixed-sensitivity performance index. The pitch attitude control is formulated as a mixed-sensitivity problem, which is solved by the model matching technique in the framework of the HOC-control theory to yield H^ -optimal pilot models. It is shown that the obtained H^ -optimal complementary sensitivity function meets the Neal-Smith criteria so perfectly that the task difficulty is exhibited solely by the HOC-optimal pilot models. Surveying all of the flight configurations of the Neal-Smith flight test, a method using the maximum gain gradient and the phase at a particular frequency of the HOC-optimal pilot model is proposed to correlate pilot ratings with the pilot compensation efforts. ANDLING qualities specifications for highly augmented air- craft can no longer be based on their basic dynamics, but rely heavily on closed-loop analyses of pilot-aircraft systems. Attempts at correlating pilot ratings with the quantities resulting from the closed-loop analyses have yielded important relationships between pilot ratings and pilot compensation efforts, culminating in the well- known Neal-Smith criteria.1 The criteria state that pilot rating is primarily a function of the pilot's compensation required to achieve good low-frequency performance and the closed-loop oscillatory tendencies that result. In the original work by Neal and Smith,1 the pilot compensation required to meet the criteria is found on the Nichols chart by adjusting the gain and the lead-lag time constants of a simple pilot transfer function model. Thus, the determination of the pilot-related parameters depends on the researcher's graphical skill and insight into the system. Despite all of the efforts to en- hance usefulness of Neal and Smith's work,24 the works based on assumed pilot transfer functions have incurred difficulties in finding the pilot compensation uniquely. One way to cope with the difficulties is to employ the optimal control approach to modeling the pilot loop closures.5 In this ap- proach, a stochastic optimal control problem with a performance index of the quadratic form is solved to yield an optimal control model (OCM) of the pilot including a Kalman filter as the state esti- mator. The OCM in turn may lead to the unique determination of the pilot compensation. It has been shown that with proper modeling the magnitude of the model's cost function correlates with subjective rating in a variety of tasks,6 and the optimal pilot compensation and the optimal closed-loop parameters may duplicate Neal and Smith's results.7 Although past works try to reflect the Neal-Smith criteria in their resulting OCMs, the performance index itself is not a mani- festation of the criteria. Furthermore, the robustness implied by the criteria is not well addressed in the conventional OCM approach. Because the Neal-Smith criteria describe the standard of perfor- mance that the pilot is trying to achieve for a required tracking task, a performance index can be formed out of the criteria. Drawing on these past works, an attempt was made at rewriting the Neal-Smith criteria in terms of H^ norms to yield a mixed- sensitivity problem.8 The problem was solved in the framework