I AD-Ri39 976 A PILOTED SIMULATOR INVESTIGATION OF DECOUPLINGi/I HELICOPTERS BY USING A MO..(U) ARMY RESEARCH ANDI TECHNOLOGY LAOS MOFFETT FIELD CA AEROMECHAN.

I UNCLASSIFIED G BOUldER ET AL- MAY 84 F/6 1/3 H

EOMEONEEhE

:7 -77 -7 T~ 7,7

N i s _ _-...

mmii W 22.20|u111111 L 1U.8

11 1.25 111111.4 11 .111W-

MICROCOPY RESOLUTION TEST CHART

NATIONAL BURESAU OF STANDARDS - IA3 - A

iz*I 1* . h.J*

A PILOTED SIMULATOR INVESTIGATION OF DECOUPLING HELICOPTERSBY USING A MODEL FOLLOWING CONTROL SYSTEM

Gerhard BouwerDeutsche Forschungs- und Versuchsanstalt fur Luft- und

Raumfahrt e.V. (DFVLR)Institut fur Flugmechanik

Braunschweig, West Germany

Kathryn B. HilbertAeromechanics Laboratory

U.S. Army Research and Technology Laboratories (AVSCOM)NASA Ames Research CenterMoffett Field, California

PRESENTED AT THE 40TH ANNUAL FORUMOF THE

AMERICAN HELICOPTER SOCIETY

ARLINGTON, VIRGINIA

MAY 16-18, 1984

* .

.

ELECTE

* _ A."

h - b"-:-'.N-.. hj; r I; n-:. .. i i A H S, 21;7 n. 8%.-l , t , l... -- . . - -

\ This document has beenCfox public release and sale; its

ditribution_. is unlimite& ,

PAPER NO. A-84-40-07-4000

84 04 03 14"

If

4

Table I Simulated BO-105 and UH-lH actuator characteristics

UH-lH V/STOLANDBO-l05-S3

Axis Position limit, Rate limit,Position Rate limit, in. (t) in./sec (%/sec)limit, in./secin. (%) (%/sec) Series Parallel Series Parallel

Longitudinal ±6(100) 9(75) ±1.6(27) ±6(100) 11(92) 1.5(13)Lateral ±6(100) 9(75) ±1.8(30) ±6(100) 12(100) 1.7(14)Collective 0-10(100) 3(30) ±0.9(18) 0-10(100) 13.3(133) 0.9(9)Pedals ±3.5(100) 3.5(50) ±0.9(26) ±3.5(100) 5.8(83) 0.9(13)

Table 2 Pilot experience

Total flight hours

Pilot AffiliationHelicopter Fixed wing BO-105 UH-lH

A DFVLR 5300 400 900 2500B DFVLR 510 330 400 0

Table 3 Test matrix

.4 TaskHelicopter Status Actuating Task

system 60-knot dolphin 60-knot slalom

BO-105 Baseline Ideal Pilots A,B Pilots A,BWith MFCS S-3 Pilots A,B Pilots A,B

UH-lH Baseline Ideal Pilot A Pilot AWith MFCS V/STOLAND Pilots A,B Pilots A,B

Table 4 Averaged pilot ratings

HELICOPTER STATUS RATING

80-105 BASELINE . [] 1WITH MFCS - . ..UH-IH BASELINE !.1; ).

UHH WITH MFCS [11213141516 7 8 9 10,F, . r

HANDLING QUALITIES

HELICOPTER STATUS RATING Acossi0n Fer

90-105 BASELINE 1: ] -' NIS OKRA&I_O-105 WITH MFCS ]IC TAB

BASELINE I ' .....

WITH MFCS

213141567 910

TASK PERFORMANCE f 1 ITASK: LKNOTDOLPHIN _ ..

4i

- .-.- - . j. " T- . .;-W J.- -. - *6.

A PILOTED SIMULATOR INVESTIGATION OF DECOUPLING HELICOPTERSBY USING A MODEL FOLLOWING CONTROL SYSTEM

Gerhard BouwerDeutsche Forschungs- und Versuchsanstalt fur Luft- und Raumfahrt e.V. (DFVLR)

Institut fur FlugmechanikBraunschweig, West Germany

Kathryn B. HilbertAeromechanics Laboratory

U.S. Army Research and Technology Laboratories (AVSCOM)NASA Ames Research CenterMoffett Field, California

Abstract K = gain

A U.S. and German piloted simulation Q = torque, ft-lbexperiment conducted to evaluate the per-formance of a model following control sys- r - yaw rate in the body-c.g. axestem by applying it to hingeless-rotor and system, deg/secteetering-rotor helicopters is reported.The explicit model was a linear, decoupled s = Laplace transform variablemodel such that the pilot commanded pitchattitude with the longitudinal cyclic, T - sample time, secroll attitude with the lateral cyclic, yaw crate with the pedals, and Earth-fixed U = control vector, ' - (6esadc,6P)downward velocity with the collective.The results of the simulation indicate w - Earth-fixed vertical velocity,that the performance of the model follow- ft/secing control system is primarily dependenton the limitations of the actuating system. wf - fuel flow rate, lb/hrSatisfactory handling qualities wereachieved for both augmented helicopters X - helicopter state vector,flying two specified evaluation tasks: (u,w,q,9,v,p, ,r)Tdolphin and slalom maneuvers. The sig-nificant improvements in task performance XM = model state vector,and handling qualities achieved for these IM - (eM,oM,wM,rM)Ttwo radically different helicopters, aug-mented with the designed model following z - unit delaycontrol system, indicates the flexibilityand versatility of this control technique. 6a - lateral cyclic stick movement,

positive to right, in.

Notation dc - collective control input, positiveup, in.

A - helicopter dynamics matrix wherex-Ax + Bu 6e - longitudinal cyclic stick movement,

positive aft, in.B - helicopter control matrix

6p - pedal movement, positive right, in.C - model sensitivity matrix

* A( ) .- increment in-e - error between the helicopter state

and model state 0 Euler pitch angle, deg

G - control system matrices - time constant, sec

HP - power - Euler roll angle, deg

I - identity matrix a = rotor angular velocity, rad/sec

J - rotor moment of inertia about C ] - matrix inverseshaft, slug-ft' h IT - matrix transpose

Presented at the 40th Annual Forum of theAmerican Helicopter Society, Crystal City,Virginia, May 16-18, 1984.

J% %*

%*

Subscripts and a recently developed quantitative mea-sure of model following performance. This

c - commanded paper discusses the results of the pilotedportion of the experiment. Analysis of the

D - discretized model following control system using thenew performance criterion will be reported

E - engine in a subsequent paper.

- - modelExperiment Design

0 - trim-or reference condition

R - reuiredMathematical modelsR -required

Simulation of the baseline helicop-ters was accomplished using two versions

Introduction of a ten-degree-of-freedom, nonlinear,generic helicopter mathematical model.'

A joint research between the One version of the mathematical model wasDeutsche Forschungs- und Versuchsanstalt configured to simulate the flight dynamicfur Luft- und Raumfahrt e.V. (DFVLR) and characteristics of the DFVLR BO-105the U.S. Army, under the Helicopter Flight S3 helicopter and the other to simulateControls Memorandum of Understanding, was those of the NASA/Army UH-lH V/STOLAND

-q established to develop and evaluate model helicopter without stabilizer bar.following control methods. The long-termobjective of the program is to develop and To make possible the transfer of theimplement a model following control system simulation results to future flight teston a flight research facility for use in programs, the different actuating systemshandling qualities, parameter identifica- of these two variable-stability helicop-tion, and control system experiments. The ters were also simulated. The fly-by-wireadvantage of using a model following con- actuators of the BO-105 S3, shown introl system on a fly-by-wire helicopter is Fig. 1, are installed in all four axes andthat the characteristics of the model to mechanically connected to the safetybe followed can be varied quickly and pilot's controls. These actuators areeasily depending on the desired task to full-authority and quick in response. Abe performed. simple second order model of the BO-105

fly-by-wire actuators was simulated forA model following control system was this experiment (Fig. 2). Because of

designed at the DFVLR, and a preliminary limits in the hydraulic force booster, theflight test program, using the MBB BO-105 rates of all the simulated actuators wereS3 helicopter, was conducted to determine constrained.

** *the feasibility of using such a controlsystem for variable stability applications. The UH-lH V/STOLAND actuating systemResults of this experiment demonstrated utilizes a combination of a parallel servothat decoupling of the DO-105 was possible and a series servo in the linkage of each

. using the model following control system, control.' Figure 3 shows the servo instal-and it was recommended that further lations for the cyclic controls. Theresearch be conducted in this area. series servos, which are severely limited

in authority, are the faster respondingS.Before undergoing more extensive actuators and thus act primarily on the

flight test programs, a U.S./German transient behavior. The parallel actua-piloted simulation experiment, the topic tors are full-authority rate servos whichof this paper, was designed and conducted, act to off-load the series servos and thusThe purpose of the experiment was to inves- provide a trimming function. In addition,tigate the decoupling performance and lim- there are disconnect devices in the evalu-itations of the designed model following ation pilot's pitch and roll cyclic con-control system and to determine how easily trols to allow for fly-by-wire operation.it could be adapted to a new baseline Figure 4 shows the simulated UH-IHhelicopter, the NASA/Army variable stabil- V/STOLAND actuating system. The seriesity UH-H V/STOLAND. The BO-lO5 (hinge- servos were modeled by a second order

less rotor) and UH-IH (teetering rotor) filter, and the parallel servos were mod-helicopters were chosen because they have eled by an integrator. If the seriesextreme differences in control system servo operates at its position limit, theactuation, as well as in control power and gain K is used to force the parallelflight dynamics corresponding to their servo to operate at its maximum rate.different rotor systems. Two methods ofassessing the decoupling performance of A comparison of the simulated actuatorthe model following control system were characteristics and limitations for theused in this simulation: pilot ratings BO-105 and UH-lH is shown in Table 1. The

1% N oN %

main limitations in the UH-IH actuating 20 knots by commanding both roll angle andsystem are the limited-authority series yaw rate with the lateral cyclic stick.servos. The main restriction in theB0-105 actuating system is the collective Model Following Control Systemrate limit of 30%/sec. This limitationis primarily the result of the rotor rpm In a typical model following controldynamics and limits. To represent accu- system (MFCS) (Fig. 7) the pilot's commandsrately the influence of the rotor rpm on are disconnected from the actual aircraftthe model following control system, a and fed into a model. This model may, forsimple model of the 8O-105's engine and example, be the equations of motion of anrpm governor was implemented in the simu- aircraft to be investigated or the dynam-lation mathematical model. Figure 5a ics of a system with good handling quali-

* shows the standard rpm governor used in ties. The errors between the states ofthe ten-degree-of-freedom generic heli- the model and those of the base system arecopter mathematical model (from Ref. 1). fed into the control system, which attemptsThis model was modified for this simula- to minimize the state errors by generatingtion experiment based on actual DFVLR control signals for the actuators. If theBO-105 flight test data (Fig. 5b). Sev- state errors are always zero, the con-eral simplifying assumptions were made in trolled vehicle exhibits the dynamics ofderiving the block diagram shown in the model.Fig. 5b from the one in Fig. 5a:

The main objective of the designed1) The throttle is held constant at MFCS shown in Fig. 8 is to control the

100% attitudes to be followed in the outer loopby controlling the rates in the inner loop.

2) The required torque, OR , is The feedback matrices, G, and G2, are useddirectly proportional to the pilot's col- by the designer to select the desiredlective stick input, 6c outer-loop and inner-loop states to be

controlled. The elements of G, and G.3) The relationship between engine are either l's or O's, and each matrix

power, HP, and engine torque, OE, is contains a maximum of four elements equallinear so that the engine produces engine to 1 (the number of states to be followedtorque directly from the change in fuel must equal the number of controls avail-flow able). Once G, and G. are defined, the

controller matrix G, is dependent onlyExplicit Model on the base system's control matrix BD

for the four commanded variables. ThePrevious research programs3 - have controllers used are feed-forward gains

demonstrated that the handling qualities calculated using a nonreal-time identifi-of a helicopter flying typical mission cation procedure which incorporates ascenarios can be improved if the interaxis linearized version of the actual helicop-coupling is minimized. Interaxis coupling, ter. In addition, the elements of thesuch as pitch-roll cross-coupling and col- controller matrix are adjusted with air-lective input coupling to pitch and yaw, speed to improve the accuracy and robust-generally requires complex control coordi- ness of the system. The output of thenation by the pilot in terrain flight and controller matrix is integrated to sup-is, therefore, a major factor influencing press constant errors and hold the basethe flying qualities of the helicoptet. system in the trim state over the whole

flight envelope. It was found that inAccordingly, the explicit model cases in which either the rate or position

chosen for this experiment was a linear, limit of one of the four actuators isdecoupled model; the pilot commands pitch exceeded, the error between the controlledattitude with longitudinal cyclic, roll state and the actual helicopter state can

% attitude with lateral cyclic, yaw rate build up very quickly and cause instabili-with pedals, and Earth-fixed downward ties owing to saturated integrators. Thisvelocity with collective. A block diagram instability problem was overcome by stop-of the explicit model is shown in Fig. 6. ping the integration of the controller

. The feedback and feed-forward matrices, output for the axis in which the actuatorAM and BM , are diagonal matrices for a is limited by zeroing the correspondingdecoupled model. The model time constants element of the diagonal G, matrix.and sensitivities were determined in pre- Finally, the elements of the diagonal G,simulation checkout and adjusted, based on are used to convert the attitude errors topilot comments, to give good handling a specified amount of commanded rate.qualities. A special turn coordinationfeature was included in the model. The Initially, the controller matrix Gturn coordination feature attempts to mini- was equal to the inverse of the base hell-mize lateral acceleration in turns above copter's control matrix BD . However, for

. ....

tasks requiring large and rapid control Conduct of the Experimentinputs where the rate or position limitsof one or more of the actuators is Facility Descriptionexceeded, this linear control systemcauses overcontrolling in the unlimited This experiment was conducted on aaxes. For this case of limited actuators, fixed-base simulator at Ames Researchan optimal control law was developed: Center. The simulator facility includes

a single-seat cockpit cab equipped withTconventional helicopter controls and

uunlimited (BIB-BIllimited) typical instrument panel (Fig. 10).Switches in the simulator cab were used by

BD(i, 3 ): third column " matrix BD the pilots to enable the various featuresof the control system. A 600:1 scale ter-

(1) rain board and camera visual system depict-ing a runway with two 100-ft obstacles

In Eq. (1), BI denotes the base iys- positioned on the centerline (Fig. 11) wastem's control matrix corresponding to the presented through the cab window on aunlimited portion of the control vector, color television monitor with a collimat-The optimal control law minimizes all four ing lens.state errors using only the unlimited con-trols while holding the limited controls Evaluation Task Descriptionat either the maximum rate or position.The controller matrix G, is redefined as Evaluation of the model followingthe pseudo-inverse of the matrix B.: control system for the two helicopters

under investigation was accomplished usingT_-IT T two typical nap-of-the-Earth mission tasks:(BB'B () a dolphin (Fig. 12) and a slalom (Fig. 13)

task. The dolphin task required the pilotAn example calculation of the controller to fly over two 100-ft obstacles spacedmatrix G for the BO-105 in 60-knot 1150 ft apart while attempting to maintainlevel flight is given in the Appendix. a 60-knot airspeed and an altitude of

75 ft between the obstacles. The controlTo illustrate the difference between strategy the pilots were instructed to

the linear control system and the linear follow was to use primarily collectivecontrol system plus the controller opti- stick inputs to perform the task and tomization feature, Fig. 9 shows the response minimize deviations in airspeed, heading,of the BO-105, augmented with the two dif- pitch attitude, and roll attitude with theferent control-systems, to a 3-2-1-1 col- remaining three controls. The slalom tasklective stick input of 1.0-in. magnitude was a lateral avoidance task which requiredduring 60-knot level flight. Figure 9a the pilot to maneuver around 100-ft high

- shows the collective input and correspond- obstacles spaced 1150 ft apart on theing response of the collective actuator, runway centerline while maintaining alti-For this input, the MFCS commands a verti- tude (75 ft AGL) and airspeed (60 knots).cal velocity of ±8 ft/sec, filtered with afrequency of 1.67 rad/sec. In addition, Evaluation Pilots' Background andthe MFCS attempts to hold the three other Experiencestates at their 60-knot trim positionsthrough manipulation of their actuator Two DFVLR pilots participated as eval-positions. It can be seen that the collec- uation pilots in this simulation study. Ative actuator was rate-limited during a summary of their flight time is presentedconsiderable portion of the input. Fig- in Table 2.ure 9b shows the time histories for thefour helicopter states for both the linear Data Acquisitioncontrol system and the optimized linearcontrol system. As expected, errors Pilot evaluation data, time histories,between the desired model states and the and magnetic tape recordings of specifiedhelicopter states occur when the collec- flight parameters were collected. Var-tive actuator is operating at its rate ables recorded included pilot controllimit. Maximum errors of 10*/sec, inputs, actuator positions, model atti-40, and 30 in yaw rate, pitch attitude, tudes and rates, helicopter attitudes andand roll attitude, respectively, were rates, airspeed, torque, and rotor rpm.observed for the linear control system. The pilot evaluation data consist of

* The advantage of the linear control system Cooper-Harper handling qualities ratings?with the controller optimization feature and two additional ratings: one of pilotis apparent in the decrease in interaxis workload and stress and the other of over-coupling, particularly the collective-to- all quality of the task performance. Theseyaw-rate coupling, ratings were obtained using a modified

Cooper-Harper rating scale (Fig. 14),

"5 .

developed and used at the DFVLR (Ref. 8); control in flying the task. The pedalsthey are used to give more insight into were used as a secondary control to mini-the secondary factors affecting the pilot mize the collective-to-yaw-rate coupling.during the task. In addition to the three Longitudinal cyclic inputs were made toratings, the pilots were asked to provide minimize collective-to-pitch couplingwritten commentary to help identify those which necessitated lateral cyclic inputsaspects of the system that most heavily to compensate for pitch-roll coupling. Oninfluenced the ratings. the right are shown the four control posi-

tions for the BO-105 augmented with theTest Matrix MFCS performing the identical dolphin.

These time histories indicate that theThe test matrix for this experiment pilot needed only the collective control

is shown in Table 3. Evaluation of the to perform the task.baseline helicopter without the modelfollowing control system was done assuming A comparison of the BO-105's states,an ideal actuating system (no rate or with and without the MFCS, during theseposition limits). Generally, only one same dolphin evaluation runs is shown intask was performed by the pilot in a typi- Fig. 16. In general, the criterion forcal simulation session. Three evaluation acceptable model following performance isruns per pilot were collected for each based on the error between the commandedtask. As indicated in the test matrix, and measured response. Maximum errors ofthe pilots evaluated the baseline helicop- 5/sec in yaw rate, 7.20 in pitch attitude,ter (without MFCS) for both tasks first and 9.60 in roll attitude were observedand then evaluated the augmented helicop- for the unaugmented BO-105. These errorster (with MFCS}. reduced to 2.40 /sec, 1.8*, and 30 in yaw

rate, pitch attitude, and roll attitude,Other Experimental Considerations respectively, for the augmented BO-105.

In addition, the airspeed errors went fromA significant amount of simulation 5 knots in the uncontrolled case to nearly

time was allotted for pilot familiariza- 0 knots in the controlled case. Thistion with the simulation facility and result shows that without wind and turbu-simulated helicopters. In order to mini- lence, the desired flightpath and 60-knotmize pilot learning-curve effects, evalu- airspeed were held constant, even thoughation data were not collected until the those parameters were not commandedpilots demonstrated a consistent level of directly by the MFCS.proficiency with the system. In general,it took the pilots a great deal more time Figure 17 shows the difference in theto become familiar with the baseline heli- control system activity needed to decouplecopter than with the baseline helicopter the BO-105 and UH-lH for the 60-knot dol-augmented with the MFCS. Using the exten- phin task. The BO-105's actuator positionssive pilot familiarity and experience with are strikingly similar to the pilot's con-the actual MBB BO-105 helicopter, the trol activity for the uncontrolled BO-105dynamics of the simulated baseline BO-105 shown in Fig. 15. Following a controlwere adjusted to match the actual BO-105 strategy similar to that used by thebefore any evaluation data were -lected. pilots, the MFCS uses the collectiveThis adjustment was not done wi the simu- actuator as the primary control and thelated UH-1H since the pilots were not pedal actuator as a secondary control infamiliar with the characteristics of the performing the task. Longitudinal andUH-IH without its stabilizer bar. lateral cyclic actuators are used to mini-

mize interaxis coupling.

Results For the UH-lH, the activity of thefour series servos is shown since these

Since the decoupling effects of the are more demanding on the control systemmodel following control system are more than the parallel servos. Because thepronounced in the dolphin task, time his- UH-IH simulation model did not include thetories and pilot ratings of that task are stabilizer bar, the dynamics and controlla-used to illustrate the major results. bility of the helicopter were severelySimilar results and trends were observed degraded. The strong coupling betweenfor the slalom task. collective inputs and pitch rate exhibited

by the simulated uncontrolled UH-lH isFigure 15 illustrates the effects on very apparent in the response of the longi-

pilot's control activity for the 60-knot tudinal series servo position. Highdolphin task when augmenting the SO-105 activity in all four series servos waswith the MFCS. On the left are shown the required in order to decouple the stronglyrequired control positions for the unaug- coupled and extremely nonlinear dynamicsmented 80-105. As instructed, the pilot of this helicopter.used the collective stick as the primary

Table 4 summarizes the averaged pilot being a totally uncontrollable vehicle inratings in handling qualities, pilot work- the baseline configuration to one thatload, and task performance for the dolphin exhibited excellent handling qualities,task. These results should not be inter- allowing satisfactory task performancepreted as constituting a comparison of the with minimal pilot workload, with the addi-handling qualities of the BO-105 and the tion of the MFCS. This drastic change inUH-1H. Rather they should be interpreted ratings demonstrates the ability of theas a comparison of the baseline simulation designed model following control system tomathematical model with and without the decouple a helicopter.designed model following control system.

Both pilots commented on the lack of Conclusionsvisual information for height and rate ofdescent. They found it particularly dif- A new model following control tech-ficult to estimate the point over the nique has been developed for potentialhurdles at which the descent should be application to a variable-stability heli-initiated, copter. Presented in this paper are some

of the results of a fixed-base simulationBoth pilots considered the baseline experiment which evaluated the performance

UH-lH mathematical model uncontrollable and limitations of this control systemand unflyable (pilot rating of 10) for the when used to make large changes in the60-knot dolphin task. This result was response characteristics of two very dif-attributed to the high couplings and low ferent helicopters, the BO-105 and thedynamic stability of the simulated vehicle UH-lH. The helicopters were evaluated,without the stabilizer bar. The baseline with and without the MFCS, for a 60-knot.0-105 received only adequate ratings, a dolphin task and a 60-knot slalom task.result of the strong couplings, collective- The following general tendencies and con-to-yaw and collective-to-pitch, which made clusions are noted:it difficult to hold airspeed and heading.In addition, attempts by the pilot to 1) The changes in response character-minimize airspeed deviations between the istics resulted in significant improve-hurdles created lateral deviations and, ments in the performance and handlingtherefore, increased pilot workload, qualities of both vehicles, flying typical

nap-of-the-Earth mission tasks, because ofSatisfactory ratings in all three a substantial decrease in interaxis cou-

categories were achieved when the MFCS was pling and, therefore, less pilot controladded to both baseline helicopters. The activity and workload.improvement in the handling qualities isprimarily a result of the decrease in the 2) Optimization of the control lawspilot's control activity and, therefore, was required to improve the model follow-workload, as illustrated in Fig. 15. For ing performance in situations in which thethe BO-105, the addition of the MFCS actuators are position or rate limited.improved pilot ratings 3 points for han-dling qualities, 5 points for pilot work- 3) The flexibility and simplicity ofload, and 4.5 points for task performance, the model following control system designPilot ratings of 1 were achieved in all allows for quick and easy adaptation tothree categories for the augmented UH-lH. helicopters with different dynamic charac-The simulated UH-IH helicopter went from teristics and actuating systems.

Appendix

Control System Design Example: BO-105 in 60-knot Level Flight

Step 1: Equations of motion: i - Ax + Bu

'0.0322 -0.0210 2.5125 -32.1990 0.0018 -0.5693 0.0000 -0.03191-0.0534 -0.7974 1.1900 -0.2019 -0.0043 -1.6577 0.9670 -0.1329 I0.0135 0.0003 -4.0010 0.0000 -0.0025 2.2435 0.0000 0.1145 "

A 0.0000 0.0000 0.9995 0.0000 0.0000 0.0000 0.0000 0.0300 I0.0026 -0.0056 -0.5699 0.0061 -0.1440 -2.7438 32.1850 0.9100

-0.0046 -0.0021 -5.9050 0.0000 -0.0689 -10.0650 0.0000 0.14320.0000 0.0000 -0.0002 0.0000 0.0000 1.0000 0.0000 0.0063

,-0.0052 -0.0090 0.1708 0.0000 0.0402 -0.0085 0.0000 -0.8351

0.000 0000 -0.002 0.000 0000 1.000 .000 0006

% 4

Step 1 (contd) Acknowledgments

-0.7507 -0.0129 0.0994 0.02761 The authors gratefully acknowledge the-1.9859 0.3991 -8.4075 0.0277 contributions of simulation applications1.0243 0.0627 0.2892 0.0210 engineer, R. Leeper, and the services of

B 0.0000 0.0000 0.0000 0.0000 the two participating pilots, K. Sanders-0.0096 0.7165 -0.0505 -1.7734 and M. Roessing.-0.1913 2.3651 -0.1032 -1.0503

0.0000 0.0000 0.0000 0.0000

-0.0282 0.0298 0.2243 1.5009 References

'Talbot, P. D., Tinling, B. E.,Step 2: Discretize the equations of Decker, W. A., and Chen, R. T. N., "Amotion using the backward rectangular Mathematical Model of a Single Main Rotorz-transform: x(i + 1] = ADX(i] + BDu[ij Helicopter for Piloted Simulation," NASA

TM-84281, 1982.

"-0.2230 -0.0510 -0.0082 0.00861

-0.3074 0.0420 -1.4411 0.0128 2Baker, F. A., Jaynes, D. N., Corliss,0.1005 0.0417 0.0283 -0.0098 L. D., Liden, S., Merrick, R. B., and

- 0.0201 0.0084 0.0059 -0.0004 Dugan, D. C., "V/STOLAND Avionics System0.BD L0.0502 0.2363 -0.0161 -0.3411 Flight-Test Data on a UH-lH Helicopter,"-0.0518 0.1396 -0.0173 -0.0619 NASA TM-78591, 1980.-0.0104 0.0279 -0.0034 -0.0121-0.0015 0.0077 0.0414 0.2546 'Chen, R. T. N., "Unified Results of

Several Analytical and Experimental Studies- (I - A * T)-1 • B • T of Helicopter Handling Qualities in Visual

Terrain Flight," Helicopter Handling Quali-Sample time of HFCS T - 0.2 sec ties, NASA CP-2219, 1982.

-Pausder, H.-J. and Hummes, D.,Step 3: Retain only the four desired "Flight Tests for the Assessment of Taskrows of B to achieve a square matrix - Performance and Control Activity," Heli-convert the angle measurements from copter Handlin Qualities, NASA CP-2219,radians to degrees. 1982.

The resulting controller matrix is $Corliss, L. D. and Carico, G. D.,"A Preliminary Flight Investigation of

r0.3987 0.1428 -0.0808 0.011.. Cross-Coupling and Lateral Damping for Nap-0.1209 0.2655 0.0334 -0.0142 of-the-Earth Helicopter Operations," Pre-0 0.1015 0.0187 -0.3675 0.0598 print No. 81-28, 37th Annual National Forum0.0052 0.0128 0.0018 0.0336J of the American Helicopter Society,

- (r ;)T New Orleans, La., May 1981.

6Chen, R. T. N., "Selection of Somer -0.5 Rotor Parameters to Reduce Pitch-Roll

Coupling of Helicopter Flight Dynamics,"where r is a correction factor to opti- Preprint No. 1-6, National Specialistsmix* the iontrol activity versus the con- Conference on Rotor Systems Design of the

-." trol system performance. American Helicopter Society, Philadelphia,Pa., Oct. 1980.

3ep 4: For the case in which the collec- 7Cooper, G. E. and Harper, R. P., Jr.,tive actuator is operating at its rate "The Use of Pilot Ratings in the Evaluationlimit, modify OD of Aircraft Handling Qualities," NASA

TN D-5153, 1969.1 .5 11 0 .4 06 - .02 41-. 0.5937 1.6006 -0.69121 'Sanders, K., Pausder, H.-J., and

'1 "-0.3074 0.0420 0.0128 ] Humes, D., *Flight Tests and Statistical0.0850 0.4432 14.590 Data Analysis for Flying Qualities Inves-

tiqations," Paper No. 56, 6th EuropeanRotorcraft and Powered Lift Aircraft Forum,S*ep_5: The controller matrix is redefined Bristol, England, Sept. 1980.

as tim pseudo-inverse of B1

r 0.3571 -0.1037 -0.0879 -0.0042]

33 a /0.1375 0.2677 -0.0058 0.0129|10.0021 -0.0088 -0.0002 0.0339J

a r((BI )' B

--

BASIC SWASHCONTROL PLATHSYSTEM PLATE

.1 II CONTROLS HYDR.SPR

SAFETY-PILOT FORCE SYST

FBW

ACTUATOR MI ENGAGE

DISENGAGE

-I -FLY-BY.WIRESCONTROL

I SYSTEM

CONTROLSLITESIMULATION-PILOT

rT ,y CONTROLFORCE

_ION BOARD

POTENTIOMETER COMPUTER

Fig. 1 BO-105 S3 control system.

COMMANDED

MODEL CONTROL ACTUATORS TAT ANSEPEFOLLOWING RATE AND HELICOPTERSF~CNOLLOIG IPT 2526.6 " POSITION

-,SYSTEM S2 -9.5 2526.6 LMT

Fig. 2 Simulated BO-105 S3 actuator dynamics.

C.• h'* 4,. .. , -. ,:. . -.- . ,- . -- - . .. , , , .... . ,...,.. ... . . , .. . . .

*~ - .. . . . . . . . . .

STIC

CONTRO

SERIESSERVO

STTICK

- - FORCE PITCH AND ROLLSENSOR POSITION SENSORS

PARALLELROL SRV

DISCONNECTPIC

ROLDISCONNECT SFT

OL RESEARCH MAGNETICPARALLEL

SERVO RSEARCH BUNGEESRESEARCH MAGNETIC (PITCH AND ROLL)

BRAKE MPICH

Fig. 3 UH-lH V/STOLAND control system longitudinal and lateral cyclic controls.

PI LOT'S INPUT (S AND S p ONLY)

COTO S2 + 105.63 + 5685.0 LIPOITO HELICOPTER

PARALLEL SERVO

Fiq. 4 Simulated UH-IN V/STOLAND actuating system.

THROTTLE (ANADTI OO

JONIOFFITORUE REQUIRED)

-an ENGINE S50 wt GE n40 -INfnnSTn +z OERNOR KE-

11 + E G

n~3S2

MO+ S ENITRIX CNIN + 0.

*

a 0 CI 80 oj

Fig. 5 Pgr. a Geeic bodel. ted

+ .'- (it

*•.-.~.-.~~~.~ *7*i j7*r~rV ° w._. - .-. * % . •. ** _~ . •

-, .,

-..

I LO'S /MODEL'S STATE COTO A C T U A T IN G

IOSSTATE ERROR CONTROL SIGNAL ACTUATING SIGNAL BASE BASESYTE SSTM SYSTEM STATE

Fig. 7 Basic model following control system concept.

l e.+ .. i 2 + B A S E-O E G3. .+ HELICOPTER

as +" AND-,','.'.ACTUATORS

"X2.11

EXAMPLE: 80 - 106 IN 60 knot LEVEL FLIGHT; 1 9 " I , , 0 1o T tl o ,0 0 w . r | T

G [00 0 1 0 0 01 G2' [o 0 b 0 0 0 0 010 0 0 0 0 1 0 0 0 0 0 00 00 0 0 0 0 000000

G3 -[0.400 0. 143 -0.0382 0 :.011 04 .05 0 0 0S0121 0.266 0.037 _.014 [0 0.2 0 0

-0.102 0.019 -0.368 0.060 0 0 0.2 000o.o5 0.013 0.002 0.034 0 0 0 0.3]

as. G - [1 0 0 0-g=.Fi 0 1 0

v 0 0 0O 1

Fig. 8 Model following control system.

,.-.-

*,%*,.

*..-'_

- ...

10 10-

COLLECTIVE COLLECTIVESTICK ACTUATOR

POSITION. POSITION, 5

In. in.

0l 0f

25

EARTH-FIXEDVERTICAL 0VELOCITY.ftiSGa

-251.-- 45

PITCH _0___ATTITUDE, 0

dog

ROLLATTITUDE, 0d"

k YAW RATE,%dagAtc 0

(b) s-c-ft

Fig. 9 Comparison of linear control systems for BO-105 at 60 knots. a) 1-in. 3-2-1-1

collective stick input and collective actuator response; b) linear control system vsoptimized linear control system.

1%

.No

• 4 .

"1*

p.

- p. .l

-% ', •;..'; " ,",..-",,'- -- "- -- ' , ." -' "D'',.'.,'/ ., ',-' .'. ., .";.i-."", R ." " ,' ,.-,".!D,'.,,., ,," ',. ., r t .:. .

.4.41

AIRSPEED *60 knots

* 75

ALL DIMENSIONS IN ft

-Fig. 10 Simulator cockpit and instrument Fig. 12 Dolphin task.panel.

MW.

q4V AIRSPEED a 60knots- u 9'ALTITUDE a 75 ft

115 --- T

ALL DIMENSIONS IN ft

rig. 11 Terrain board model. Fig. 13 Slalom task.

e-'~~~. A:. E -.-,. ;JZ

-'p

IAOEOUACY FO SIElDTS EIOTER CHARACTERISTICS PILOT STRESS IN PERFORMANCE PILOOR SLFOR ECTED TASK SELECTED TASK OF THE TASK I RATING

EXCELLENT 0 NO PARTICULAR *TASK PERFORMANCE IIS OPTIMALI

GOOD NO PARTICULAR * GOOD ACCURACY 2

FAIR 0 MINIMUM 0 FAIR ACCURACY 3

' ' "JiYES

MINOR BUT ANNOYING @ LOW 0 MINOR INACCURACIES 1 4DEFICIENCIES4

SATISFACTORY IMPROVEMENT MODERATELY OBJECTIONABLE* MODERATE 0 FAIR INACCURACIES SSTSATRJUSTIFIED | DFICIENCIES M

VERY OBJECTIONABLE BUT 0 HIGH 0 HIGH BUT TOLERABLETOLERABLE DEFICIENCIES INACCURACIES 6

SIGNIFICANT DEFICIENCIES 0 VERY HIGH 0 DEFICIENCIES IN THlETASK PERFORMANCE f

ACCEPTABLE No IMPROVEMENT C CONSIDERABLE DEFICIENCIES a,%,,., ~NECESSARY /IE CONSIDERABLE DEFICIENCIES NESV NTETS EFRAC

NECESARYIN THE TASK PERFORMANCE

VERY CONSIDERABLE * MAXIMUM 0 TASK PERFORMANCEDEFICIENCIES IS NOT WARRANTED 9

YES

I T AS K

NO.F IMPROVEMENT VERY CONS DEABE TASK( PARTIAL. NOT.,PRACTICABLE [-]MANDATORY DEFICIENCIES f MAXIMUM FULFILLED 10

DECISIONS

Fig. 14 Pilot stress and task performance rating scale.

J

.1'.

I

a..-

10

COLLECTIVESTICK 5

POSITION. in.

0

PEDALPOSITION, in. 0

-3.5

LONGITUDINALCYCLIC STICK 0POSITION, in.

LATERAL

CYCLIC STICK 0

BASELINE 80-105 80-105 WITH MFCS

Fig. 15 Comparison of Pilot's Control activity: 60-knot dolphin.

25EARTH-FIXED

VERTICALVELOCITY,

ft/nc-25MUU

YAW RATE, 0

*PITCH___ATTITUDE, 0

dig

ROLLATTI1TUCE. 0

BASELINE 80-105 M105 WITH MFCS

rig. 16 Comparison of states: 60-knot dolphin.

a- a .,=i

.* * - .. ,. .~& . - a - - - -. -~6 *

*-0-

COLLECTIVE

ACTUATOR 5o0POSITION.

In. 0 -

3.5 1

a.-:

'9 OPEDALACTUATOR 0POSITION, 0

-6n. . -2 _ _ _ _ __ _ _ _ _ __ _ _ _LONGITUDINAL 2-

CYCLICACTUATOR 0 --__- -- 0POSITION.

in. - -2

-9.LATERAL 2-

CYCLICACTUATOR 0 0 . ..............

A POSITION. -. 4.-1 sc sac__ _ _ __ _ _

B0.105 UH-1H

(SERIES SERVOS)

Fig. 17 Comparison of model following control system activity: 60-knot dolphin.

o.,

- .6""1- I", l l," ., l - . ".b . "

S.

I+ .r " ' + - " I .. . . , t .

it" Ilk *i

*-'- -"".4' -,f ,.44 "i.

V, -- c tit

A

4-

Vv f

4*

L'1S .. 4 .. ;• , V .,i; - .',. . ...

4' .. .+ ... • -, . .

. ....~ 'I .; .. ,,,. . ,'

ii it '.;. 4'A k ' ,'

• .. * ,-J9 .;. - ,,

..- . ,.;. . . .+ . --

• . ., ,,, -,- ,

'- . , . . 4..

47.a . .. . +' . .

, . , . r"t.;

,-,,'-. .. ., , " ' ... ..

.. crS',.C,, ,

lis ,. * . - ..> . "- .1

::"; * - 4' % -- 3 ;.3 :d+ , - "A ". , -

IJ

• - . •~'I . - , . ,

p. 4 ' t ., "

-v .,- .,4 .'-. -. .. ......