In Flight Evaluation of Active Inceptor Force-Feel Characteristics and Handling Qualities

In Flight Evaluation of Active Inceptor Force-Feel

Characteristics and Handling Qualities

Jeff A. Lusardi, Chris L. Blanken, and MAJ Carl R. Ott Aeroflightdynamics Directorate (AMRDEC) US Army Research, Development, and Engineering Command Jeff.Lusardi@us.army.mil, Chris.L.Blanken@us.army.mil, Carl.Ott@us.army.mil Moffett Field, CA, USA Carlos A. Malpica NASA Ames Research Center Carlos.A.Malpica@nasa.gov Moffett Field, CA, USA Wolfgang von Grünhagen Deutsches Zentrum für Luft- und Raumfahrt e.V., DLR ± Institute of Flight Systems Wolfgang.Gruenhagen@dlr.de Braunschweig, Germany Abstract The effect of inceptor feel-system characteristics on piloted handling qualities has been a research topic of interest for many years. Most of the research efforts have focused on advanced fly-by-wire fixed-wing aircraft with only a few studies investigating the effects on rotorcraft. Consequently, only limited guidance is available on how cyclic force-feel characteristics should be set to obtain optimal handling qualities for rotorcraft. To study this effect, the U.S. Army Aeroflightdynamics Directorate working with the DLR Institute of Flight Systems in Germany under Task X of the U.S.

German Memorandum of Understanding have been conducting flight test evaluations. In the U.S., five experimental test pilots have completed evaluations of two Mission Task Elements (MTEs) from ADS-33E-PRF and two command/response types for a matrix of center-stick cyclic force-feel characteristics at Moffett Field. In Germany, three experimental test Pilots have conducted initial evaluations of the two MTEs with two command/response types for a parallel matrix of side-stick cyclic force-feel characteristics at WTD-61 in Manching. The resulting data set is used to correlate the effect of changes in natural frequency and damping ratio of the cyclic inceptor on the piloted handling qualities. Existing criteria in ADS-33E and a proposed Handling Qualities Sensitivity Function that includes the effects of the cyclic force-feel characteristics are also evaluated against the data set and discussed.

systems and active inceptors in helicopters, the force-feel Introduction characteristics are now determined by the closed-loop For most helicopters, the force-feel system characteristics response of the active inceptor itself as defined by the of the cyclic inceptors are set based on the characteristics inertia, force/displacement gradient, damping, breakout of the mechanical components in the control system force and detent shape configuration parameters in the (mass, springs, friction dampers, etc.). For these inceptor control laws. These systems give the flexibility helicopters, the force-feel characteristics typically remain to dynamically prescribe different feel characteristics for constant over the entire flight envelope, with perhaps a different control modes or flight conditions, and the ability trim release to minimize control forces while to provide tactile cueing to the pilot through the actively maneuvering. With the advent of fly-by-wire control controlled side-stick or center-stick cyclic inceptor. A number of studies have been conducted to assess the th Presented at the American Helicopter Society 68 Annual Forum, Fort impact of controller force-feel characteristics on the pilot- Worth, TX, May 1-3, 2012. This is a work of the U.S. Government and vehicle flying qualities in high performance fixed wing is not subject to copyright protection in the U.S. Approved for public fly-by-wire aircraft, primarily directed toward minimizing release; distribution is unlimited.

pilot induced oscillations and roll ratcheting [1][2]. There cyclic, pedals and collective. Due to the safety monitors has been much less research into the effects of force-feel on the aircraft, the acceleration, rate and attitude characteristics on rotorcraft handling qualities. A brief capabilities were limited necessitating the use of a overview of a few of these studies is given in the relatively benign sum-of-sines input compared to the input following paragraphs. used in other studies [1][2]. Although not in the published paper, the presentation by Watson and Schroeder showed One of the major elements studied by Boeing Vertol under a proposed requirement on the feel system characteristics.

the Army's Advanced Digital/Optical Control System The requirement set boundaries based on the cyclic natural (ADOCS) program was the pilot's integrated side-stick frequency and inertia, with the stipulation of a lower controller [3]. This simulation study looked at a range of damping limit of 0.3. An updated version of the force displacement gradients from stiff (40 lb/deg) to large requirement is published in [7].

deflection (0.6 lb/deg) with functionality ranging from 4- axis (lateral, longitudinal, directional and vertical) to 2- The second study was conducted by the Canadian Institute axis (lateral and longitudinal only) side sticks with pedals for Aerospace Research using their variable-stability Bell and left hand collective. This study provided valuable 205A helicopter [8]. This study evaluated isometric sticks insight into force-deflection characteristics and the and variations in damping ratio and natural frequency of number of axes controlled by the side-stick controller for displacement sticks. The pilots evaluated both a sum-of- the ADOCS demonstrator aircraft. However, the study sines tracking task and various low-speed maneuvering recommended provisions for evaluations of multiple tasks. One of the outcomes of this research was a controller configurations in the flight demonstration suggested boundary for stick dynamics based on natural aircraft due to differences between simulation and flight. frequency and damping ratio. While these two studies produced boundaries for acceptable/unacceptable stick More recently, Sikorsky Aircraft working on a Technical dynamics for rotorcraft, they were not able to provide Area of Joint Interest (TAJI) funded under the National guidance on how variations of the stick dynamics in the Rotorcraft Technology Center (NRTC) performed a acceptable region impact handling qualities.

simulation study to gather data in support of the Under Task X, Handling Qualities for Active Controlled development of handling qualities specifications for side- stick feel characteristics [4]. This study looked at Rotorcraft of the U.S. German Memorandum of variations of stick travel, breakout forces, damping and Understanding for cooperative research on helicopter force gradient (with fixed stick inertia) in Sikorsky's aeromechanics, the U.S. Army Aeroflightdynamics motion base simulator. The simulation model was based Directorate (AFDD) and the DLR Institute of Flight on early CH-53K control laws with both rate command Systems Germany are conducting an active inceptor attitude hold, and attitude command velocity hold control characteristics flight test study. In the U.S., AFDD is modes. This study provided valuable insight into the utilizing the Rotorcraft Aircrew Systems Concepts effects of changes in control travel and force gradient, but Airborne Laboratory (RASCAL) JUH-60A in-flight cautioned that the results should be validated and refined simulator with an active center stick, and in Germany in flight test before being incorporated into a future update DLR is utilizing their Active Control Technology/Flying of ADS-33E [5]. Helicopter Simulator (ACT/FHS) with an active side stick.

Evaluations of the ADS-33E Hover Mission Task Element Studies have been conducted to assess the effects of cyclic (MTE) and Slalom MTE are being performed on both force-feel characteristics in flight, two of which are aircraft with a common matrix of inceptor natural discussed herein. The first study was conducted on the frequencies and damping ratios with both attitude NASA/Army CH-47B variable-stability helicopter [6].

command and rate command response types. In the U.S., The aircraft was equipped with a programmable active evaluations have been completed on the RASCAL by four center stick and rate command, attitude hold response U.S. Army experimental test pilots (XPs) and one German types. The cyclic damping was varied, and the lateral military XP at Moffett Field California. In Germany, natural frequency was varied by varying the stick inertia preliminary flight tests have been conducted on the while keeping the stick gradient constant. The maneuver ACT/FHS at WTD-61 in Manching Germany by two performed by the evaluation pilot was a roll attitude German military XPs and one U.S. Army XP.

regulation task while the copilot flew the longitudinal Coverage of paper The pilot comments collected at the end of each configuration evaluation included the assignment of a This paper presents the results of a flight test study Cooper-Harper handling qualities rating (HQR) and conducted to collect data to investigate how changes in answers to a structured questionnaire about task cyclic inceptor force-feel characteristics effect piloted performance, aircraft characteristics and demands on the handling qualities, to evaluate existing handling qualities pilot. In addition, the pilots were asked to assign a criteria, and to provide a basis for developing new criteria numerical score from one to nine rating the level of that account for the cyclic inceptor force-feel precision obtainable, their ability to be aggressive, and on characteristics. A description of the RASCAL and ride quality where higher numerical scores were ACT/FHS as configured for these tests is presented, considered to be the best. The pilots were also asked to followed by an overview of the matrix of cyclic force-feel assign numerical scores for characteristics of the cyclic characteristics evaluated with both Attitude Command inceptor feel, forces, and response sensitivity where five (AC) and Rate Command (RC) response types. Results was considered to be optimal. This choice of quantities to include handling qualities ratings, a set of quantitative be rated and the adjectives used to describe the quantities ratings designed to augment the HQR scale, and pilot was based on recurring adjectives recorded during comments.

numerous other flight tests conducted by AFDD. The application of a numerical rating scale based on common The current ADS-33E short term response (bandwidth) adjectives allowed for quantitative analysis of the requirements as applied to response due to both otherwise qualitative comments.

displacement and force input is presented for the RASCAL center-stick configurations to assess the Center-Stick Cyclic Testing applicability of the current criteria when using cyclic force as the input. A comparison of results of the center stick The flight testing on the RASCAL with an active center evaluations of the Slalom maneuver against predicted stick cyclic was conducted by AFDD on the ADS-33 handling qualities levels from the Handling Qualities course at Moffett Field [10]. The RASCAL has a full- Sensitivity Function (HQSF) proposed in [9] is also authority, fly-by-wire research flight control system for the right seat evaluation pilot, while maintaining the presented to investigate the viability of the HQSF as a predictive tool. standard UH-60 mechanical controls for the safety pilot in the left seat [11]. The control laws used for the Conduct of Test evaluations were model following control laws with both RC and AC response types for the lateral and longitudinal Two Mission Task Elements (MTEs) from ADS-33E were axes and are described in detail in [12] [13]. The gains for chosen as evaluation maneuvers for this testing to study these control laws were optimized to provide Level 1 the effects of force-feel characteristics on handling handling qualities; the optimization did not consider the qualities, the Hover MTE for low-speed maneuvering cyclic force-feel characteristics. The control laws featured using small precise inputs, and the Slalom MTE for high- height hold in the vertical axis, heading hold at hover/low speed maneuvering when making large inputs. To the speeds and zero side-slip hold at high speeds in the extent possible, the testing always began with a baseline directional axes. This allowed the pilots to fly the Hover cyclic force-feel configuration. The pilot then performed and Slalom MTEs using only the cyclic. For the Slalom as many practice runs as desired to become familiar with MTE, velocity hold was enabled in the pitch axis which the task using the baseline configuration. The pilot then allowed the pilot to use only the lateral cyclic to conduct performed a minimum of three evaluation runs "for the the evaluations. Position hold was disabled for all record" and provided feedback. The remaining evaluations.

configurations were evaluated "blind", in a random order and rated by the pilot using the same procedure. At the The characteristics of the cyclic inceptor dynamics are end of the evaluations the pilot was allowed to go back defined by the inertia and the force displacement features and look at any of the configurations and update their shown in Figure 1. The inceptor displacement due to evaluation if desired. The pilot then ranked their order of force input can be modeled as a simple second order preference of the inceptor configurations for the MTE and system: response type being evaluated. Not all pilots were able to n disp k evaluate all configurations.

2 2 2 s s n n force where the k is the force gradient shown in Figure 1 and Mass k . A matrix of two undamped natural 2.3 lb/in 2.9 lb/in 5.9 lb/in n frequencies ( = 7 and 23 rad/sec) and two damping n ratios ( = 0.7 and 1.5) were selected that defined the 40 Mitchell boundaries of the test space to be evaluated. Within this Side Stk - lon Aponso Side Stk - lat (1995) Center Stk 0.75 lb/in Inertia (lbm) DEGRADED Width detent Watson Schroeder (1990) ACCEPTABLE 0 Force 25 15 109 8 7 6 5 4 gradient Inceptor Natural Frequency (rad/sec) F detent Figure 2. Boundaries on inceptor inertia and natural F breakout frequency from [5] and [6] Displacement Figure 1. Cyclic force displacement diagram ACCEPTABLE 1.75 test space, an additional interior point ( = 9 rad/sec, = n 0.9) that had been used for the testing in [13] was selected 1.5 A B as the baseline configuration. The force-displacement 1.25 gradient was set to 0.75 lb/in for all configurations and the F side stick inertia was adjusted to change the undamped natural frequency. For evaluations with AC, the breakout force F center stick was set to 0.1 lb and the detent was set to 1 lb with a width 0.75 D C of 0.14 inches. When the AC breakout and detent settings were evaluated with RC the detent was found to be 0.5 Morgan objectionable, so the detent was removed and the breakout (1990) was set at 1 lb. The lateral and longitudinal cyclic force- 0.25 UNACCEPTABLE feel characteristics were constrained to be equal for each configuration. The five cyclic configurations used for the 0 10 20 30 evaluations are shown in Table 1, plotted against the Inceptor Natural Frequency (rad/sec) boundaries from [6] [7] in Figure 2, and plotted against the Figure 3. Boundaries on inceptor damping ratio and boundaries from [8] in Figure 3. None of the center-stick natural frequency from [7] configurations fall in the degraded regions of either figure.

Table 1. RASCAL center-stick cyclic configurations a Config. Inertia Breakout Detent (rad/sec) n (lbm) (lb) (lb, in) AC RC AC RC A 5.9 7.0 1.5 0.1 1.0 1.0, 0.14 0, 0 B 0.6 23.0 1.5 0.1 1.0 1.0, 0.14 0, 0 C 0.6 23.0 0.7 0.1 1.0 1.0, 0.14 0, 0 D 5.9 7.0 0.7 0.1 1.0 1.0, 0.14 0, 0 F 3.4 9.2 0.9 0.1 1.0 1.0, 0.14 0, 0 a Inertia (lbm) is listed to be consistent with the criteria of Figure 2 Side-Stick Cyclic Testing for different capabilities of the human arm and wrist. The intent was to provided the pilot with a side stick that felt Flight test evaluations with an active side-stick were qualitatively the same laterally and longitudinally, and conducted by DLR at the Technical and Airworthiness symmetric about trim. The undamped natural frequencies Center for Aircraft (WTD 61) in Manching Germany on and damping of configurations A through D were selected the ACT/FHS [14][15][16]. The test aircraft research to be approximately the same as the configurations system features a full authority, four times redundant evaluated with the center stick on the RASCAL, but the (quadruplex) fly-by-light primary flight control system gradient and breakout were set to values more appropriate that incorporates a simplex experimental flight control for a side stick. For testing of the side stick, configuration computer. The control laws used for the evaluations were F was selected to be in the center of the test space. As a developed at DLR, and were also used during an Empire result, the damping and natural frequency of this Test Pilot School rotorcraft exercise on the ACT/FHS, to configuration differs from configuration F that was give the trainees the opportunity to tune their own control evaluated with the center stick on RASCAL (Figure 3).

laws. The control laws were AC in pitch and roll, with a The test procedure with the side stick was similar to the selectable RC in the roll axis. Rate command was not procedure used for testing with the center stick described selected in the pitch axis to reduce the need for in the previous section. A different pilot questionnaire compensation inputs to maintain velocity during was used for testing of the side stick, which contained maneuvers with large bank angles changes. The AC only a subset of the questions from the center stick testing.

control laws and the mixed RC roll, AC pitch control laws were predicted to provide Level 1 handling qualities. Results Altitude hold performance was dependant on setting the The results of the testing with the center stick and side collective trim position in the detent at the initiation of stick are presented in the following sections. As stated each evaluation, and occasionally required small earlier, the results with the side stick are preliminary corrections by the pilot to maintain altitude. Heading hold results so only the HQRs are presented herein. A more in was not available for these tests so the pilot had to depth analysis of the results from the testing on the manually maintain heading during the evaluations. ACT/FHS will be published in [17].

The side-stick force-feel configurations that were tested on Both the RASCAL and the ACT/FHS have safety the ACT/FHS are tabulated in Table 2. The side-stick monitoring systems that disengage the research flight longitudinal characteristics were symmetric about trim; control systems when limits on command input magnitude the lateral characteristics were set differently from the or rate are encountered. During evaluations of the Slalom longitudinal characteristics and were not symmetric about task the presence of these monitors forced the pilots to trim. The natural frequency of the lateral side stick constrain their technique to prevent trips of the safety reported in Table 2 and plotted in the following figures is monitors. This was more of a factor on the RASCAL than based on the average of the left and right natural on the ACT/FHS.

frequencies. The difference in longitudinal and lateral, All of the evaluations of the Hover MTE were performed and left and right force-feel characteristics are to account Table 2. ACT/FHS side-stick cyclic configurations a b Gradient Breakout Detent Config. Inertia n (lbm) (lb/in) (lb) (lb, in) (rad/sec) lat lon lat lon lat lon lat lon left right left right A 20.7 45.9 2.9 2.3 5.8 6.3 1.5 0.76 0.76 4.8, 1.7 3.8, 1.7 0, 0 B 1.9 4.2 2.9 2.3 5.8 25.1 1.5 0.76 0.76 4.8, 1.7 3.8, 1.7 0, 0 C 1.9 4.2 2.9 2.3 5.8 25.1 0.7 0.76 0.76 4.8, 1.7 3.8, 1.7 0, 0 D 20.7 45.9 2.9 2.3 5.8 6.3 0.7 0.76 0.76 4.8, 1.7 3.8, 1.7 0, 0 F 4.0 8.8 2.9 2.3 5.8 15.7 1.1 0.76 0.76 4.8, 1.7 3.8, 1.7 0, 0 a Inertia (lbm) is listed to be consistent with the criteria of Figure 2 b lateral natural frequency is average of left and right natural frequencies in the day (GVE) when the winds were 10 kt or less, and evaluations of the Slalom MTE were conducted when the winds were 15 kt or less. During the pilot debriefings B A 1.6 conducted at the end of each flight, none of the pilots considered winds to be a factor during any of the evaluations. In addition to HQRs, when available the 1.2 numerical ratings for ability to be precise, limitations on F ability to be aggressive, and ride quality collected for each configuration were fit to a regression plane model. A two- 0.8 way Analysis of Variance (ANOVA) was also performed D C on the data to test the statistical significance of the 0.4 influences of damping and natural frequency, and the interaction effect of the two. Results with a p-level of less than 0.05 were considered statistically significant, and with levels between 0.05 and 0.1 to be marginally 0 10 20 30 significant. When available, pilot ratings of the stick feel, (rad/sec) n forces and sensitivity were averaged and plotted with 95 Figure 4. Regression fit of HQRs for Hover, AC, percent confidence intervals.

center stick Hover MTE Evaluation, Center Stick (RASCAL) Q The results of the regression plane fits of the HQRs from evaluations of the Hover MTE for the five cyclic inceptor B A 1.6 configurations with AC are shown in Figure 4, and with RC in Figure 5. The lines of constant HQR are plotted with the numerical rating, and the arrow shows the 1.2 direction of improvement. The figures show that F configuration B produced the best average HQR (Level 1), and that configuration D produced the worst average HQR 0.8 (Level 2) for both AC and RC. The ANOVA showed that D C the influence of damping ( ) was significant, the influence 0.4 of natural frequency ( ) was marginally significant for n AC, and that the interaction effect of the two ( and ) n was not significant. The figures also show that for all 0 10 20 30 cyclic configurations there is almost one HQR improvement for AC compared to RC. (rad/sec) n Figure 5. Regression fit of HQRs for Hover, RC, These results are consistent with the results reported in center stick [13] which utilized configuration F. The main difference between the two tests is that position hold was enabled for rating, a higher numerical value is better, and the direction the results of [13] where position hold was disabled for of improvement is indicated by the arrow.

this test, forcing the pilot to maintain position using the The results show that the pilot's perceived ability to be cyclic during the entire 30 seconds of the Hover MTE.

precise was the best for configuration B, and worst for The pilots also rated their level of precision obtainable on configuration D. The results of the ANOVA showed that a nine point scale (one = low precision, and nine = high the influences of damping and natural frequency were precision), limitations on their ability to be aggressive significant for AC, but were not significant for RC.

(one = limited, nine = unlimited), and ride quality (one = Interaction effects were not significant for either AC or jerky, nine = smooth). The regression fit of the scores RC. The figures also show that configuration D has about from the precision rating for the Hover MTE with AC is the same precision rating for both AC and RC. While the shown in Figure 6, and with RC in Figure 7. For this B A 1.6 B A 1.6 1.2 1.2 F F 0.8 0.8 D C D C 0.4 0.4 0 10 20 30 0 10 20 30 (rad/sec) (rad/sec) n n Figure 8. Regression fit of aggressiveness rating, Figure 6. Regression fit of precision rating, Hover, AC, center stick Hover, AC, center stick B A 1.6 1.6 B A 1.2 1.2 F F 0.8 0.8 D C D C 0.4 0.4 0 10 20 30 0 10 20 30 (rad/sec) (rad/sec) n n Figure 9. Regression fit of aggressiveness rating, Figure 7. Regression fit of precision rating, Hover, Hover, RC, center stick RC, center stick regression plane fit for RC (Figure 7) shows only a small configuration D the lowest for both AC and RC. The change in numerical scores between D and B (relatively results of the ANOVA showed that interaction effects flat slope), the fit for AC (Figure 6) shows a much larger were not significant, that the influences of damping and change in the numerical scores (steeper slope) toward high the influence of natural frequency were significant for AC, precision for configuration B when performing the Hover but were not significant for RC. Again this is consistent MTE. with the relatively steep slope of the plane for AC, and the relatively flat slope of the plane for RC. It is interesting to The regression fit of the aggressiveness rating for the note that the ratings for both precision and aggressiveness Hover MTE with AC is shown in Figure 8, and with RC in for the best configuration (B) are rated much higher for Figure 9. The figures show that configuration B received AC than for RC.

the highest average aggressiveness rating and The regression fit of the ride quality rating for the Hover from evaluations of the Hover MTE with a center stick are MTE with AC is shown in Figure 10, and with RC in presented in Table 3. Statistically significant results are in Figure 11. The figures show that the pilots were able to bold, and the marginally significant results are shaded.

perceive an improvement in ride quality with increasing damping of the cyclic. For AC, the ANOVA results did Table 3. Summary of ANOVA p-levels, Hover, center not show that the influence of damping, natural frequency, stick or the interaction effect were significant. For RC, the ANOVA results showed that the influence of damping on Response Rating scale Main effect Interaction ride quality was marginally significant, and that the type x n n influence of natural frequency and the interaction effect HQR 0.016 0.057 1.00 were not significant. A summary of the ANOVA p-levels Precision 0.001 0.014 0.177 AC Aggressiveness 0.032 0.032 0.558 Ride quality 0.356 0.576 0.356 B A 1.6 HQR 0.124 0.518 0.012 Precision 0.432 0.563 0.495 RC Aggressiveness 0.165 0.407 0.407 1.2 Ride quality 0.067 0.156 0.257 F 0.8 Slalom MTE Evaluation, Center Stick (RASCAL) D C The regression plane fits of the HQRs for the Slalom MTE 0.4 for AC are shown in Figure 12, and for RC in Figure 13.

Both figures indicate that there is a slight improvement in handling qualities with increasing damping. However the 0 10 20 30 results of the ANOVA showed that neither the influence (rad/sec) n of damping, natural frequency, nor the interaction effect Figure 10. Regression fit of ride quality rating, Hover, were statistically significant for both AC and RC. The AC, center stick same is true for the precision ratings, and the aggressiveness ratings for both AC and RC. The improvement in ride quality rating with increasing damping was the only parameter for the Slalom that did show a statistically significant effect for AC and B A 1.6 marginally significant effect for RC.

The reasons for the difference in the effect of the 1.2 configurations on handling qualities between the Slalom F and Hover MTEs lie in the requirements of the maneuver, 0.8 which drive the character of the pilot inputs. Two metrics that can be used to characterize the pilot inputs are the D C RMS and cutoff frequency ( ) of the cyclic input time co 0.4 histories. The cutoff frequency is calculated from the autospectra of the pilot control time history, and is defined as the upper end of the frequency range that encompasses 0 0 10 20 30 one half of the total area under the curve. This parameter (rad/sec) is a direct measure of the pilot's operating frequency and n has been shown to be a good estimate of the piloted Figure 11. Regression fit of ride quality rating, Hover, crossover frequency [18] [19].

RC, center stick AC Hov er RC Hov er 1.6 B A AC Slalom 2.5 RC Salom 1.2 F 1.5 0.8 D C 0.4 Cutoff Frequency (rad/sec) 0.5 0 10 20 30 A B C D F (rad/sec) n Figure 14. Pilot lateral cyclic cutoff frequencies with Figure 12. Regression fit of HQRs, Slalom, AC, 95% confidence interval, center stick center stick AC Hov er RC Hov er B A 1.6 AC Slalom RC Salom 1.2 F 0.8 RMS (%) D C 0.4 0 10 20 30 A B C D F (rad/sec) n Figure 15. Pilot lateral cyclic RMS with 95% Figure 13. Regression fit of HQRs, Slalom, RC, confidence interval, center stick center stick The mean lateral cyclic cutoff frequencies are plotted in This indicates that the lower operating frequency and Figure 14 and the mean RMS are plotted in Figure 15 larger magnitude of the pilot inputs associated with the along with the 95 percent confidence intervals for all the Slalom MTE did not expose any significant benefit, or evaluations of the Hover and Slalom MTEs. Figure 14 deficiency of any inceptor configuration with respect to and Figure 15 show quantitatively that the pilots adopted a task performance. However, the improvement in ride high frequency/small amplitude control strategy for the quality with increased damping of the inceptor is an Hover MTE, and a low frequency/high amplitude control important result that can impact pilot fatigue and overall strategy for the Slalom MTE.

mission performance. A summary of the ANOVA p-levels from evaluations of the Slalom MTE with a center stick are presented in Table 4.

Table 4. Summary of ANOVA p-levels, Slalom, Feel: 1 slow -9 fast Force: 1 low -9 high center stick 8 Resp: 1 insensitive-9 overly sensitive Response Rating scale Main effect Interaction type x n n HQR 0.218 0.909 0.427 Precision 0.164 0.279 0.279 AC Aggressiveness 0.683 0.435 0.875 4 rating Ride quality 0.752 0.964 0.007 HQR 0.574 0.574 0.198 Precision 0.737 0.666 0.149 RC Aggressiveness 0.804 0.804 0.172 Ride quality 0.081 0.496 0.649 A B C D F Additional Pilot Ratings, Center Stick (RASCAL) Figure 16. Average cyclic qualitative ratings with 95% confidence interval, AC, center stick In addition to the ratings presented above, the pilots were asked to provide ratings of three qualitative parameters on Feel: 1 slow -9 fast Force: 1 low -9 high a nine point scale, with five being optimal. The parameters 8 Resp: 1 insensitive-9 overly sensitive were the feel of the cyclic (1=too slow, 9=too fast), the cyclic forces (1=too low, 9=too high), and the sensitivity (1 = too small of a response for a given input, 9 = too large of a response for a given input) with five corresponding to optimal. The intent of these questions was to expose objectionable characteristics (e.g. sluggish dynamics, excessive force gradients or response sensitivity gains) so they could be corrected early in the testing.

Figure 16 shows the average ratings from the combined 2 evaluations of Hover and Slalom MTEs with AC. The error bars on the plots correspond to the 95% confidence intervals of the mean values. The figure shows that 0 A B C D F configurations A and D with the lowest natural frequency were rated slightly slower than the other configurations.

Figure 17. Average cyclic qualitative ratings with 95% Configuration F was rated as fast as configurations B and confidence intervals, RC, center stick C, even though the natural frequency of configuration F is AC and RC, these results show that the feel rating tracked close to configuration D. The figure shows that in general the force-deflection gradient was rated near optimal for the inceptor natural frequency, and that the force- AC. It is interesting to note that configuration D was displacement gradient and stick sensitivities were perceived to have slightly lower than optimal forces and satisfactory for the experiment.

sensitivity even though the force-displacement gradient Hover MTE Handling Qualities Ratings, Side Stick and sensitivity did not change between configurations.

(ACT/FHS) The average qualitative ratings for RC are shown in Figure The regression plane fit of the HQRs for the Hover MTE 17. For RC, there appears to be more variation in the pilot conducted on the ACT/FHS with AC is shown in Figure ratings than there were for AC, in particular for 18. Once again, configuration B received the best overall configuration D. The feel ratings for RC follow a similar average HQR. In contrast to the results from the center trend to the trend observed for AC. In general for both ANOVA did not show that either of the influences of damping or natural frequency, nor the interaction effect were statistically significant for the evaluations with AC.

1.6 B A For the evaluations with RC, only the influence of natural frequency was statistically significant. Again, this may be a characteristic of side sticks that is attributable to the F 1.2 action of the wrist (versus the arm for center sticks), and could be influenced by the large inertia required to obtain 0.8 the desired natural frequency for configurations A and D.

D C 0.4 B A 1.6 0 10 20 30 (rad/sec) n F 1.2 Figure 18. Regression fit of HQRs, Hover, AC, side stick stick, the slope of the regression plane for the side stick 0.8 shows the greatest improvement in handling qualities with D C increasing natural frequency. This observation is 0.4 confirmed by the ANOVA analysis which showed that the influence of natural frequency was significant, but that neither the influence of damping, nor the interaction effect was significant. Results are not presented here from the 0 10 20 30 evaluations of the Hover MTE with RC due to the mixed (rad/sec) n mode and the limited number of pilots evaluations Figure 19. Regression fit of HQRs, Slalom, AC, side currently available.

stick The reduction of the influence of damping for the side stick as compared to the center stick may be related to the presence of the arm rest for the side stick [16], which would tend to stabilize the pilot's arm. The increased 1.6 B A influence of natural frequency for the side stick may be attributable to the use of wrist motion for the side sticks 1.2 F versus arm motion for center sticks. In addition, the inertia of the different stick configurations could also be a factor. For the center stick, all of the configurations fall in 0.8 the acceptable region of Figure 2. For the side stick, D C configurations B, C and F are in the acceptable region of Figure 2, while configurations A and D are in the 0.4 degraded region due to the large mass required to obtain a natural frequency of 6.3 rad/sec for these two configurations.

0 10 20 30 (rad/sec) Slalom MTE Handling Qualities Ratings, Side Stick n (ACT/FHS) Figure 20. Regression fit of HQRs, Slalom, RC, side stick The regression plane fits of the HQRs for the Slalom with AC is shown in Figure 19, and with RC in Figure 20. The Pilot Comments, Center Stick and Side Stick be noted that when flying the slalom task in the AFDD JUH-60A RASCAL the evaluation pilot control inputs had The two ADS-33E maneuvers flown were well suited for to be slightly restrained to avoid tripping the aircraft ¶ s revealing differences since they required different control internal lateral rate safety monitors on the Research Flight techniques: quick, small, precise inputs for stabilizing and Control System (RFCS) which would result in the RFCS maintaining the aircraft in a hover, and large, moderate to disengaging. Because of this, some pilots felt that they highly aggressive inputs for the slalom course. For the could have been more aggressive with their inputs in Hover MTE, the following inceptor characteristics were several of the stick configurations.

generally found desirable: Qualitatively, for flight maneuvers requiring larger, A light, quick feel sustained stick displacements, such as the slalom when Well damped to allow precise small inputs around trim flown in AC, the side stick configuration was preferred Little to no perceived delay in aircraft response to since the force required to hold the stick out of detent was control input less objectionable than with the center stick. Additionally the effects of bio-feedback were less perceptible with the There seemed to be benefits on increased damping such side arm controller due in part to the integrated arm rest that greater levels of precision were achieved resulting in providing a more stable platform for the pilot ¶ s arm. The the lowest workload. While the force gradient and asymmetrical lateral force characteristics of the side stick sensitivity of the inceptor remained unchanged between felt symmetrical in all but configuration B (low inertia, the different configurations, the pilots ¶ perception was that high damping) when flying high gain maneuvers.

inceptor configurations with lower natural frequencies presented a heavier feel, and were less sensitive making Pilot Ranking of Configurations the workload to capture and maintain a hover more difficult. An inceptor with the combination of a heavy As part of the questionnaire, the pilots were asked to rank feel and low damping (configurations D) provided the the configurations from best to worst for each command least precision, felt wobbly when making small rapid type (AC and RC) and each MTE (Hover and Slalom).

inputs, and was the most prone to over-controlling the Generally, the pilots' rankings correlated well with their aircraft.

HQRs. For the center stick, configuration B was rated as best by most pilots, and as second best by the remaining For the Slalom MTE, desirable characteristics were: pilots. This was the case for both response types and both MTEs. Configuration D was generally rated at or near the An inceptor that tracked well with the aircraft bottom. The exception was for the Slalom where one pilot movement (especially for AC) rated configuration D as best for RC and AC (B was the Little to no perceived delay in aircraft response to pilot's second choice) and another pilot rated it second control input best for AC (behind B).

Well damped to prevent over-controlling resulting in a jerky ride quality For the side stick evaluations of both response types and No susceptibility to bio-feedback MTEs, configuration B was again rated as best by most pilots and second best by all but one of the remaining Pilot preference varied somewhat between the pilots. This pilot rated configuration B as the third best configurations when flying the slalom task but the for Hover, RC. Again, configuration D was rated dominant factor that affected pilot perception was how th th or 5 ), with the best consistently near the bottom (4 precisely the aircraft tracked or responded to control rating of third best by one pilot for Slalom, RC.

inputs. Bio-feedback (aircraft vibrations being fed back through the pilot ¶ s arm into the inceptor) and its effect was Predictive Criteria more noticeable in the attitude command configuration since the lateral cyclic inputs had to remain displaced and Assessment of ADS-33 Criteria held from the detent in order to hold the desired aircraft The attitude bandwidths for pitch and roll as defined in attitudes. The lighter, less-damped configurations (configurations C) proved to be the most susceptible to ADS-33E section 3.3.2, small-amplitude pitch and roll bio-feedback interference with the slalom task. It should attitude changes, were calculated for the center-stick configurations evaluated on the RASCAL. This the same limits. The numbers in brackets are the average requirement states that pitch (roll) response to the HQRs from the Hover MTE for each center-stick longitudinal (lateral) cyclic control position inputs shall configuration. The differences between the two meet the specified limits. It also states that it is desirable displacement points on the plots are due to the location of to also meet the criteria for controller force inputs. The the control position measurement. The displacement roll bandwidths for AC are plotted in Figure 21, and the inputs for the CLAW are at the input to the control laws, pitch bandwidths for AC are plotted in Figure 22 with the where the inceptor inputs are the unfiltered outputs of the specified limits for UCE = 1 and fully attended operations. cyclic inceptor rotary potentiometers. The differences The roll bandwidths for RC are plotted in Figure 23, and between the two are primarily due to anti-alias filters on the pitch bandwidths for RC are plotted in Figure 24 with the inceptor signals.

Disp 0.4 0.4 Disp Level Level Force Force 3 Level [4.3] 0.3 0.3 Level [3.5] D [3.8] F Level [2.6] A [5.3] B C [3.6] 0.2 0.2 A p D F p [4.6] [inceptor] [3.4] C (sec) (sec) [4.0] B [inceptor] 0.1 0.1 [3.5] Level [CLAW] [CLAW] 1 2 3 4 5 1 2 3 4 5 (rad/sec) (rad/sec) BW BW Figure 21. Roll bandwidth from displacement input Figure 23. Roll bandwidths from displacement and and from force inputs, AC, center stick force inputs, RC, center stick 0.4 Disp 0.4 Disp Level Level Force [3.4] Force 2 [4.3] [5.3] [3.8] A D D 0.3 0.3 Level A Level F [3.5] [4.0] F B [2.6] [3.5] B C [3.6] [4.6] C 0.2 0.2 p p [inceptor] (sec) (sec) [inceptor] 0.1 0.1 [CLAW] [CLAW] 1 2 3 4 5 1 2 3 4 5 (rad/sec) (rad/sec) BW BW Figure 22. Pitch bandwidth from displacement input Figure 24. Pitch bandwidths from displacement input and from force inputs, AC, center stick and force inputs, RC, center stick All of the bandwidths and phase delays from the position The use of the structural pilot model for handling qualities inputs are solidly in the Level 1 regions of the criteria, prediction is predicated on the value of the Handling although there is a significant loss of bandwidth and Qualities Sensitivity Function (HQSF) increase in phase delay due to the anti-alias filters. All of the plots suggest that configuration C would confer the best ratings as the C inceptor characteristics impart the least reduction in the bandwidth, and the least amount of being contained within a prescribed set of boundaries additional delay. This prediction does not agree with the (Figure 26) over the typical frequency range of pilot HQRs from the Hover MTE which show that control, i.e., 1 ± 10 rad/sec. Here is the pilot configuration B confers the best HQRs. These results compensation in response to proprioceptive and vestibular show that all of the force points that fall in the Level 2 feedback, and is the proportional component of the region received Level 2 HQRs, however all the force visual compensation strategy. At its core the fundamental points in the Level 1 region did not receive Level 1 HQRs.

concept of the HQSF is to quantify the compensation This result indicate that the bandwidth/phase delay criteria required from proprioceptive and vestibular feedback, in should be evaluated using displacement inputs, and the response to a desired attitude command. Handling force-feel characteristics should be considered seperatly.

qualities predictions from this model approach can be tailored to be task dependent. This is achieved by Handling Qualities Sensitivity Function Criteria specifying the pilot crossover frequency, which An alternative method of predicting the handling qualities of the closed-loop system that includes the inceptor characteristics is the Handling Qualities Sensitivity Function (HQSF) proposed by Hess [9]. The block diagram shown in Figure 25 illustrates a mathematical Level 3 representation of the feedback structure employed by a pilot flying a compensatory lateral position tracking task Level 2 from [9]. This compensatory pilot-aircraft model is seen to rely on a structural pilot model for inner-loop attitude (roll) compensation. Sub-components for the structural 2 Level 1 pilot include neuromuscular dynamics ( ), proprioceptive and vestibular feedback ( and respectively), and the visual error compensation ( ). A key feature of this approach is the modeling of proprioceptive feedback, which accounts for the ability of 0 2 4 6 8 10 the pilot to make corrections to his control inputs based on Frequency (rad/sec) the perception of stick displacement. This allows for the Figure 26. Handling qualities prediction modeling of the inceptor dynamic response to pilot force boundaries inputs ( .

Figure 25. Compensatory feedback pilot model fundamentally serves as an independent tuning parameter Table 5. Comparison of assigned and predicted to the model. This assumes an a priori estimate of the handling qualities levels, Slalom, AC, center stick crossover frequency for a given task is available. Herein Experimental this was achieved a posteriori, from estimates of the Analytical Average Cooper-Harper rating crossover frequency based on the spectral analysis Config. A 2.8 (Level 1) (4 pilots) Level 1 computation of the pilot control input cutoff frequency Config. B 2.5 (Level 1) (4 pilots) Level 1 (Figure 14) to validate the methodology.

Config. C 3.4 (Level 1) (4 pilots) Level 1 As a starting point, selection of the pilot model sub- Config. D 3.0 (Level 1) (3 pilots) Level 1 component parameters was done to match the literature Config. F 3.2 (Level 1) (4 pilots) Level 1 [9]. Lower Order Equivalent System (LOES) models of the RASCAL RC and AC control modes were used for Table 6. Comparison of experimental and analytical handling qualities levels, Slalom, RC, center stick this analysis. The implementation of the proprioceptive loop is critical to the success of the HQ analyses, and the Experimental method is sensitive to selection of the proprioceptive Analytical Average Cooper-Harper rating feedback model, i.e., , or . At the Config. A 3.4 (Level 1) (3 pilots) Level 2 experimental cut off frequencies, aircraft dynamics for AC Config. B 3.0 (Level 1) (3 pilots) Level 1 indicate a transition between and .

Config. C 3.8 (Level 2) (3 pilots) Level 1 Results for AC shown below are based on the selection of Config. D 3.0 (Level 1) (2 pilots) Level 1 (i.e., a = 0) as the best approximation at the nominal Config. F 3.5 (Level 1) (3 pilots) Level 1 cut-off frequencies. Selection of the proprioceptive feedback model for RC was quite straightforward with K experimental handling qualities in the RC evaluations for representing a perfect fit of the aircraft dynamics over a two configurations in particular: configurations A and C.

wide range of frequencies. Selection of pilot low Configuration A received a Level 1 rating, whereas the frequency integral compensation can also have an impact structural pilot model HQ sensitivity function predicted on the analytical handling qualities prediction. Results Level 2 handling qualities. Conversely, configuration C, below assume zero integral compensation. Also, because which was predicted to be Level 1, was assigned validation of the handling qualities boundaries had been convincing Level 2 ratings.

achieved from fixed-base simulation [9], the vestibular feedback was assumed to be zero, which may not be Figure 27 illustrates the qualitative differences between appropriate for the current flight test activity.

the two configurations in question. The discrepancy with configuration A is actually seen to not be that significant.

Handling Qualities Sensitivity Function Predictions Assigned ratings were in actuality reflective of borderline Level 1-2 handling qualities. The Level 2 prediction for Comparisons against flight data showed that the proposed configuration A is based on the HQSF breaching the Level compensatory tracking model could adequately represent 1-2 boundary. This breach is, however, weak and far pilot inceptor input activity amplitude and general removed from the actual operating frequency of the pilot.

qualitative character for the slalom maneuver with a center The HQSF for configuration C, interestingly, is the stick. These results were encouraging and provided farthest away from the boundary, and presumably should increased confidence in the values selected for the have provided the best handling qualities. It is noted that structural model parameters.

evaluation pilots indicated a tendency for this configuration to be prone to biodynamic feedback in Table 5 and Table 6 summarize the results obtained for the response to the motion of the aircraft. The particular setup slalom task, with AC and RC control modes with a center of the structural model does not include the effect of stick. It is noted that, based on the cutoff frequency vestibular feedback, which could contribute to this approximations obtained from flight test data, the discrepancy.

structural pilot model predicts Level 1 handling qualities for all five active inceptor configurations in AC (Table 5).

Table 6 indicates discrepancies between the predicted and precision and aggressiveness ratings showed 6 improvement with increasing damping and increasing natural frequency. Only the results for AC proved to be statistically significant Level 3 ride quality improved with increasing damping and Config. A reduced natural frequency for both AC and RC, but Level 2 the improvements were not shown to be statistically significant no statistically significant interaction effect of 2 Level 1 damping and natural frequency were observed While results for evaluations of the Slalom MTE with a center stick showed similar trends, the results were not as Config. C significant. This is likely due to the task requirements of 0 2 4 6 8 10 the Slalom MTE which require the pilots to make larger Frequency (rad/sec) amplitude inputs at a lower frequency than for the Hover Figure 27. HQSF for configurations A and C, Slalom, MTE, reducing the benefits of a fast inceptor with high RC, center stick damping. The one result from the Slalom MTE that did Discussion show a statistically significant effect was the improvement in ride quality rating with increased cyclic damping for A flight test evaluation of the interaction between cyclic AC; the improvement for RC was marginally significant.

inceptor force-feel characteristics and handling qualities This could be an important consideration that would not has been conducted with a center stick cyclic on the normally be exposed through the use of the Cooper- RASCAL JUH-60A, and initiated with a side stick on the Harper scale alone.

ACT/FHS EC-135. In addition to collecting HQRs, a set of numerical ratings of how changes in inceptor In contrast to the center stick results, the side stick results characteristics affect the pilots ability to perform the task showed a tendency for improved HQRs with increasing were collected. The intent being to develop a set of natural frequency for both the Hover MTE (AC) and the ratings to complement the Cooper-Harper scale in cases Slalom (AC and RC). This could be attributable to the where the pilots have a strong preference for a one wrist action when controlling the side stick as opposed to configuration over another, yet the HQRs assigned to the arm action when controlling a center stick. This result two configurations are nearly the same. The results may also be influenced by the fact that the two presented in this paper support the criteria published in [6] configurations with the lowest natural frequency are [7] for cyclic inertia and damping, and in [8] for cyclic located in the degraded region of Figure 2.

damping and natural frequency in defining regions where For evaluations of the Hover MTE with a side stick, only the cyclic force-feel system will degrade handling results from AC were presented herein. Analysis of HQRs qualities.

resulted in the following observations: The results of the testing show that for center sticks AC provided almost one full HQR improvement over RC for increasing natural frequency resulted in a statistically corresponding cyclic configurations, which is consistent significant improvement in HQRs with the results presented in [13]. From evaluations of the the influence of damping was not strong and did not Hover MTE with a center stick, the following results were prove to be statistically significant observed: For evaluations of the Slalom MTE with both AC and RC, increasing the damping of the cyclic resulted in a similar trends were observed. The only statistically statistically significant improvement in HQRs for significant result was the improvement in HQRs with both AC and RC increasing natural frequency. A more in-depth analysis of the results of evaluations of the side-stick on the increasing natural frequency resulted in a nearly ACT/FHS will be presented in [17].

statistically significant improvement in HQR for AC (P=0.057), but not for RC The bandwidth and phase delay from force inputs were Conclusions calculated for each center stick inceptor configuration and Based on the results of these tests, the following plotted against the hover/low-speed short-term response to conclusions are made: control inputs requirement from ADS-33E. A comparison of HQRs from the Hover MTE for the corresponding 1) The effect of cyclic force-feel characteristics points on the plot showed that meeting the Level 1 have been shown to have a significant impact on boundary did not always result in Level 1 handling the handling qualities of rotorcraft.

qualities ratings. These results indicate that the current 2) For tasks that require high precision such as the guidance in ADS-33E is appropriate; the bandwidth/phase Hover MTE, Attitude Command was preferred delay criteria should be assessed using displacement as the over Rate Command in the good visual input, and checked using force input. To reliably include environment with a center stick.

the important effects of cyclic force-feel characteristics on 3) The damping of the center stick cyclic inceptor predicted handling qualities of rotorcraft, other approaches can have an impact on the pilot's perception of need to be investigated.

the aircraft ride quality.

4) Meeting the current ADS-33E Level 1 bandwidth To this end, an analytical closed-loop pilot-vehicle requirements from force inputs is necessary, but methodology [9] for predicting helicopter handling not sufficient to ensure Level 1 handling qualities was evaluated against data collected during qualities. Therefore, bandwidth requirements evaluations of the Slalom MTE with a center-stick cyclic.

should be assessed using displacement as the The helicopter control laws (altitude hold, velocity hold input.

and turn coordination) made this a lateral cyclic only task.

This consideration made the maneuver ideal for analysis Future Plans through this method. The HQSF approach, based on the structural model of pilot-vehicle coupling, was chosen Additional flight testing on the ACT/FHS with a side-stick because it encompasses some of the key physical elements inceptor is anticipated along with publication of a detailed of the proprioceptive feedback loop which is absent from analysis of the results. In the U.S., a simulation study is other analytical approaches: nominal neuromuscular planned in the NASA Ames Vertical Motion Simulator dynamics and simplified inceptor dynamics (including which will expand the current matrix of inceptor inertia, damping and force gradient). Generally good configurations and allow for evaluations of both a center- agreement between the analytical predictions and the stick and a side-stick inceptor. The study will also explore flight test results was achieved, particularly for the AC the use of tactile cueing to provide the pilot feedback configurations. Discrepancies in the results for the RC when approaching the actuator command and rate limits of configurations could be attributed to the absence of the RASCAL safety monitors.

vestibular feedback in the model setup.

Acknowledgements However, other questions regarding the feedback model setup remain to be addressed. Validity of the HQ Level This work was conducted under the U.S./German boundaries in the literature could also be questioned, for Memorandum of Understanding. This agreement allowed example. These proposed boundaries have been predicated the U.S and German PIs to seamlessly conduct this on the use of very specific set of parameter values for the cooperative research study which benefited greatly from structural pilot model. Deviation from any of these the exchange of evaluation pilots and engineers. The parameters would render the HQ predictions invalid based authors would like to thank all the individuals that on these boundaries. An example of one such parameter contributed to this research effort; the XPs from the U.S.

would be the pilot low frequency integral compensation and Germany, the RASCAL team and L3 Vertex ground mentioned above. Selection of this parameter was and aircrew support at AFDD, and the ACT/FHS team arbitrary; however, it can have a direct impact on the from DLR and the flight support personal at WTD 61.

analytical handling qualities prediction. Further analysis Thanks also go to Professor Ron Hess from UC Davis for will be required to specifically isolate appropriate values his consultation about the HQSF modeling approach and for this and other configuration parameters.

insights into the interaction of the human operator with the inceptors.

References "Flight Research Qualification of the Army/NASA Rascal Variable-Stability Helicopter,". Montréal, [1] Johnston, D. E., Aponso, B. L. "Design Considerations Canada: Presented at the 58th annual forum of the of Manipulator and Feel System Characteristics in Roll American Helicopter Society, May, 2002.

Tracking". NASA CR-4111, Feb,1988.

[12] Mansur, M. H., Lusardi, J. A., Tischler, M. B., and [2] Bailey, R. E., Knotts, L. H. "Interaction of Feel System Berger, T. ³ Achieving the Best Compromise Between and Flight Control System Dynamics on Lateral Flying Stability Margins and Disturbance Rejection Qualities". NASA CR-179445, Dec, 1990.

Performance ´ . Grapevine, TX: Presented at the 65th annual forum of the American Helicopter Society, [3] Landis, K., Glusman, S. "Development of ADOCS May, 2009.

controllers and control laws: Vols 1-3".

USAAVSCOM TR 84-A-7, Mar, 1987.

[13] Fujizawa, B. T., et al. "Response Type Tradeoffs in Helicopter Handling Qualities for the GVE". Virginia [4] Greenfield, A., Sahasrabudhe, V. "Side-Stick Force- Beach, VA: Presented at the 67th annual forum of the Feel Parametric Study of a Cargo-Class Helicopter".

American Helicopter Society, May 3-5, 2011.

Virginia Beach, VA: Presented at the 67th annual forum of the American Helicopter Society, May 3-5, [14] Kaletka, J., H. Kurscheid, and U. Butter. "FHS, the 2011.

New Research Helicopter: Ready for Service".

Presented at the 29th European Rotorcraft Forum, [5] Anon. "Handling Qualities Requirements for Military Friedrichshafen, Germany, Sep, 2003.

Rotorcraft, ADS-33E-PRF". U.S. Army Aviation and Missle Command, Mar, 2000.

[15] Hamers, M., Lantzsch, R., Wolfram, J. "First Control System Evaluation of the Research Helicopter FHS".

[6] Watson, D. C., Schroeder, J. A. "Effects of Stick Kazan, Russia: Presented at the 33rd European Dynamics on Helicopter Flying Qualities". AIAA-90- Rotorcraft Forum, Sep 11-13, 2007.

3477-CP, Presented at the AIAA Guidance, Navigation and Control Conference, Aug, 1990.

[16] von Grünhagen, W., Müllhäuser, M., Abildgaard, M., Lantzsch, R. "Active Inceptors in FHS For Pilot [7] Mitchell, D. D., Aponso, B. L., Klyde, D. H. "Feel Assistance Systems". Paris France: Presented at the Systems and Flying Qualities". AIAA-95-3425-CP, 36th European Rotorcraft Forum, Sep 7-9, 2010.

Presented at the AIAA Atmospheric Flight Mechanics Conference, Baltimore, MD, Aug 7-10, 1995.

[17] von Grünhagen, W., Schönenberg, T., Lantzsch, R., Müllhäuser, M., Fischer, H., Lee, D. "Handling [8] Morgan, M. J. "An Initial Study into the Influence of Qualities Studies Into the Interaction Between Active Control Stick Characteristics on the Handling Sidestick Parameters and Helicopter Response Types".

Qualities of a Fly-By-Wire Helicopter". AGARD-CP- Amsterdam, The Netherlands: To be presented at the 508, Feb, 1991.

38th European Rotorcraft Forum, Sep 4-7, 2012.

[9] Hess, R. A., Zeyada, Y., Heffley, R. K. "Modeling and [18] Atencio, A. "Fidelity Assessment of a UH-60A Siumlation for Helicopter Task Analysis". Vol. 47, No.

Simulation on the NASA Ames Vertical Motion 4, pp. 243-252: Journal of the American Helicopter Simulator". N ASA Technical Memorandum, 104016, Society, Oct 2002.

Sep, 1993.

[10] Kasper, E. F. "Design and Establishment of a [19] Tischler, M. B., Remple, R. K. "Aircraft and Permanent Aeronautical Design Standard-33 Low- Rotorcraft System Identification, Engineering Methods Altitude Rotorcraft Flight Test Facility". AMRDEC with Flight Test Examples". Reston, Virginia: Technical Report RDMR-AF-10-02, Apr, 2010.

American Institute of Aeronautics and Astronautics, Inc., 2006.

[11] Moralez, E., Hindson, W. S., Frost, C. R., Tucker, G.

E., Arterburn, D. R., Kalinowski, K. F., and Dones, F.