Synthesis and evaluation of an H2 control law for a hovering helicopter

The design and simulator evaluation of a rate-command flight-control law for a UH-60 helicopter in near-hover flight conditions are described. The multivariable control law was synthesized using an //2 method, which, through weighting functions, directly shapes the singular values of the sensitivity (/ + GK)~l, comple- mentary sensitivity GK(I + GK)~ l, and control input K(I + GK)~l transfer-function matrices. The design was implemented on the vertical motion simulator, and four low-speed hover tasks were used to evaluate the control system characteristics. The pilot comments indicated good decoupling and quick response characteristics, but also revealed a mild pilot-induced oscillation tendency in the roll axis. HE purpose of the research documented in this paper is to investigate the application of an H2 multivariable de- sign method to synthesize flight-control laws for helicopters. The objective of the work is to determine through analysis and piloted simulation the strengths and weaknesses of the H2 method within the context of a modern rotorcraft design specification. The choice of a multivariable design method is natural to the problem presented because of the highly coupled nature of helicopter dynamics and the desired level of feedback dictated by modern design specifications. Helicopter flight-control laws designed through classical single-input, single-output techniques are satisfactory when requirements do not infringe on the frequency range of significant interaxis cross coupling due to the rotor. Previous generation helicopters and their design requirements have complied with this assumption. However, future helicopters will use hingeless and bearingless rotors with more complex coupling, and these helicopters will have control-law design requirements demanding higher levels of feedback. To ensure that the design criteria are sufficiently stringent to represent future requirements, the control law here was de- signed according to the Section 3 hover requirements in the Aeronautical Design Standard (ADS-33C).1 Previous work on multivariable techniques for rotorcraft flight-control design2'4 complied with part or none of the ADS-33C requirements, making these works difficult to interpret in terms of this new standard. These previous works also used a low-order model (usually a 6-degree-of-freedom model) to design the flight-control law, and the final design was usually checked on a higher-order model in analysis or simulation. These higher-order models did not fully represent the rotor-fuselage dynamics, lacking either the coupled nature of the rotor dynamics (between the fuselage and the rotor and/or between the rotor degrees of freedom) or lacking the rotor lead-lag degree of freedom. These effects are known limiting factors in the feedback gains of helicopter flight-control laws and are represented in the analysis and piloted simulator model here. The importance of these effects was illustrated in advanced digital optical control system (ADOCS) research, where feedback gain reductions of 40% were necessary to avoid lead-lag instabilities.5 This paper continues with a description of the models used in the design process and piloted simulation, followed by a description of the method used to synthesize the flight-control law. The salient features of a rate-command control-law de- sign are then discussed, followed by descriptions of the simu- lation environment and task arrangement. Finally, the results of the simulation are discussed and conclusions drawn. More details on the design method, design models, and flight-con- trol design can be found in Ref. 6, whereas more details related to the simulation experiment can be located in Ref. 7.