Document
AN OVERVIEW OF CONTROLS AND FLYING QUALITIES TECHNOLOGY ON THE F/A-18 HIGH ALPHA RESEARCH VEHICLE Joseph W. Pahle and Keith D. Wichman NASA Dryden Flight Research Center Edwards, CA John V. Foster and W. Thomas Bundick NASA Langley Research Center Hampton, VA ABSTRACT The NASA F/A-18 High Alpha Research Vehicle (HARV) has been the flight test bed of a focused technology effort to significantly increase maneuvering capability at high angles of attack.
Development and flight test of control law design methodologies, handling qualities metrics, per- formance guidelines, and flight evaluation maneuvers are described. The HARV has been modi- fied to include two research control effectors, thrust vectoring, and actuated forebody strakes in order to provide increased control power at high angles of attack. A research flight control system has been used to provide a flexible, easily modified capability for high-angle-of-attack research controls. Different control law design techniques have been implemented and flight-tested, includ- ing eigenstructure assignment, variable gain output feedback, pseudo controls, and model- following. Extensive piloted simulation has been used to develop nonlinear performance guide- lines and handling qualities criteria for high angles of attack. This paper reviews the development and evaluation of technologies useful for high-angle-of-attack control. Design, development, and flight test of the research flight control system, control laws, flying qualities specifications, and flight test maneuvers are described. Flight test results are used to illustrate some of the lessons learned during flight test and handling qualities evaluations.
NOMENCLATURE ANSER actuated nose strakes for enhanced rolling CRAFT control power, robustness, agility, and flying qualities tradeoffs HAIRRY High-Alpha Investigation of Requirements for Roll and Yaw HANG High-Alpha Nosedown Guidelines HARV High Alpha Research Vehicle HATP High-Alpha Technology Program KIAS knots indicated airspeed MDA McDonnell Douglas Aerospace normal acceleration, g PIO pilot-induced oscillation stability-axis roll rate, deg/sec P stability wind-axis roll rate, deg/sec Pwind body-axis pitch rate, deg/sec q stability-axis yaw rate, deg/sec rstability RFCS research flight control system standard evaluation maneuvers set STEMS t time, sec O_ angle of attack, deg angle of sideslip, deg pitch angle, deg wind-axis bank angle, deg ¢Pwind INTRODUCTION Prior to the 1970's, U.S. fighter airplanes exhibited poor stability characteristics at high angles of attack. As a result, maneuvering was often limited by abrupt departure boundaries, and stall and spin accidents were a major cause of loss of aircraft and crew. 1 With the emergence of "all-aspect weapons," close-in combat scenarios of future fighter aircraft are predicted to be dominated by the aircraft that can most rapidly point the nose to obtain the first weapons firing opportunity. Thus, the demand for increased agility and carefree maneuvering throughout the envelope led to a signif- icant change in philosophy toward high-angle-of-attack flight. This change in attitude from "avoid- ance" to one of "exploitation" spawned the development of research programs such as the X-31A Enhanced Fighter Maneuverability, 2'3 F-16 Multi-Axis Thrust Vectoring, 4 X-29A vortex flight control system, 5 and NASA High-Alpha Technology Program (HATP), 6 which were designed to explore and exploit various aspects of the high-angle-of-attack flight regime.
The HATP is a multicenter, multidisciplinary program designed to take advantage of the unique facilities and expertise at the NASA research centers and combine development of analyt- ical tools, ground tests, and flight testing. The computational fluid dynamics research and wind- tunnel testing were supported by the NASA Ames Research Center and the NASA Langley Research Center. Inlet aerodynamics work was led by the NASA Lewis Research Center, and con- trol law research and development was supported by NASA Dryden Flight Research Center and NASA Langley Research Center. Flight testing was conducted at the NASA Dryden Flight Research Center.
As figure 1 shows, the HATP includes research in computational fluid dynamics techniques for computing the behavior of new control effectors and aircraft aerodynamics at high angles of attack, wind-tunnel tests for experimental investigation of these same phenomena, analytical and simula- tion studies of control power requirements, and a concurrent flight test activity to focus the evalu- ation of all the technology areas. The shaded bubbles (fig. 1) show those HATP technologies primarily contributing to control law and handling qualities development at high angles of attack.
requirements Advanced controls Wind-tunnel tests strskl Thrust vectoring Enhanced agility fluid dynamics / Joint and shoot Positioning 960563 Figure 1. HATP research disciplines.
The approach to control system design for current- and previous-generation fighter configura- tions has been primarily driven by two key requirements: achieving angular accelerations and rates for maneuvering, and achieving closed-loop dynamics that provide the desired piloted flying qual- ities for precision tasks. Maneuvering requirements were addressed primarily by proper control sizing to provide the necessary control moment (often called control power) where needed. On configurations using conventional aerodynamic controls, well-defined control system require- ments were typically specified for low-angle-of-attack conditions. Requirements for high angles of attack were not defined other than requiring sufficient margin from entering out-of-control flight.
This lack of definition reflects the common characteristic of conventional aerodynamic controls where control effectiveness rapidly degrades at high angles of attack.
As the demand for high-angle-of-attack maneuverability has increased, the reliance on advanced control effectors and high-authority control augmentation has increased to compensate for the loss of airframe stability. In addition, nonlinear dynamic effects, such as inertial and kine- matic coupling and yaw coordination, are greatly amplified at high angles of attack. As a result,
high-angle-of-attack control power and flying qualitiesrequirements have becomeinseparably
linked andmustbeconsidered togetherwith controlsystemdesignmethodologies andthemaneu-
versusedto evaluatethe overall result.The designof controlsystems for high angleof attack,ct, presents newchallenges andrequiresnewandintegratedapproaches to ensure that theflying qual- ities andmaneuvering performance areoptimized.
Thepurposeof this paperis toprovideanoverviewof the resultsusedto developandevaluate
someof the technologies usefulfor high-_ control.Design,development, andflight testof the re- search flight controlsystems, controllaws,flying qualitiesspecifications, andflight testmaneuvers will be briefly described. Someaspects of the concurrent development of control laws,specifica- tions,andevaluationmaneuvers will bediscussed. Flight testresultswill beusedto illustratesome of the lessons learnedduringflight testandpilotedevaluations.
CONTROLSAND FLYING QUALITIES RESEARCHSCOPE
Oneof thegoalsof theHATP hasbeento developaflight-validatedcontrolsystemdesignpro-
cessfor high anglesof attack.As figure 2 shows,flight controldesigncan be represented as an
iterativeprocess beginningwith airplanemodeldevelopment andaninitial controllaw designand
progressingthrough simulationand flight test stages.Resultsat each stage,requiring specific
methodsandcriteria, arefed backto retunethe designor perhapsevencompletelyredesignthe
controllaw. Because high _ wasnot a well-understood flight regimewhentheHATP began,many
elements of thedesignprocess did notexistor werenot yetflight tested. The approach in the HATP / / / / / \ \ #, Model design and Batch _ Piloted and flight Control analysis ____ Implementation I simulation simulation test 960564 Figure 2. Flight control design process.
wasto address thecriticaldesignguidelines andmethods andto exercise eachelementin thedesign processusingthe NASA F/A-18High Alpha Research Vehicle (HARV) asthe flying testbed.
Although high-c_ flight wasnot a completeunknown,little informationwasavailableon how
to effectivelyintegrateinnovativeeontroleffectorswith advanced controllawsto achievea signif-
icant increasein high-o_ maneuvering capability.The HATP wasstructuredto evaluatethe tech-
nologiesthatcouldbe usedto answersomeinitial questions:
Canthrustvectoringandforebodystrakes beusedto provideadequate, predictablecontrol
powerat high anglesof attack? Whatstrategies canbeusedto effectivelyallocatethe con-
trol poweravailable?
What flight control designtechniques canbe usedor arewell-suitedfor the high-_ flight
regime?Are existing linear control designtechniquesadequate, or will nonlinear tech-
niquesbe required?
• What level of increased controllability andagility is requiredfor a "usable" high-o_ flight
envelope? Whatrequirements or guidelinescanbe usedto definea "usable"envelope?
What flying qualitiesmetricsshouldbe usedto designandevaluatethe control laws?Can
we extendthe existinglow-o_ criteria?Whatnewor evolvingapproaches shouldbeevalu-
ated? What maneuvers shouldbeusedin theevaluation? Whatparameters shouldthe pilot
control with the stick andruddersathigh anglesof attack?
A numberof issuesregarding military utility wereoutsidethe scopeof the HATP andwill not
be coveredin this paper.Theseareasincludespecifictacticsfor successful air combat,weapon/
airframeintegration,anddisplaysfor increased tacticalawareness at high o_.
SIMULATION FACILITIES
A broad range of ground-testand simulationfacilities were usedconcurrently during the
HARV program.The primarypiloted simulationusedin the development of performanceguide-
lines andhandlingqualitiesevaluations wasthe fixed-base, 40-ft domedifferential maneuvering
simulatorat NASA Langley.The differentialmaneuvering simulatorwasalsousedin the control
law designprocessandflight maneuver development. 7 The piloted simulationat NASA Dryden
was limited to forward visualsonly, but could be linked with an all-software,hardware-in-the-
loop,or iron-bird capability.The NASA Drydensimulationwasusedfor flight planning,engineer-
ing, and softwaredevelopment and was the primary site for softwareand hardwaretesting. A
configuration-controlled batchsimulation,commonto both sites,was usedas a benchmarkwith
which to compareotherdissimilarsimulations.
DESCRIPTION AND EVALUATION OF CONTROL LAWS During the HARV program, different flight control laws were planned to be designed and evaluated in flight to cover a broad scope of controls and handling qualities research. To facilitate design and implementation of control laws using various design techniques, the overall control law structure was separated into modules (fig. 3). This modularity allowed the longitudinal and the lateral-directional control laws to be designed using different methodologies. Additionally, the mixer and thrust estimator were designed and modified independent of the control laws, reducing the gain scheduling requirements within each control law by isolating those functions dependent on engine parameters.
Longitudinal surface v control law __ Longltudinal I commends Sensor _[ Thrust- vectoring vectoring v estimator inputs Thrust- commands =,_ Thrust I mixer Lateral- directional ,.._ directional surface -_ I Lateral- control law commands 960565 Figure 3. Modular control law structure.
Four research control laws were designed during the HARV program (table 1). The control laws are described very briefly in the following section. In addition to the major control law releas- es, modifications were required for some of these control laws during flight testing. Only up-and- away flight within a limited envelope (at less than Mach 0.7 and an altitude between 15,000 and 45,000 ft) was considered during the design and flight test because the emphasis in the HATP was on high-or research. The different methodologies used to accomplish these designs will be dis- cussed briefly in a later section.
For all of the control laws, angle of attack was a critical parameter used for feedback and gain scheduling. Unfortunately, angle of attack is not available in the production F/A-18 system at greater than approximately 35 ° . Angle of attack, angle-of-attack rate, angle of sideslip, and sideslip rate were computed in the HARV mission computer using information from the inertial navigation system. 8,9 The signals were then passed to the flight control computer over a 1553 bus and trans- ferred to the research flight control system (RFCS) through the dual-port random-access memory interface. Significant time delay was present in these signals (from approximately 40 to 80 msec), and various techniques were used to compensate for it. _° Initially, using the computed sideslip angle as an inner-loop feedback was desired, but flight test showed poor comparison with the wingtip probes and none of the control law designs evaluated in flight used the signal as feedback.
Although severalmethodsto obtain sideslip at high-o_ were proposed, no viable altemative to this computed sideslip was tested during the program.
Table 1. HARV control laws.
Control First Design Control law flight Axes technique parameter Longitudinal Nonlinear model- Blend of pseudo NASA-0 July following N z and o_ Lateral-directional Eigenstructure assignment Pstability; Pstabili_, Longitudinal Variable-gain output Blend of N z and o_ June feedback NASA- 1A 1994 Lateral-directional CRAFT and pseudo Pstability; controls Longitudinal Variable-gain output Blend of N z and oc ANSER July feedback 1995 Lateral-directional CRAFT and pseudo Pstability; rstability controls Blend of q and o_ Nonlinear dynamic Longitudinal inversion NASA-2 None Lateral-directional Nonlinear dynamic Pstability; inversion The control laws actually used in flight were tested using a progression of maneuvers designed to evaluate performance, handling qualities, and agility. A wide spectrum of open-loop and closed- loop tasks were used. Examples of the open-loop maneuvers are o_ and g captures, pullup- pushovers, stability-axis rolls, and roll reversals, all at various angles of attack to a maximum 65 °.
Air-to-air tracking and gross-acquisition tasks in longitudinal and lateral-directional axes were incorporated early into the flight test plan to identify flying qualities problems. In addition, a lim- ited number of basic fighter maneuvers and close-in combat engagements were flown. This rela- tively quick progression from open-loop to tracking was integrated into the more typical controls envelope expansion to give a "quick-look" evaluation. The tradeoff is that the closer the tasks got to actual air combat, the better the evaluation but the more difficult the engineering interpretation.
To enable correlation of flight results with simulation results, most of these maneuvers were also flown in piloted simulation.
The U.S. Air Force-sponsored standard evaluation maneuvers set (STEMS) study conducted by McDonnell Douglas Aerospace (MDA) (St. Louis, Missouri) focused on developing critical evaluation maneuvers for control system design. Several of these maneuvers were developed spe- cifically for high angles of attack. The purpose of these maneuvers was to definitively show vari- ations in critical control system characteristics and expose problem areas. Obviously, one requirement was that the maneuvers be suitable for flight test by being repeatable and reasonably
easyto perform.As an example,thelateralgrossacquisition(fig. 4) hasbeenshownto be an ex-
cellent maneuverfor evaluatinglateral--directional controlsystemcharacteristics. This maneuver
wasoriginally developedfor theMDA linearflying qualitiesguidelinesstudy.Themaneuverwas
then usedfor control law development andlateralcontrol power evaluations 11beforebecoming
part of the STEMS.Anothermaneuver usedextensivelywasa combinedlongitudinal andlateral-
directionalfine tracking.Figure5 showsanoverplotof two fine-trackingmaneuvers performedon
differentflights by differentpilotsandcontrollaws.Thefigureclearlyshowsthatconsistentresults
within a desiredct region canbe obtainedfor predominantlyheads-uphigh-ctmaneuvers. The
desired tx during these maneuvers was 45 °, and both maneuvers show significant fine-tracking evaluation time where angle of attack was within +10 ° of the desired value.
(x = 25 ° Maintain 180 KIAS
/
Roll to
/
capture (]) target Mach = 0.5, J Altitude = 25,000 ft / (x = 45 °
/
@ 1,500 ft @ Pull to (x = 45 ° Desired performance: • Acquire target within 25 mrad • No overshoot • Desirable time to complete task Mach = 0.5, Altitude = 24,000 ft 960566 Figure 4. Illustration of a lateral gross-acquisition maneuver for 45 ° _.
Lateral and longitudinal fine tracking at 45 (x Pitch 3 .... '," .......__t_/ll-_1- -It- I- ..... _1_'"t-_ !11i-ii; ......._/'i- ][ h"l- ...... i"............ _AA "-..... _ A stick, in. 2
:/__: .................... __t77_
: ' i "l i _1 i i i i i -_ Gross acquisition v I_ Fine tracking 6O
i.................. ...... ............ _-i
deg 30 ............... i 'J ..................... '................... -i................... i-................... '- ...................
0 ' i i _ i 15 20 25 30 35 40 Time, sec 960567 Figure 5. Time histories of two fine-tracking maneuvers flown by different pilots using different control laws.
NASA-0 Control Laws The first set of RFCS control laws, known as NASA-0, were initially developed by MDA and NASA to demonstrate the research utility of the thrust-vectoring control system and to allow initial RFCS flight envelope expansion. 8.9 Integrating aerodynamic and propulsive controls, the control laws were designed for large-amplitude maneuvering as well as stabilized flight at high angles of attack for, data acquisition. The pilot commanded _ and stability-axis roll rate in the longitudinal and lateral-directional axes, respectively. In the low-or, high-dynamic pressure envelope where a blended pitch-rate and normal-acceleration response is desired, the commanded normal accelera- tion was converted to an ot command from a simple model of the lift curve at a fixed, nominal gross weight. In this way, the NASA-0 control law was an o_ command system throughout the envelope.
This control law was used to conduct the first documented closed-loop, multiaxis thrust- vectoring flight. During envelope expansion, considerable high-or data were obtained in steady- state flight (to a maximum 70 ° ct), and preliminary performance results were obtained in maneu- vers similar to those described in the previous section. Control law modifications were required afterinitial envelope expansion to eliminatecontrollaw deficiencies thatwould haveaffectedsub-
sequentflying qualitiesevaluations,implementan on-boardexcitationsystemusedinitially for
aeroservoelastic clearance, andaddthecapabilityto parametricallyvary the nosedownpitch con-
trol powerto supportcontrolpower_2 research. Theon-boardexcitationsystemcapabilitywasused
extensivelyasa genericresearch tool andwasretainedfor all subsequent control law designs.
As a resultof someof the controllaw changes, gross-acquisition tasksin all axesseemed to
improve,but trackingperformance in thepitch axis haddegraded. Postflightanalysisusing some
of the low-o_linear handlingqualitiestools (Smith-Geddes andNeal-Smithprocedures)showed
similar trendsas thoseseenin flight. 13,14 The Smith-Geddes andNeal-Smith tools, developed
primarily for analysis,were usedin a redesignprocessto fine-tune control law modifications.
Although only a limited numberof comparisons were flown, results from flight test indicated
improvedtrackingfor thiscontrollaw (Version28)overthepreviouscontrollaw (Version27)with
no degradation in the gross-acquisition results.Figure6 showsa summaryof the limited handling qualitiesevaluationsaccomplished with all versionsof the NASA-0 control law, generallyresult-
ing in a Level 1-2 result.The histogramshowsCooper-Harper ratingsfor both longitudinal and
lateral-directionalaxessummarized acrossall pilots, flight conditions(includingangleof attack), andevaluationmaneuvers described previously.
[] Longitudinal 14 • Lateral Number of pilot ratings 0 1 2 3 4 5 6 7 8 9 10 Cooper-Harper rating 960568 Figure 6. Summary of Cooper-Harper ratings for the NASA-0 control laws (all versions).
NASA-1A Control Laws A prime objective in the NASA-1A control law design was to demonstrate that enhanced agil- ity, poststall maneuvering, and good handling qualities could be achieved simultaneously. Steady- state flight at high o_ was still required to support flight research in other HATP disciplines. This requirement was a major determinant in the choice of an c_ command system at high o_ and normal acceleration at low oc for the longitudinal axis. The pilot commanded stability-axis roll rate in the lateral axis and angle of sideslip (approximate) in the directional axis.
The NASA-1A control law was designed in-house by the NASA Langley/NASA Dryden control law design team. This intercenter approach was deemed an advantage in that it brought togetherexpertiseandexperience in control theory, simulation,flight dynamics,control system
implementation,andflight testtechniques. The high-_xlineardesignandopen-loopperformance
guidelines(to be discussedin a later section)wereusedasthey becameavailableto guide the
NASA-1A design.In particular,the linearguidelinesfor 30 ° and 45 ° o_ were directly applied to the design of the lateral-directional control law. Using engineering pilots and NASA research pilots, including the HARV project pilots, the NASA-1A control law was evaluated extensively in piloted simulation in the NASA Langley differential maneuvering simulator, and the results were incor- porated into the design process.
Flight performance of the NASA- 1A longitudinal control law was mixed. The pilot disengaged the RFCS on the first flight of this control law because of high-frequency vibration. After some analysis, it was determined that a units error had been made during the aeroservoelastic analysis, and some unmodelled vane dynamics were not adequately suppressed. Structural filters were designed, implemented, and tested within 2 weeks, highlighting one benefit of the RFCS. Envelope expansion resumed with no further structural interactions noted. Steady-state _ control proved to be excellent, and pitch authority was rapid and crisp. For example, the maximum pitch rates achieved in flight were within the performance guidelines with the only significant deviation being at 55 ° _, where pitch-rate capability is limited by insufficient control power.15 The most significant deficiency observed during the flight tests was the overall sensitivity of the control law, resulting in increased surface activity, rate and position saturation, and pilot-induced oscillations (PIOs) during tracking tasks. The pitch sensitivity in flight was much greater than that expected from the piloted simulation results. This flight test experience led to additional research in the areas of PIO analysis and prediction tools and improved piloted simulation maneuvers and techniques.16.17 Flight performance of the lateral--directional control law was generally good. For example, the maximum wind-axis roll rates achieved in flight were very close to the design guideline except at angles of attack of approximately 60 ° , and good lateral-directional tracking and predictability was obtained. However, some discrepancies were noted during the flight evaluation. The pilots did not like having the rudderpedals command sideslip and preferred a conventional pedal response. For angles of attack in the 600-70 ° range, an asymmetry was observed in the airplane roll response.
Stability-axis rolls to the right initiated at 60 ° t_ were completed, whereas rolls to the left could not be completed. In fact, rolls to the left reversed direction to the right, although left stick was main- tained throughout. Grit strips were applied to the forebody in an attempt to alleviate this asymme- try, and rolls to the left at 60 ° ct were successful. However, after the grit was applied, a 1-Hz limit- cycle oscillation in roll with a peak bank-angle amplitude of approximately 0.5 ° was observed on several occasions. Attempts to reproduce this oscillation in the simulation were unsuccessful. 15 Actuated Nose Strakes for Enhanced Rolling Control Laws The actuated nose strakes for enhanced rolling (ANSER) control law was the only control law flown on the HARV designed to command the actuated forebody strakes as well as the thrust- vectoring and conventional aerodynamic effectors. The ANSER control law was developed to accomplish the same broad scope of objectives as the NASA-1A control law with the additional requirement of expanding the ANSER flight envelope and acquiring aerodynamic and flow
visualizationdataregardingthe forebodystrakes.Thecontrol law commanded the standard aero-
dynamicsurfaces andhadthreemodesof operationfor theresearch effectorsselectable by the pi-
lot: TV modeusespitchandyaw thrustvectoring;Smodeusesactuated forebodystrakesandpitch
thrustvectoring;andSTV modeusesactuated forebodystrakes pluspitch andyaw thrust vector-
ing. The mode-switchingfeatureperformedwell andallowedback-to-backcomparisonof agility
andhandlingqualitiesfor the TV, S, andSTV modesduring oneflight. Flight resultsindicated
thrust vectoringand actuated forebodystrakeswereboth effective for controlling the aircraft in
body-axisyaw at high anglesof attack. 18
To minimize designtime andchangesto the flight software,the initial plan was to use the
NASA-1A designasthe ANSERcontrollaw TV mode.Somechanges to the NASA-1A lateral-
directional control laws were requiredin addition to the strakemodifications. Thesechanges
includedreducingthe feedback gain magnitudes at55° and60° c_to eliminatethe 1-Hzroll oscil- lation observed with the NASA- 1A controllaw, limiting thetrailing-edge-down ailerondeflection
asa functionof o_ to reduceadverse yaw, andchangingtherudderpedalpathto providea conven-
tional pedalresponse. Because theeffectof thestrakes onairplanelongitudinaldynamicswasmin-
imal, the plan hadbeento usethe NASA-1A longitudinalcontrollaw in all threeANSER modes
without change.However,the longitudinal sensitivityobservedduring NASA-1A flight testing
forcedsomechanges to the longitudinalcontrollerprior to ANSER flight test.To aid in flight test
evaluation,a dial-a-gainfeature was implementedin the ANSER control law to provide the
capabilityfor the pilot to selectoneof threesetsof gains(low, medium,andhigh) for the longitu- dinal controller.Additionally, the dial-a-gainfeatureallowedfor a wider variationof comparison
betweenflight andpiloted simulationresultsandproduceddatafor evaluationof PIO prediction
andanalysistools.
Fourreleases of the ANSERcontrollaw wererequiredto providethe necessary changes in the
on-boardexcitationsystemto accommodate themanyparameter identification,aerodynamic, flow
visualization,andaeroservoelastic research maneuvers. Onedifferenceof significancein the four
versions wasa change in thec_-scheduled symmetricdeployment of thenosestrakes in Smodeand
STV mode.This changewaspromptedby the occurrence of a small-amplitudeoscillation in the
strakedeflectionsin the 15°-20° o_ region.Thenewsymmetricschedule slightly changedthechar-
acterof the oscillationsbut did not eliminatethem.The inability to reproduce theseoscillationsin
simulationis thoughtto be a resultof modelingerrors,perhapsin the aerodynamic modelsor in
signaldelays.
Target-tracking performanceof the ANSER longitudinal control law was significantly
improvedoverthe NASA-1A control laws.Evaluationof the threegain setsearly in the ANSER
flight testsresultedin themediumgainsetbeingselected asthe defaultfeedback gainsfor mostof
the flight testing. Overalllongitudinal andlateral-directionalperformanceof the controller was
good as evidencedby the clusteringof pilot ratingsat 3 and4 in the histogramof the Cooper-
Harperratings(fig. 7) receivedduringANSERflight testing.Theseratingsaresummarizedfor all
pilots, flight conditions, and rated tasks combined. Ratings were found to be strongly
pilot-dependent, especiallyfor the target-tracking taskat 30° and45° ct. Ostroff, Murphy, Murri,
andHoffler providedetailson thebreakdownof ratingsby task,angleof attack,andpilot andpro-
vide comparisonof flight resultswith piloted simulationresults. 16,_8-2°
60E m [] Longitudinal • Lateral 5O Number of pilot 30 ratings 2O 0 1 2 3 4 5 6 7 8 9 10 Cooper-Harper rating 960569 Figure 7. Summary of Cooper-Harper ratings for the ANSER control laws.
NASA-2 Control Law Initially, it was thought that because the high-ix aerodynamics are dominated by nonlinear effects, a completely nonlinear control law design technique would be required to take full advan- tage of the flight regime. For this reason, a nonlinear dynamic inversion control law was designed by Honeywell Systems and Research Center (Minneapolis, Minnesota) and NASA engineers. 21 Although preliminary all-software piloted simulation results were very positive, this particular nonlinear dynamic inversion control law showed unexpected design and implementation flaws when tested in the hardware-in-the-loop simulation. The control law was very sensitive to differ- ences allowed among the four channels by the quad input signal management, a function resident in the basic F/A-18 flight control computers. Eventually the differences (multiplied by high surface-command gains) would grow large enough to cause failure enunciation at the actuator because of a force fight between the channels. These flaws could not be rectified in time to meet the schedule, and the flight test of this control law was abandoned.
HIGH-ANGLE-OF-ATTACK FLIGHT CONTROL DESIGN GUIDELINES The approach used in the HATP to address flying qualities requirements was to categorize the requirements under three topics: linear closed-loop handling qualities guidelines, agility and open- loop control power guidelines, and nonlinear maneuvering performance guidelines. For the purposes of this research, addressing each category separately was necessary, but it was recognized that these areas are closely linked and must be considered together in the development of final guidelines.
Linear Flying Qualities Guidelines Well-accepted flying qualities design guidelines for low angles of attack, such as MIL-STD- 1797, have been available for many years, but little flying qualities design criteria were available for high anglesof attack.Preliminarylongitudinalandlateral-directionalflying qualitiesrequire- mentsathigh t_ were addressed by NASA-sponsored simulation studies conducted by MDA. 22 The MDA flying qualities studies focused on closed-loop flying qualities requirements from 30 ° to 60 ° or. These studies were conducted to identify critical flying qualities requirements, evaluation ma- neuvers, and piloted test techniques for high angles of attack. These guidelines and maneuvers were used as they became available in the development and testing of the HARV control law designs. Maneuvers developed in the simulation and used for handling qualities evaluation were further refined and used successfully in flight at high angles of attack.
One objective of the MDA studies was to identify key figures of merit that define the flying characteristics pilots desire for acceptable performance. Conventional parameters such as roll- mode time constant and short-period frequency and damping were evaluated to determine if low-or measures were suitable for high-o_ applications. In addition, agility and open-loop control power guidelines focused on measures of maximum maneuvering performance such as peak pitch and roll rates and time to roll to a bank angle. The importance of meaningful, definitive figures of merit was highlighted by Murphy, 23 where numerous candidate parameters were analyzed and correlated to pilot opinion.
In general, results of these studies indicated that typical figures of merit used for low angles of attack were suitable for high angles of attack, but significant differences in the requirements between the two regimes were clear. For example, lateral-axis dynamics criteria (fig. 8) indicated large variations in requirements with or. The implication of this large variation is that the control law design must adjust to provide the desired dynamic response.
.... I .... I .... I .... I .... I ....
A _ 30° ct guideline _ / \ ----- 45 ° (x guideline / \ .oo gu.in. .....
/ 3o oct \ V 4s° aLevell _ Stability-axis / Level 1 region _._ regi°n roll rate sensitivity,
/ /7" -\. -
(deg/sec)/Ibf
/ f ...'"'... _
J...'. "'I
_'_.." " Level 1 region s °" , -- .... 1 .... I .... I .... I .... I ....
Roll-mode time constant, aec 960570 Figure 8. Example linear design guidelines for the lateral axis showing variations of the Level 1 regions with o_.
Correlationof theflight datawith this simulationdatabase is still ongoing.Preliminaryresults,
however,indicatesimilar trendsbut suggest differentboundaries betweenthe levels._4Oneprob-
lematicaspectof this particularsetof criteriais thatit is fundamentally based onlinear lower-order
equivalentsystems. The F/A-18 HARV flight dynamicsare not always represented well in this
frameworkwhen evaluatedwith maneuvers that traversea large o_rangeandthus violate linear
assumptions. Additionally, maneuvers at high o_areoften dominatedby nonlineareffectssuchas
control rate or position limits. For a typical high-o_full-stick 360° roll in flight, the lateral-
directional aerodynamic surfacesandyaw thrustvectoringaresaturated throughoutmuch of the
maneuver. As a resultof this controllimiting, wind-axisroll-rateresponse canbe fairly linearwith time andis not easilyrepresented asa uniquefirst-orderresponse.
Agility andOpen-LoopControl PowerGuidelines
The definition of agility hasbeenundercontinuingdebatefor many years;however,general
agreementexists that agility involves the ability to achieveangularrates or accelerationsfor maneuvering. Examplesincludepeakpitch-acceleration, peakroll-rate,andtime-to-bankcriteria.
Because thesecriteria arecloselytied to maximumavailablecontrol power,agility andopen-loop
controlpowerguidelinesareinterrelated andwereaddressed togetherin this research. Preliminary agility andcontrol powerguidelinesweredevelopedin a varietyof simulationStudies, and flight
validationwasaccomplished on manyof thesecriteria.Theimportanceof well-understood control
powerrequirements wasrecognized because of its strongimpacton airplanedesign.For example,
nosedown controlpowercandetermine the horizontaltail sizeandcenter-of-gravitylocation.Sim-
ilar to thelinearflying qualitiesstudies,amajorobjectivein controlpowerandagility research was to identify key figures of merit that define the maneuveringperformancedesired by pilots.
Also, maneuversthat define maneuveringperformanceand are suitablefor flight testing were
highly desired.
Open-loopmaneuvering goalswereaddressed earlyin the HATP by Hoffler in the develop-
mentof maneuvering andagility goalsfor theHARV.24 Theseguidelineswereintendedto provide
preliminary guidancefor designof the thrust-vectoringsystem.Although theseguidelineswere
developedprimarily for the HARV, they weregenerallysuitablefor preliminary assessment of
configurationsusingadvanced controls.Figure9 showsflight resultsusingthe NASA-1A control
law (at 1 g and an altitude of 25,000 ft) plotted against the 1-g maximum pitch-rate guideline.
Agility metrics were addressed through simulation and flight experiments to identify key fig- ures of merit that the pilot uses to judge airplane response. 25 The approach involved piloted eval- uation of definitive maneuvers for a range of performance levels. Figure 10 shows examples of maneuvers that were evaluated. One of the objectives of this study was to evaluate a broad range of agility and handling qualities levels to assess the tradeoff between these two areas. Analysis of the flight data and correlation with simulation is still in progress.
C) Simulation Flight Maximum pitch rate, deg/sec Initial _, deg 960571 with pitch-rate guideline Figure 9. Comparison of NASA-1A simulation and flight results (maximum afterburner power at an altitude of 25,000 ft).
Roll Pitch v Noseup: maximum positive pitch rate Maximum rate roll through (x capture I_wlndl : 180 ° 0 capture Nosedown: maximum negaive pitch rate I_)windl : 90 °, 180 ° capture (x = 0 ° capture e = 0 ° capture 960572 Figure 10. Example agility maneuvers used in simulation and flight test.
Beginning in 1990, broad-scope control power assessments were initiated as part of a joint NASA-U.S. Navy study addressing controls design requirements for next-generation fighter air- craft. The High-Alpha Nosedown Guidelines (HANG) program addressed the minimum nosedown control response required for safety of flight. This study used piloted simulation to develop a database focused on nosedown response requirements. Results showed that pilots judged minimum nosedown requirements on the short-term response such as pitch acceleration and pitch-rate build- up, and a single design value was identified. 25 The NASA-0 control laws included pilot-selectable variability in nosedown pitch authority and command shaping using variable-rate and position- limiting logic. The piloted simulation results using this capability were validated through a series of flight evaluations. Figure 11 shows pitch control power variations from three maneuvers flown sequentially on a single flight. The three identical maneuvers, full-stick pushovers from initial con- ditions of 1 g, an altitude of 25,000 ft, and 50 ° or, are overplotted to show the dramatically different aircraft responses that were evaluated by the pilot and compared to the design values selected from simulation. Results of this study are in use for next-generation designs, and tactical nosedown requirements are now being addressed in simulation studies.
Sequential HANG maneuvers using different on-board excitation system settings, Flight 172 4 ............... 4 ................ ', ................ "_ ................ ', ................ ,'- .........
Pitch stick, 0 ............... "i ............................... ; ................ i ..............
in.
-2 .......... ; ................ ; ........ _ ..... ; ............... ; .........
ql -10 deg/sec - 20 - 30 2O Pitch acceleration, deg/sec 2 - 20 deg 0 1 2 3 4 5 6 Time, sec 960573 Figure 11. Three sequential HANG pushover maneuvers from 50 ° ot showing aircraft response for different levels of pitch control power.
The High-Alpha Investigation of Requirements for Roll and Yaw (HAIRRY) program ad- dressed high-_ roll maneuvering requirements for various performance levels. ]] The HAIRRY study also used piloted simulation to develop an extensive database of roll maneuvering require- ments from 15 ° to 60 ° (x. Whereas the HANG study focused on safety-of-flight considerations, the HAIRRY study addressed the tradeoff of roll maneuverability with tactical effectiveness.
Generally, a clear variation in roll maneuvering requirements with angle of attack was identified.
Figure 12 shows a preliminary criterion, based on the simulation results, where time to roll is used asa figure of merit.Limited flight evaluations usingtheHARV werecompleted, providing prelim- inary validation of the simulationresults. 26In general,the flight results(fig. 12)indicatereason- ablecorrelationwith the simulationcriteria,but acomprehensive flight programis needed to fully validatethe simulationresults.
i I i i i i Flight results ',
: /!
...... O Enhancing ................................... i ..... _¢ .... T ...........
[] Satisfactory ..... : _!
Z_ M unacceptame ,_ , ; .................... J .......... J ........... 1 ...........
........... p............ t- - ° i i i i t _ i i i i i i _ i i i i i _ i i i i i i i i q argjnal "_i:i_iiiii ........... ........................ i......... y..........
i _ t i I tA(_win d = 90 °, i _ i i p i , r i i t aec j b I i t i , p t i _ ........... f- ....................... , ......... "1 ........... 1 ......... • ...........
• • , , ,.,aruma, Stmulat,on results "-r.. - " " ' " i i p i • ........... _ ............. r'l- ancing ................ ..... i.......................
t i i i i i _, deg 960574 Figure 12. Preliminary roll-maneuvering criteria from the HAIRRY study.
Nonlinear Maneuvering Performance Guidelines Nonlinear performance guidelines addressed handling qualities issues related to nonlinear characteristics of aircraft dynamics. Examples included roll overshoot during aggressive bank- angle captures and sideslip excursions during rolling maneuvers. In the HATP, numerous guide- lines were developed, primarily from piloted simulation. 7 Use of these guidelines involved a "cut- and-try" approach because explicitly integrating the guidelines into control law design is difficult.
However, these requirements have been shown to have a significant impact on flying qualities and highly influenced the HARV control law design throughout the program. 15'27 Figure 13 shows simulation and flight results from one control law configuration plotted with the combined 1-g roll performance and overshoot guidelines. Initially, the roll overshoot guideline was not included, and roll rate was maximized without consideration to roll overshoot. Pilot comments during simulation and flight evaluation of the NASA-0 (Version 27) control laws, however, indicated that the roll overshoot was a key factor in determining the tradeoff between performance and controllability.
The NASA-1A and ANSER lateral control law designs significantly reduced the roll overshoot, and this reduction resulted in improved pilot comments.
0 Flight [] Simulation ------ Guideline
................... :==_==_-_==_-_-
--_% ....... I, ...............................
Maximum o!
Pwind
-D...... -_-R----'_ .... __e_____o__
deg 0 within .... ............ ........... ......
_(_wind = 90°
g_°_
-5 degJsec ................. L.......... : _ .a...o- _-Q 10 1- : : : i i ....
:/ deg o[ ........... _. ..... g--__,_--:----_-P.---.
............ i- i // "_, : ............ i-/--Q ..... i ........... :, ......
tA_win d = 90 °, 80 I
/( [] " ! ............ -............i...........i--u-
deg/sec [] : : : ................... 'l °-- _1_.._, L ........ .;......
@wind
.,d _ :. E
overshoot,
__.¢._.u_=__O.__ a__ ___
deg/aec --_ ..... J ........... J ......
i i 20 ............ _............ ! ........... ! ......
i i i i I i i ' o ' 20 40 60 0 20 4 60 0 _, deg C(, deg 960575 Figure 13. Comparison of simulation and flight results with nonlinear performance guideline: maximum Pwind and time to achieve _wind = 90° plotted with AOc, AI], and I_win d overshoot criterion.
CONTROL LAW DESIGN METHODOLOGIES A goal of the HARV program was to develop flight-validated control design methodologies for application to the very nonlinear high-_ regime. A number of different control law design meth- odologies were used and implemented over the life of the HARV program in order to cover a broad spectrum of effective ideas and to increase the experience base at high c_. The design methodolo- gies were selected in order to evaluate technologies that were thought to be critical to next- generation aircraft control law development. These technologies include: • the capability to integrate handling qualities specifications easily in the design phase.
• robust stability and maneuvering performance throughout the envelope, particularly at high o_.
• efficient control power allocation.
• integrated aerodynamic, forebody vortex, and propulsive controls.
Longitudinal Control Laws Two significantly different techniques were used to design the longitudinal axis in the NASA-0 and NASA-1A and ANSER control laws. For both control laws, symmetric leading and trailing flap were commanded as a function of 0_, identical to the standard F/A- 18 control laws. Horizontal stabilator and pitch-vectoring commands were blended with a washout to eliminate steady-state vectoring commands caused by thermal limits on the vanes.
The NASA-0 technique was a continuous, nonlinear model-following scheme that tried to force the aircraft response to a desired second-order transfer function. 28 Desired eigenvalues were defined by the selection of the short-period parameters for the lower-order transfer function. Open- loop dynamics were computed from a nonlinear aerodynamic model contained in the control laws.
Control system gains were computed from the algebraic difference between the desired and open- loop short-period approximations. Compensation was added to account for the high-frequency effects (actuator and sensor dynamics and structural filtering) neglected in the simplified nonlinear aerodynamic model. Pitch rate, estimated _ rate, and o_ were used as the primary feedback. Non- linear compensation included inertial coupling (roll rate times yaw rate) and gyroscopic coupling.
At low dynamic pressures and high throttle settings typical of high-o_ thrust-vectored flight, gyro- scopic coupling was a significant part of the cross-axis vehicle dynamics.
The NASA-1A and ANSER longitudinal control laws were a direct digital design accom- plished using variable-gain output feedback, an approach initially developed for NASA Langley under contract. The technique is derived from a stochastic, optimal, discrete, output-feedback design approach developed by Halyo and Broussard 29 and is useful for extending the operating envelope of linear control laws. Traditional design methods involve performing constant-gain designs at several different operating points using linear techniques, then creating a gain schedule between points using some type of curve fit. With the variable-gain output feedback technique, all design points are handled simultaneously. Linear models at several aircraft operating points are integrated into a single design problem to obtain a variable-gain global controller with gains that are functionals of aircraft parameters (fig. 14). Gains are adjusted for varying operating conditions by evaluating these functionals, which may be linear, such asf(Mach), or nonlinear, such asf((x3).
This technique can be used with many different control structures, but for the HARV, a proportional-integral filter structure was used with the control equations implemented in incremen- tal form. 3° For NASA-1A and ANSER, feedback gains were computed simultaneously for 39 de- sign points, or flight conditions, and a gain-scheduling algorithm for computing feedback gains between design points was designed at the same time. Feedback parameters were pitch rate, 0_, and normal acceleration. A detailed discussion of the design, simulation, and flight results can be found in references. 3°,31 Variable-gain output feedback is an optimal control design technique that computes feedback gains by minimizing a quadratic cost function. A drawback to this technique is the difficulty in directly incorporating handling qualities criteria such as the MDA linear guidelines. Tuning the variable-gain output feedback design in the sense of adjusting a single feedback path is also diffi- cult because the technique does not readily accommodate changing one gain at a time. This diffi- culty was encountered in the adjustment of the integrator gain, which reduced the longitudinal
sensitivityin theNASA-1A redesign. By adjustingweightingmatrices,however,the designerwas
ableto adjustgainssothe dominantchangewasin theintegratorgain.
Linear design models I Plant 1 I o Measurements Control inputs r o Plant n Nonlinear feedback I_ 1.)
q !
K(p) = K0 + _, Pi (_) Ki i=1 960576 Figure 14. Variable-gain output feedback design method.
Lateral-Directional Control Laws The NASA-0 and NASA-1A and ANSER control laws used eigenstructure assignment 32-35 for the basis of the lateral--directional control law design, although the control laws varied greatly in focus and actual implementation. Both control laws used a fairly conventional set of feedback parameters that included roll and yaw rates, lateral acceleration, and estimated sideslip rate.
The NASA-0 control law approach was a fairly standard eigenstructure-assignment design. Control power allocation was through a fixed ratio of surface deflection. At a dynamic pres- sure greater than approximately 150 lbf/ft 2, however, the control laws transitioned to the standard F/A-18 lateral-directional commands with some additional yaw thrust vectoring. This transition was required because of inadequate stability margins and excessive gains resulting from this eigenstructure-assignment design.
The NASA-1A and ANSER lateral-directional control laws were divided into two modules to try to separate the tasks of designing the feedback gains from allocation of the multiple control effectors. Feedback gains were designed at 12 flight conditions using a methodology called control power, robustness, agility, and flying qualities tradeoffs (CRAFT) 15 and then scheduled with flight condition.
The CRAFT technique addresses the design objectives of satisfying the control power con- straints, providing adequate robustness, maximizing agility, and providing satisfactory flying qual- ities. The CRAFT technique provides a graphical method to allow the designer to perform the
tradeoffsrequiredduring the linear designphaseto achievea designthat is the bestcompromise
amongthe four designobjectives. _5 As figure 15shows,thesetradeoffsareaccomplished by sys-
tematicallyevaluatingtheclosed-loopsystem atspecifieddesignpointsoveranappropriate design
space. At eachdesignpoint, closed-loop metrics,which quantifythe designobjectives,areevalu-
atedandplottedoverthedesignspace. In asense, theCRAFTtechnique is a "bruteforce" approach
in that the designerattemptsto find the bestsolutionby designingandcomputingmetrics for a
largenumberof designs. Theadvantages of theCRAFTtechnique arethatit automates the process
of designingover the matrix of pointsandprovidesa compositegraphicaldisplay of the results.
Theseplotsindicatethedesirable regionsin designspace based onmetric values.Graphicallyover- laying desirableregionsgivesthe designer a clearview of thetradeoffsandsensitivitiesinvolved.
Gainschedulingthe resultingsingle-pointdesigns, however,is still required.
llllll Control design metrics
Ic°ntr°' °"gn c'°"d"°°°'"t'm F'n.'O..'on I
• algorithm analysis _ Best design compromise 960577 Figure 15. CRAFT design method.
The feed-forward path includes inertial compensation and a distributor to apportion the control law commands to the appropriate aerodynamic and thrust-vectoring control effectors. The distrib- utor was designed using a technique known as pseudo controls. 36 For example, the rudder, yaw thrust-vectoring system, and actuated nose strakes effectively produce body-axis yawing moment.
The relative control effectiveness of each effector, though, is not constant over the entire flight envelope. Using a model of the relative control effectiveness of each surface over the flight enve- lope, pseudo controls allow the designer to develop surface schedules that apportion the command to the control effectors in a manner that maximizes the moment in the desired axis while minimiz- ing the moment produced in the other axes.
THRUST-VECTORING VANE MIXER
TheHARV produces multiaxis thrustvectoringusinganexperimental thrust-vectoringsystem
with six thrust-vectoring vanes. 9,37 These thrust-vectoring vanes are interfaced with the fli.ght con- trol laws through a separate function known as the mixer. The control effectiveness of each vane is highly nonlinear, dependent on engine parameters and flight condition, and--unlike convention- al aerodynamic surfaces--very dependent on the position of the other vanes around the same engine. This latter behavior made conventional surface scheduling impractical. Isolating the thrust- vectoring control allocation within the mixer allows the control laws to be designed or modified separately, as mentioned previously, with three moment commands (pitch, roll, and yaw) rather than six vane commands. In addition, the mixer adjusts the thrust-vectoring commands to account for changes in thrust level and losses in thrust caused by thrust vectoring 38 and limits the commands as a function of flight condition to avoid excessive structural loads.
Because direct measurement of in-flight gross thrust was not available to the flight controls, in- flight gross thrust was calculated from a model of ideal gross thrust based on nozzle throat area, nozzle pressure ratio, and ambient pressure. 28 Comparison with in-flight real-time thrust measurement 39 values for each engine indicated that results of this estimated gross thrust were within approximately 7 percent of measured values.
Two mixer designs were used in the HARV program. The first was developed by the contractor 28 and delivered with the first release of the NASA-0 control laws. A second mixer (Mix- er 4.2) was developed by the HARV controls team to improve the vectoring performance, add the capability of roll vectoring, and implement a priority scheme between the vectoring commands when the thrust-vectoring system was not capable of satisfying those commands simultaneously, n° Mixer 4.2 was used in the NASA-1A and ANSER control laws, and a modified version was used in the NASA-2 design. Even though Mixer 4.2 was designed with roll thrust-vectoring capability, roll thrust vectoring was not used in the NASA-1A and ANSER control laws.
Both mixers were based on ground tests with a 14.25-percent model of the HARV thrust- vectoring system, nl This high-pressure cold-jet test was conducted to measure the thrust-vectoring effectiveness of the vane system (that is, to measure the thrust vectoring as a function of the vane deflection angles). Recent flight results from parameter identification show excellent comparisons with the cold-jet results. For the Mixer 4.2 design, a numerical optimization procedure was used to invert the cold-jet data to obtain vane deflection angles as a function of desired pitch- and yaw- moment commands. To conserve memory in the flight computer, these inverted data were stored in variable-density, nonrectangular arrays such that data outside the irregularly shaped boundary of achievable thrust vectoring were not included.
ROLE OF PILOTED SIMULATION IN DESIGN PROCESS A program goal had been to demonstrate that even in the relatively unknown high-or flight regime, the amount of flight test time required to develop a new control law and the number of
designchanges duringflight testingcouldbeminimizedby extensivelyusingpiloted simulationin
an iterativeprocesswith control designandtuning andwith newhigh-o_ designguidelines.Obvi-
ously,this goalwasnot accomplished with unqualifiedsuccess because the occurrence of PIO with
the NASA-1A longitudinalcontrol law wasnot predictedby simulation,althoughthe control law
wasevaluated extensivelyin theNASA Langleydifferentialmaneuvering simulatorwith multiple
pilots, including theprojectflight testpilots.The reasons for this arenot clear,but possibilitiesin- clude lack of motion cues,pilot adaptation because of familiarity with the tasksandthe control laws,andgenerallylow pilot gain in simulationcompared to flight. Reliability of simulatorresults may beimprovedby carefullydefiningthetask,giving carefulconsideration to all pilot comments,
andfully investigatingthe reasons for isolatedpoorpilot ratings.16
CONCLUDING REMARKS
The High-Alpha TechnologyProgram(HATP) was considered uniquebecauseof the broad
scopeof research thatwasconducted duringtheprogram.Significantprogress wasmadein devel-
oping designguidelinesthat did not exist prior to the HATP, andtheseresultsshouldprovide a
foundationfor future designs.
Conventionalflying qualitiesmetricsweresuccessfully evaluated athigh anglesof attack.Cor-
relation to flight for thosemetrics basedon low-order linear systemswas a complicatedtask
becauseof nonlinear responsecharacteristics.Good handling qualities were achieved, but
numerousissuessuchas pilot commandvariableswereleft unresolved.Appropriateevaluation
maneuvers wereshownto becritical to expose controlsystemproblems.
Performance in the conventionaland poststallflight regimescanbe significantly improved
with a thrust-vectoringandforebody-vortexcontrol capability and advancedcontrol laws. The
large increasein controllability was critical in achievingthe unprecedented levels of agility and carefreemaneuverability demonstrated. Flight resultsindicatedthrustvectoringandactuated fore- bodystrakeswereeffective for controlling the aircraftin body-axisyaw at high anglesof attack.
Linearcontrol law designtechniques canbeusedsuccessfully in nonlinearflight regimeswith
somenonlinearcompensation techniques to accountfor cross-axis couplingtermsreadily apparent
at high angleof attack.Easeof modificationis an importantconsiderationin implementationand
subsequent useof research flight controllaws.Capabilitiessuchasan on-boardexcitation system anddial-a-gaincanbe effectivelyusedto reduce the flight testtime required.
The importanceof anaccuratesimulationmodel wasclearly demonstrated. Every stepof the
designprocessreliedheavilyon a modelthataccounted for the nonlinearaerodynamics aswell as
flight computerand sensorcharacteristics. Piloted simulationwasshown to be a key elementin
helping to maturea designprior to flight andaddressing problemareas.However,pilot-induced
oscillationprediction,using simulationandcurrentanalyticalmethods,wasshownto be a signifi-
cantproblemareathatwill requirefocusedstudyfor futuredesigns.
A primary goal of the controlsresearch in the HATP wasto accelerate andmaturecontrols
designtechnologyfor futurefighter aircraft.This paperhassummarized the approach to achieving this goal by discussingresearch activitieson the variouselements in the designprocess.Several control designmethodologies wereevaluated in flight, illustratingthe strengthsandweaknesses in thedesignprocess.
REFERENCES
_Chambers, Joseph R., "High-Angle-of-AttackTechnology:Progress andChallenges," High-
Angle-of-Attack Technology, NASA CP-3149, vol. 1, May 1992, pp. 1-22.
2Herbst, W. B., "X-31A at First Flight" Flying Qualities, AGARD CP-508, Feb. 1991, pp. 29-1- 29-8.
3Canter, Dave E., "X-31 Post-Stall Envelope Expansion and Tactical Utility Testing," Fourth High Alpha Conference, NASA CP-10143, vol. 2, July 1994.
4Kidman, David S., Vickers, Jesse E., Olson, Brian P., and Gerzanics, Michael A., Evaluation of the F-16 Multi-Axis Thrust Vectoring Aircraft, AFFTC-TR-95-12, Sept. 1995.
5Smith, W., "'X-29 High AOA Flight Test Results: An Overview," SAE Aerospace Atlantic Conference and Exposition, April 1993.
6Gilbert, William P., Nguyen, Luat T., and Gera, Joseph, "Control Research in the NASA High- Alpha Technology Program," AGARD Fluid Dynamics Panel Symposium on Aerodynamics of Combat Aircraft Control and of Ground Effects, Oct. 1989.
7Hoffler, Keith D., et al, "Evaluation Maneuver and Guideline Development for High Alpha Control Law Design Using Piloted Simulation," AIAA-94-3512, Aug. 1994.
8McDonnell Aircraft Company, NASA High Alpha Research Vehicle Thrust Vectoring Control System (HARV/TVCS), Control Law Specification Document, Mar. 1989.
9pahle, Joseph W., Powers, Bruce, Regenie, Victoria, Chacon, Vince, Degroote, Steve, and Murnyak, Steven, Research Flight-Control System Development for the F-18 High Alpha Research Vehicle, NASA TM- 104232, April 1991.
l°Carter, John F., "Inertially Derived Flow Angle Measurements," High-Angle-of-Attack Projects and Technology Conference, NASA CP-3207, April 1992, pp. 51-68.
l lFoster, John V., Ross, Holly M., and Ashley, Patrick A., "Investigation of High-Alpha Lateral- Directional Control Power Requirements for High-Performance Aircraft," AIAA-93-3647, Aug. 1993.
12Ogburn, Marilyn, Ross, Holly M., et al, "Flight Validation of Ground-Based Assessment for Control Power Requirements at High Angles of Attack," Fourth High Alpha Conference, NASA CP-10143, July 1994.
_3Wichman, Keith D., Pahle, Joseph W., Wilson, R. Joseph, Connelly, Patrick J., and Kelly, John W., "High Alpha Handling Qualities and Agility Flight Research on the F/A-18 High Alpha Research Vehicle," Fourth High Alpha Conference, NASA CP-10143, vol. 2, July 1994.
14Wichman, Keith D., Pahle, Joseph W., Bahm, Catherine M., Davidson, John B., Bacon, Barton B., Murphy, Patrick C., Ostroff, Aaron J., and Homer, Keith D., "High Alpha Handling Qualities Flight Research on the NASA F/A-18 High Alpha Research Vehicle," High-Angle-of- Attack Technology Conference, Sept. 1996.
15Murphy, Patrick C., et al, "Preliminary Evaluation of HARV NASA-1 Control Law Flight Test Results," Fourth High Alpha Conference, NASA CP-10143, vol. 2, July 1994.
_6Hoffler, Keith D. and Messina, Michael D., "Flight Test and Piloted Simulation Assessment of Evaluation Maneuvers and Nonlinear Guidelines Developed for High-Alpha Control Law Design" High Angle-of-Attack Technology Conference, Sept. 1996.
17Bacon, Barton J., "PIO Susceptibility of the F-18 HARV: Analysis/Prediction vs. Flight," High-Angle-of-Attack Technology Conference, Sept. 1996.
18Murri, Daniel G., Fisher, David E, and Lanser, Wendy R., "Flight-Test Results of Actuated Forebody Strake Controls on the F-18 High-Alpha Research Vehicle," High-Angle-of-Attack Technology Conference, Sept. 1996.
19Ostroff, Aaron J. and Wichman, Keith D., "Flight Evaluation of a Variable-Gain Output Feed- back Longitudinal Control Law for the NASA F/A-18 High Alpha Research Vehicle (HARV)," High-Angle-of-Attack Technology Conference, Sept. 1996.
2°Murphy, Patrick C., Davidson, John B., Lallman, Fredrick J., Morelli, Eugene A., Messina, Michael D., Thomson, Michael P., and Connelly, Patrick J., "An Evaluation of Design Methodol- ogy and High-Alpha Design Criteria for the ANSER Lateral-Directional Control Law," High- Angle-of-Attack Technology Conference, Sept. 1996.
21Enns, Dale, Bugajski, Dan, Hendrick, Russ, and Stein, Gunter, Dynamic Inversion: An Evolving Methodology for Flight Control Design, Jan. 1995.
22Wilson, David J. and Citurs, Kevin D., High Angle of Attack Flying Qualities Design Guide- lines, NASA CR-4681, vol. I and II, Feb. 1996.
23Murphy, Patrick C., Bailey, Melvin L., and Ostroff, Aaron J., Candidate Control Design Metrics for an Agile Fighter, NASA TM-4238, March 1991.
24Foster, John V., Bundick, W. Thomas, and Pahle, Joseph W., "Controls for Agility Research in the NASA High-Alpha Technology Program," S AE-912148, Sept. 1991.
25Ogbum, Marilyn, Batterson, James, et al, "High-c_ Agility/Handling Qualities," High Alpha Technology Program Workshop, Nov. 1989.
26Ross, Holly M., Foster, John V., and Lluch, Daniel C., "High-Angle-of-Attack Roll Perfor- mance Guidelines: Simulation and Flight Validation Results," High-Angle-of-Attack Technology Conference, Sept. 1996.
27Davidson, John B., Foster, John V., Ostroff, Aaron J., Lallman, Fredrick J., Murphy, Patrick C., Hoffler, Keith D., and Messina, Michael D., "Development of a Control Law Design Process Utilizing Advanced Synthesis Methods with Application to the NASA F-18 HARV," High- Angle-of-Attack Projects and Technology Conference, NASA CP-3137, vol. 4, April 1992, pp. 111-158.
28McDonnell Aircraft Company, NASA High Alpha Research Vehicle Thrust Vectoring Control
System (HARV/TVCS): Theory and Implementation of the HARV/TVCS Mixer Predictor, HARV- MJH-12, Aug. 1989.
29Halyo, Nesim, Moerder, Daniel D., Broussard, John R., and Taylor, Deborah B., A Variable- Gain Output Feedback Control Design Methodology, NASA CR-4226, Mar. 1989.
3°Ostroff, Aaron J. and Proffitt, Melissa S., Longitudinal-Control Design Approach for High- Angle-of-Attack Aircraft, NASA TP-3302, Feb. 1993.
31Ostroff, Aaron J., Hoffler, Keith D., and Proffitt, Melissa S., High-Alpha Research Vehicle (HARV) Longitudinal Controller: Design, Analyses, and Simulation Results, NASA TP-3446, July 1994.
32Moore, B. C., "On the Flexibility Offered by State Feedback in Multivariable Systems Beyond Closed Loop Eigenvalue Assignment," IEEE Transactions on Automatic Control, vol. 21, Oct. 1976, pp. 689-691.
33Andry, A. N., Jr., Shapiro, E. Y., and Chung, J. C., "Eigenstructure Assignment for Linear Sys- tems," IEEE Transactions on Aerospace and Electronic Systems, Sept. 1983, pp. 711-729.
34Sobel, K. M. and Shapiro, E. Y., "Application of Eigensystem Assignment to Lateral Trans- lation and Yaw Pointing Flight Control," The 23rd IEEE Conference on Decision and Control, Dec. 1984, pp. 1423-1428.
35Davidson, John B. and Schrnidt, David K., Flight Control Synthesis for Flexible Aircraft Using Eigenspace Assignment, NASA CR-178164, June 1986.
36Lallman, Frederick J., Relative Control Effectiveness Technique with Application to Airplane Control Coordination, NASA TP-2416, April 1985.
37Bowers, Albion H., Pahle, Joseph W., Wilson, R. Joseph, Flick, Bradley C., and Rood, Richard L., "An Overview of the NASA F- 18 High Alpha Research Vehicle," High-Angle-of-Attack Technology Conference, Sept. 1996.
38Johnson, Steven A., Aircraft Ground Test and Subscale Model Results of Axial Thrust Loss Caused by Thrust Vectoring Using Turning Vanes, NASA TM-4341, Jan. 1992.
39Alexander, R. I. and Ray, Ronald J., Development and Flight Test of a Real-Time Thrust Mea- surement Technique on the X-29A/F404 Advanced Technology Demonstrator, NASA TM- I 01707, Sept. 1989.
4°Bundick, W. Thomas, Pahle, Joseph W., Yeager, Jessie C., and Beissner, Fred L., Jr., Design of a Mixer for the Thrust-Vectoring System on the High-Alpha Research Vehicle, NASA TM- 110228, June 1996.
41Bowers, Albion H., Noffz, Gregory K., Grafton, Sue B., Mason, Mary L., and Peron, Lee R., Multiaxis Thrust Vectoring Using Axisymetric Nozzles and Postexit Vanes on an F/A-18 Configu- ration Vehicle, NASA TM- 101741, April 1991.