Document
NEW DEVELOPMENTS IN FLYING QUALITIES CRITERIA WITH APPLICATION TO ROTARY WING AIRCRAFT Roger H. Hoh Principal Research Engineer Systems Technology, Inc.
Hawthorne, California Abstract precise point on the runway.
In a paper presented to the AGARD Flight Mechanics Panel in 1981 the Some recent considerations and developments in authors of Ref. 4 cited a case where a pilot gave handling quality criteria are reviewed with empha- a surprisingly good rating to what should have been sis on using fixed wing experience gained in devel- a particularly poor configuration. However, the oping MIL-F-8785C and the more recent Mil Standard landings were not in the prescribed touchdown area and Handbook. Particular emphasis is placed on the and the author (who was also the safety pilot) tasks and environmental conditions used to. develop insisted that the evaluation pilot improve his the criterion boundaries, SAS failures, and poten- performance. On the very next run, in an attempt' tial fixed wing criteria that are applicable to to achieve the required precision, a severe PI0 was rotary wing aircraft.
encountered near touchdown. Needless to say, the Introduction Historically, the handling qualities of rotary wing aircraft have been vastly inferior to their Level 2 Mil-F-8785C For example, the pitch fixed wing counterparts.
//,,,,//,/,11/,,,/1/111/1~ attitude control of many operational helicopters Level 3 Mil-F-8785C - will not even meet the Level 3 requirements of MIL- F-878X. (Level 3 is defined as a Cooper-Harper rating of worse than 6-l/2 or "Flying qualities such that the airplane can be controlled safely but pilot workload is excessive or mission effective- ness is inadequate or both."). An example is 1 where it is shown that the illustrated in Fig.
time to double amplitude for several operational helicopters is in the extreme Level 3 region. The major deficiencies of rotary wing aircraft are nearly always associated with: excessive cross- axis coupling; inadequate dynamic stability; and unacceptable stick force gradients. Interestingly, the Cooper-Harper pilot ratings from many heli- copter handling quality studies (for example, Refs. 1 and 2) indicate that rotary wing pilots are AIrspeed ikt) willing to accept much less than their fixed wing This is shown in Fig. 2 where pilot counterparts.
Fig. 1 Illustration that conventional unaugmented ratings of 2 to 3-l/2 are found well into the helicopters fall well below fixed wing Level 2 region defined for pitch control in MIL-F- standards even for a failed SAS 8785C. (Level 2 corresponds to pilot ratings of (data from Ref. 3) 3-l/2 to 6-l/2 in MIL-F-8785C.) This is felt to occur for two reasons: 1) helicopter pilots are trained to cope with, and expect as "normal," severe instabilities and cross-axis coupling; and 2) the tasks used in the evaluations were not sufficiently demanding.
Consideration of Handling Note : to meet Level2 ffying Ouality Evaluation Tasks quo/i,ies in hfd - f -878X, I>0 In recent years the task used in experiments 0 to obtain handling quality pilot ratings has been 0 found to have a profound effect on the results.
For example, in the landing approach experiments of 0 0 0 I- 1 I I I 0, Ref. 4 the pilots were required to touch down at a -.3 :2 :I 0 .I .2 .3 4 .5 “Phugoid” Damping Ratio, 5 Fig. 2 Cooper-Harper pilot ratings vs. damping ratio in hover; w < 0.5 rad/sec (data from Ref. 1) tified in terms of a scale as shown in Fig. 3a.
evaluation pilot revised his rating downward con- Certain specific closed-loop considerations, which siderably. The point here is that only by insist- were considered in formulating the scale, are ing on a precision task was the experimenter able summarized below and by the generic closed-loop to expose deficient handling qualities that would have otherwise gone unnoticed. In using existing structure in Fig. 3b.
1) A requirement for closure of the attitude data to develop boundaries for the helicopter loop implies VKC conditions and must prevail for handling qualities specification, we must criti- Some suggested evaluation adequate control.
cally evaluate the task.
2) If the equivalent system dynamics require factors might be: 1) Does the task require the same precision closure of position and position rate, but not minimum set of operating conditions as required by operational missions? attitude, a quantified as OVC = 3 is defined.
2) Does the task require the same degree of 3) OVC = 4 quantifies the operating condition aggressive maneuvering as the proposed operational missions? where velocity and attitude cues are not available; 3) Are the tasks well defined, or does the that is, only the outer loop in Fig. 3b can be closed by the pilot.
task encompass a series of subtasks such as an 4) OVC = 5 indicates that no outside visual entire approach, hover, and vertical descent? If can we identify what subtask cues are available.
the latter is true, has the most impact?
Pilot workload can also be reduced via im- 4) Are the data being used as a compromise proved displays. Recent work in the control/ because no better data are available?
display tradeoff area includes the Calspan X-22 5) Are the atmospheric disturbances of low flight tests (Ref. 8) and the CH-46 variable- enough frequency and large enough magnitude to stability helicopter (Ref. 9).
displace the aircraft from its path?
Based on the above considerations, the re- 6) Are the available outside visual cues quired level of augmentation and cockpit displays consistent with the proposed mission?
Unfortunately, these factors may well elimi- can be related to the visibility levels associated nate most existing data. The last factor was found with the missions defined for the helicopter. An initial attempt was made to establish a format for to be especially important for low speed and hover specifying the augmentation and displays required in Refs. 5 and 6 and is briefly reviewed in the following section. for various levels of outside visual cues in Refs. 1. 4, and 5 and is repeated in Table 1 for _ .
convenience.
Effect of Outside Visual Cues on Required Level of Augmentation and Display Most of the available data for low-speed and criteria have been obtained with hover handling good visual outside references and with no require- ment for unattended operation. The real-life existence of secondary tasks, and intermittent to total loss of visual references, places increased demands on the pilot -- an effect which is not For example, pilot discernible from such data.
ratings for an unaugmented helicopter (Ref. 2) and a highly augmented translational rate command (TRC) system (Ref. 7) all fall within the acceptable region (pilot rating better than 3.5). This result is a consequence of experimental scenarios that tend to be tailored toward the systems being inves- al Ouonfification of Oufside Visual Cues /OVCl That is, with pure rate systems the tigated.
scenario is usually benign, thereby usually allow- ing intense, full-time attention; whereas with a translational rate command system the task tends to Pilot The most critical contributor Posttim be more demanding. Perceived Position cues, x to be the to the total pilot workload appears quality of out-the-window cues for detecting air- craft attitudes, and, to a lesser extent, position and velocity. Currently, these cues are cate- gorized in a very gross way by designating the A more discrimi- environment as either VMC or INC.
nating approach is to classify visibility in terms Requires OVC 5 2 (VMC) of the detailed attitude and position cues avail- able during the experiment (or proposed mission), Requires OVC 5 3 and to associate handling qualities requirements Reauires OVC C 4 with these finer-grained classifications.
61 Required Outside Visual Cues for Confro/ The need for certain specific outside visual cues has been inferred from closed-loop considera- tions. These OVC levels have been logically quan- Fig 3 Development of outside visual cue scale Table 1. Augmentation and displays required for various levels of outside visual cues Pilot display Integrated display- MIL-F-8785C Llight director plus flying Mechanical aircraft velocity Augmentation quality level flight director information Rate Level 1 2 3 Level 2 5 Rate command/ Level 1 3 3 attitude hold Level 2 5 5 Attitude Level 1 3 (response feedback) Level 2 5 Attitude Level 1 4 (model following) Level 2 5 Level 1 5 Translational rate with attitude Level 2 5 Translational rate Level 1 5 with direct force control Level 2 5 rotary wing aircraft are reviewed in the following SAS Failures paragraphs.
The concept of "Levels" is used in MIL-F-878X to specify the allowable degradation in handling Lower-Order Equivalent Systems qualities in the presence of failures.
The speci- fication of Level 2 and 3 handling qualities will The basic intent of lower-order equivalent tend to be more critical in rotary wing aircraft in systems is to define a very high-order system in terms of driving the cost and complexity of the terms of a few variables that describe the funda- SAS. characteristics important to the This is a result of the relatively poor mental response handling qualities of the unaugmented helicopter (see Ref. IO). This can be done in the pilot and hence the large change in dynamics before and time domain or in the frequency domain, although after a failure of the SAS. This is illustrated in all work done to date has been in the frequency Fig. 4, which shows a dramatic shift in the charac- domain. Equivalent systems are a viable concept for teristic modes after a SAS failure in the CH-53D. defining Level 1 flying qualities for helicopters.
Clearly, the specification of Level 2 handling qualities that are better than most unaugmented helicopters would have significant implications on SAS complexity and cost.
On c-l Potential Fixed Wing Criteria Applicable to Rotary Wing Aircraft The mission requirements for rotary wing air- craft have become increasingly severe to the point where marginal handling qualities can no longer be tolerated. In most cases satisfactory inherent stability and coupling cannot be obtained without some level of stability augmentation.
Indeed, many modern helicopters employ a stability augmentation SAS system. It is therefore not unreasonable to expect Off the same quality of response (to control inputs and turbulence) in helicopters that is currently enjoyed by fixed wing pilots. In fact, the rapid and precise maneuvering required in some NOE mis- sions may make it necessary to impose _more strin- gent requirements than are necessary for fixed wing aircraft.
Fig 4 Effect of SAS failure on key The applicability of some requirements cur- response modes of CR-53D rently proposed for the fixed wing MU-Standard to (Data from Ref. 3) From a pilot's point of view, a high-bandwidth However, the complexity of the responses of unaug- response would be described as "crisp" or perhaps due to inter-axis coupling, mented helicopters, "rapid and well damped." Typical commentary for a makes it unlikely that useful equivalent system low-bandwidth response might be "sluggish response forms of sufficient generality can be defined for to control input" or "tends to wallow." There the Level 2 and 3 boundaries.
is a long history of correlating such commentary with basic aircraft stability derivatives and/ Bandwidth Criterion mete s made up of such derivatives (e.g., ; 'e'&Mqr- M, , etc.). The term bandwidth comes The bandwidth criterion was developed origi- m%e naturally into play when feedbacks and cross- nally for fixed wing aircraft with direct force feeds are combined to produce aircraft responses control. Because of the almost infinite variety that are unconventional in that the classical modes due to inter-axis of responses that can occur are no longer appropriate definitions.
it was difficult to define a lower-order coupling, equivalent system form for aircraft with direct The advantage of this approach is that it does In looking for an alternative force control.
not assume a particular form of response.
Hence it solution it was hypothesized that the coupling may be suited for helicopters, where coupling tends itself was incidental, and mattered only to the to mask the classical response forms.
The defi- extent that it interfered with the pilot's ability ciency of the bandwidth criterion in its present to adequately perform tight closed-loop tracking.
form is that it does not directly account for the This of course is directly related to the band- "The band- pilot's ability to supply crossfeeds to counteract width, which was defined in Ref. 11: coupling. It seems intuitively obvious that re- width of the specified response to a particular sponses requiring only a control input is defined as the lowest frequency simple crossfeed (such as pure gain) would be more acceptable than those for which the (open-loop) phase margin is at least requiring complex shaping.
45 deg and the gain margin is at least 6 dB." (See This concept was inves- tigated in Refs. 13 and 14 for the turn coordina- Fig. 5 for a graphical description.)
tion problem in fixed wing aircraft and is reviewed briefly in the following section.
The Ref. 11 variable-stability in-flight simu- lation results indicated that the Bandwidth Hypo- Inter-Axis Coupling thesis was indeed valid, i.e., the coupling itself mattered only to the extent that it affected band- Inter-axis coupling is well recognized as one width. These results were extended to pitch atti- of the most severe handling quality problems with tude control in Ref. 12.
unaugmented rotary wing aircraft.
While fixed wing aircraft tend to be much less affected by such coupling, a significant amount of yaw response to roll control inputs is not uncommon at high angles ODen Loop Transfer Function of attack. In such cases the pilot must use rudder coordinated with aileron inputs to eliminate the (S + I/T)e-= -= undesirable heading excursions that occur.
It was 6 se+ 2Q.lJs + 2 hypothesized in Ref.
14 that the pilot opinion of roll-yaw coupling would be directly related to the magnitude and shaping of the rudder control re- quired. Such an approach is expected to be directly applicable to inter-axis the coupling characteristics of helicopters. Because of its possible direct application to helicopter coupling, the results of Ref.
13 are briefly described below.
While the use of "coordinated" aileron and rudder is accepted as common piloting technique, a w (rod /set) + I quantitative measure of what exactly is acceptable or desirable is not known.
The purpose of this study was to provide a quantitative measure of the aileron-rudder sequencing required to eliminate roll-yaw coupling and thereby achieve coordinated turns, and to correlate this with pilot opinion ratings from available data. To achieve this end Ref. 13 considered the aileron-to-rudder crossfeed i necessary to exactly cancel the inter-axis cou- ’ c&=45” pling.
This idealized crossfeed provides a measure -l80”-- of pilot acceptability of heading control because -200 .t i 4 it is indicative of: the complexity of the rudder activity necessary to achieve perfectly coordinated \ turns; and the heading excursions that occur when the pilot does not use rudder.
Note that these considerations equally well to the known apply coupling between pedal, Fig. 5 Effect of using gain and phase margins power, cyclic and collec- tive in an unaugmented helicopter.
to define bandwidth Table 2.
Parameters defining the Two parameters are defined in Ref. 13: u, aileron-rudder crossfeed which defines the shaping of the rudder crossfeed; and Ns /L's , which defines the magnitude. The freque%y r%ponse characteristics of the aileron- to-rudder shaping as a function of the sign of u Analytical Pilot-centered are shown in Fig. 6 in terms of literal expressions Parameter function function for the Bode asymptotes. These asymptotes indicate that the magnitude of the rudder required to coor- IJ Defines shape Determines complexity of dinate is a function of Ns /N& at all frequen- rudder activity necessary cies and that the shaping oy thd?udder response is Of 'CF for ideally coordinated determined by u.
These parameters are summarized turns; also defines phas- in terms of their analytical and pilot-centered ing of heading response functions in Table 2.
when rudder is not used.
The details of the criterion are presented in Defines magni- Determines magnitude of Refs. 13 and 14. The criterion boundaries and the required and/or high- tude of YCF data used to support these boundaries are given in frequency yawing induced Fig. 7. It is interesting to note that the ideal by aileron inputs.
crossfeed was not a pure gain (u = 0). Actually, a little proverse yaw (u = -1) is seen to be desir- able. Similar results could be expected with helicopters, i.e., the coupling can actually be favorable.
Conclusions A great deal of the experience gained in developing handling quality criteria for fixed wing aircraft is directly applicable to rotary wing air- craft as well. In this we have reviewed a par== for/d > 0 Lag Lead Compensation
I Nk,c
/ L’s,,~TR (I+‘) i
fLwp< 0 Lead Lag Compensation = normalized rudder control %C normalized aileron control Tat = Fig. 7 Pilot rating correlation with crossfeed parameters Fig. 6 Asymptotes of aileron-rudder crossfeed Hoh, R. H., and Ashkenas, I. "Handling summar- 6. L*, few areas that seem particularly salient.
izing, these are: Quality and Display Requirements for Low Speed and Hover in Reduced Flight Visibility," Jour- 1) The piloting task and environment are nal of the American Helicopter Society, 26, overriding considerations in developing and using (l), Jan. 1981.
handling quality criterion boundaries.
2) Helicopter pilots have historically been Bryant, W. B., Cattel, J. C., et al., "VTOL willing to put up with considerably more degraded 7.
Advanced Flight Control System Studies for handling qualities than have fixed wing pilots.
The increasing severity of helicopter missions is All-Weather Flight. Vol. I: Task I Report," reversing this trend. USAAMRDL-TR-75-13A, July 1975.
3) Outside visual cues and cockpit displays Lebacqs, J. V., and Aiken, E.W., "A Flight must be considered when structuring a helicopter 8.
handling quality specification. Investigation of Control, Display, and Gui- 4) The poor inherent handling qualities of dance Requirements for Decelerating Descending rotary wing aircraft make SAS failures more criti- VTOL Instrument Transitions Using the X-22A cal than for fixed wing aircraft. Attempting to Variable Stability Aircraft. Vol. I: Techni- cal Discussion and Results," Calspan Corp., impose fixed wing requirements for Levels 2 and 3 Buffalo, NY, Rept. AK-5336-F-1, Sept. 1975.
is probably not practical in terms of cost and complexity.
5) Many handling qualities criteria developed 9. Niessen, F. R., Kelly, J. R., Garren, J. F., for fixed wing aircraft should be directly appli- et al., "The Effect of Variations in Controls cable to helicopters with appropriate revisions in and Displays on Helicopter Instrument Approach the numerical limits and boundaries. Capability," NASA TN D-8385, Feb. 1977.
References 10. Hodgkinson, J., and LaManna, W. .I., "Equiva- lent System Approaches to Handling Qualities 1. Hoh, R. A., and Ashkenas, I. L., 'Development Analysis and Design Problems of Augmented of VIOL Flying Qualities for Low Speed and Aircraft," AIAA Atmospheric Flight Mechanics Hover," NADC-77052-30, Naval Air Development Conf., Hollywood, FL, 8-10 Aug. 1977.
Center, Warminster, PA, Dec. 1979.
11.
Hoh, R. H., Myers, T. T., Ashkenas, I. L.
2. Seckel, E., Traybar, .I. J., and Miller, G. E., Ringland, R. F., and Craig, S. J., "Develop- "Longitudinal Handling Qualities for Hover- ment of Handling Quality Criteria for Aircraft Princeton Department of ing," University, with Independent Control of Six Degrees of Aeronautical Engineering, Rept. 594, Dec. Freedom," AFWAL-TR-81-3027, Air Force Wright 1961. Aeronautical Laboratories, Wright-Patterson AFB, OH, Apr. 1981.
3. Heffley, R. K., Jewell, W. F., Lehman, .I. M., "A Compilation and 12. Hoh, R. H., Mitchell, D. G., and Hodgkinson, and Van Winkle, R. A., Analysis of Helicopter Handling Qualities "Bandwidth -- A Criterion for Highly Aug- J..
Data. Volume One: Data Compilation," NASA CR- mented Airplanes," AIAA-81-1890, AIAA Atmos- 3144, Aug. 1979.
pheric Flight Mechanics Conference, Albuquer- que, NM, 19-21 Aug. 1981.
4. Smith, R. E., "Effects of Control System Dyna- mics on Fighter Approach and Landing Longitu- 13. Ashkenas, I. L., Hoh, R. H., and Craig, S. J., dinal Flying Qualities (Volume I)," AFFDL-TR- "Recommended Revisions to Selected Portions of 78-122, Air Force Flight Dynamics Laboratory, MIL-F-8785B(ASG) and Background Data," AFFDL- Wright-Patterson AFB, OH, Mar. 1978. TR-73-76, Air Force Flight Dynamics Labora- tory, Wright-Patterson AFB, OH, Aug. 1973.
5. Hoh, R. H., and Ashkenas, I. L., "Effect of Reduced Visibility on VTOL Handling Quality 14. Hoh, R. H., and Ashkenas, I. "Handling L-9 and Display Requirements," Journal of Guidance Quality Criterion for Heading Control," Jour- and Control, 4, (2). Mar.-Apr. 1981. nal of Aircraft, 14, (2), Feb. 1977.