Document
NASA/TM-2002-210729
An Overview of Flight Test Results for a
Formation Flight Autopilot
Curtis E. Hanson, Jack Ryan, Michael J. Allen, and Steven R. Jacobson NASA Dryden Flight Research Center Edwards, California August 2002 The NASA STI Program Office...in Profile CONFERENCE PUBLICATION.
Since its founding, NASA has been dedicated to the advancement of aeronautics and space Collected papers from scientific and science. The NASA Scientific and Technical technical conferences, symposia, seminars, Information (STI) Program Office plays a key or other meetings sponsored or cosponsored part in helping NASA maintain this by NASA.
important role.
SPECIAL PUBLICATION. Scientific, technical, or historical information from The NASA STI Program Office is operated by NASA programs, projects, and mission, Langley Research Center, the lead center for often concerned with subjects having NASA's scientific and technical information.
substantial public interest.
The NASA STI Program Office provides access to the NASA STI Database, the largest collection TECHNICAL TRANSLATION. English- of aeronautical and space science STI in the language translations of foreign scientific world. The Program Office is also NASA's and technical material pertinent to institutional mechanism for disseminating the NASA's mission.
results of its research and development activities.
These results are published by NASA in the Specialized services that complement the STI NASA STI Report Series, which includes the Program Office's diverse offerings include following report types: creating custom thesauri, building customized databases, organizing and publishing research TECHNICAL PUBLICATION. Reports of results...even providing videos.
completed research or a major significant phase of research that present the results of For more information about the NASA STI NASA programs and include extensive data Program Office, see the following: or theoretical analysis. Includes compilations of significant scientific and technical data • Access the NASA STI Program Home Page and information deemed to be of continuing at http://www.sti.nasa.gov reference value. NASA's counterpart of peer-reviewed formal professional papers but • E-mail your question via the Internet to has less stringent limitations on manuscript help@sti.nasa.gov length and extent of graphic presentations.
• Fax your question to the NASA Access Help TECHNICAL MEMORANDUM. Scientific Desk at (301) 621-0134 and technical findings that are preliminary or of specialized interest, e.g., quick release • Telephone the NASA Access Help Desk at reports, working papers, and bibliographies (301) 621-0390 that contain minimal annotation. Does not contain extensive analysis.
Write to: NASA Access Help Desk CONTRACTOR REPORT. Scientific and NASA Center for AeroSpace Information 7121 Standard Drive technical findings by NASA-sponsored contractors and grantees. Hanover, MD 21076-1320 NASA/TM-2002-210729
An Overview of Flight Test Results for a
Formation Flight Autopilot
Curtis E. Hanson, Jack Ryan, Michael J. Allen, and Steven R. Jacobson NASA Dryden Flight Research Center Edwards, California National Aeronautics and Space Administration Dryden Flight Research Center Edwards, California 93523-0273 August 2002
NOTICE
Use of trade names or names of manufacturers in this document does not constitute an official endorsement of such products or manufacturers, either expressed or implied, by the National Aeronautics and Space Administration.
Available from the following: NASA Center for AeroSpace Information (CASI) National Technical Information Service (NTIS) 7121 Standard Drive 5285 Port Royal Road Hanover, MD 21076-1320 Springfield, VA 22161-2171 (301) 621-0390 (703) 487-4650 AN OVERVIEW OF FLIGHT TEST RESULTS FOR A FORMATION FLIGHT AUTOPILOT Curtis E. Hanson,* Jack Ryan, ? Michael J. Allen,_ and Steven R. Jacobson § NASA Dryden Flight Research Center Edwards, California C controller Abstract CMD command The first flight test phase of the NASA Dryden Flight e error signal Research Center Autonomous Formation Flight project has successfully demonstrated precision autonomous ERR error station-keeping of an F/A-18 research airplane with a second F/A-18 airplane. Blended inertial navigation FILT filtered system (INS) and global positioning system (GPS) measurements have been communicated across an G low-pass filter air-to-air telemetry link and used to compute GPS global positioning system relative-position estimates. A precision research formation autopilot onboard the trailing airplane I integral controls lateral and vertical spacing while the leading INS inertial navigation system airplane operates under production autopilot control.
Four research autopilot gain sets have been designed and K gain flight-tested, and each exceeds the project design requirement of steady-state tracking accuracy within 1 nl GPS relative-position error standard deviation of 10 ft. Performance also has been /12 INS relative-position error demonstrated using single- and multiple-axis inputs such as step commands and frequency sweeps. This N navigation report briefly describes the experimental formation Nz normal acceleration flight systems employed and discusses the navigation, guidance, and control algorithms that have been P position flight-tested. An overview of the flight test results of the formation autopilot during steady-state tracking and PSFCC production support flight control computer maneuvering flight is presented.
s Laplace operator Nomenclature SRA Systems Research Aircraft t time AFF Autonomous Formation Flight V velocity ARTS Airborne Research Test System x 1 GPS relative-position measurement x 2 INS relative-position measurement *Aerospace Engineer.
"_Aerospace Engineer.
X state SAerospace Engineer.
§Aerospace Engineer.
z true relative position Copyright © 2002 by the American Institute of Aeronautics and Astronautics, Inc. No copyright is asserted in the United States under uncertainty Title 17, U.S. Code. The U.S. Government has a royalty-free license d) bank angle to exercise all rights under the copyright claimed herein for Governmental purposes. All other rights are reserved by the copyright A estimate owner.
American Institute of Aeronantics and Astronautics Introduction A total of 167 test points have been accomplished in 11 research flights over a period of 2 months. This report briefly describes the experimental formation A formation autopilot capable of precise flight systems employed and discusses the navigation, station-keeping is required for applications such as guidance, and control algorithms that were flight-tested.
formation flight drag reduction. Formation flight shows An overview of the flight test results of the formation significant promise to improve efficiency through autopilot during steady-state tracking and maneuvering cooperative aircraft operations. In a manner similar to flight is presented.
migrating flocks of birds, aircraft in formation can take advantage of the vortex upwash created by the aircraft Note that use of trade names or names of ahead of them, allowing them to operate more manufacturers in this document does not constitute an efficiently. Flight tests that showed drag reduction by official endorsement of such products or manufacturers, measuring power reduction were first performed in either expressed or implied, by the National Aeronautics Germany in 1990.1 Beukenberg and Hummel and Space Administration.
successfully tested an optimization technique and measured an average power reduction of approximately Research Objectives 10 percent, l' 2 One limitation of their approach was a lack of intership communication that prevented The ultimate goal of the AFF project has been to bring coordination among the aircraft within the formation.
AFF technology and its associated drag reduction benefits to a readiness level that will be attractive to commercial cargo operators and the military. Additional The objective of the NASA Dryden Flight Research applications of AFF technology include autonomous Center (Edwards, California) Autonomous Formation aerial refueling, low-visibility formation separation Flight (AFF) project has been to develop and flight-test assurance, and uninhabited combat air vehicle pack and a formation autopilot system capable of achieving swarm operations.
sustained drag reduction by exploiting robust intership communication and precise formation navigation and The objective of the initial phase (Phase 0) of the AFF control. The AFF project had three planned phases; research has been to reduce the programmatic risks for beginning with autonomous station-keeping without achieving drag reduction through autonomous drag reduction, then proceeding to the piloted mapping formation flight. Although pilots regularly achieve very of wingtip vortex locations and effects, and finally accurate formation spacing, an automatic system is culminating with the testing of an autonomous drag desirable for reducing pilot workload, particularly on reduction system. The project, however, was cancelled long-endurance missions. Data were sought regarding shortly after completion of the second phase.
the feasibility of using GPS for formation navigation and the achievable control precision of a formation The first phase of the AFF project, named "Phase 0," autopilot.
has successfully demonstrated precision autonomous station-keeping of two aircraft in flight. Although flight Phase 0 of the experiment evaluated a wide range of within the vortex was not the intention of this phase of formation autopilot performance and robustness levels the program, the station-keeping flight tests discussed to provide guidance to future formation flight control herein are the first steps in that development. Two system designers. Flight test measurements also helped NASA Dryde_owned F/A-18 (McDonnell Douglas validate and improve the project design tools in Corporation, now The Boeing Company, St. Louis, preparation for the next phases of the project. Much of Missouri; and Northrop Corporation, now Northrop the systems integration effort accomplished and tested Grumman, Newbury Park, California) research aircraft by NASA Dryden during Phase 0 was also to be used were outfitted with identical global positioning system for the planned drag reduction tests.
(GPS) receivers and an air-to-air telemetry system for intership communication. In addition, the trailing Design Requirements airplane was equipped with an experimental precision formation autopilot control system. Using GPS and The formation autopilot controlled only the lateral inertial measurement data from both aircraft, the and vertical motion of the airplane because hardware restrictions limited the number of available command experimental system computed the equivalent of pilot stick commands to fly the trailing airplane to the desired outputs to two. The primary design goal has been to position within the formation. achieve precision tracking during steady, level flight.
American Institute of Aeronantics and Astronautics Performance requirements for the formation autopilot Research Systems Overview specify that the formation control accuracy be within 1 standard deviation of 10 ft. This position-error budget is The research systems primarily have been developed partitioned into two categories: navigation uncertainty, from existing capabilities at NASA Dryden and and controller performance. These two random and integrated onto the two F/A-18 research aircraft.
uncorrelated error sources are combined as the root of Elements from past and current research experiments the sum of their squares: include instrumentation systems on both aircraft, an air-to-air telemetry system, a research computer, a flight crew interface, and a set of specially modified F/A-18 £_N 2 = + ec (1) flight control computers. This approach reduced cost and development time for the experiment.
where e is the total uncertainty, eN is the navigation uncertainty, and ec is the uncertainty in controller Research Aircraft Description performance. Navigation uncertainty has been allocated Figure 1 shows the two NASA Dryden F/A-18 aircraft ±4 ft based upon the results of preliminary piloted used for this experiment. A two-seat chase-support risk-reduction flight tests. The remaining approximately airplane equipped with a GPS receiver and a telemetry ±9 ft of position error is the performance design goal of system served as the leading airplane; the Systems the formation autopilot.
Research Aircraft (SRA) served as the trailing one. In Three levels of stability robustness requirements addition to its research instrumentation, GPS receiver, and telemetry modifications; 5 the SRA was outfitted (table 1) have been defined to allow the full range of formation autopilot performance levels to be evaluated. with the experimental formation flight systems described in the following section. A third NASA The high and medium categories are based upon the military specification 3 robustness requirements for Dryden chase-support airplane was occasionally used for video and photographic documentation of the modes greater and less than 0.06 Hz, respectively. The low category is consistent with the minimum experiment.
flight-demonstrated margins for the X-29 aircraft. 4 Table 1. Stability robustness requirements.
Robustness Gain margin, Phase margin, level dB deg High 6.0 45 Medium 4.5 30 Low 3.0 20 The NASA Dryden F/A-18 "Class B" flight envelope defines operational limits within which simulation analysis has shown that any combination of fully deflected control surfaces will not cause structural EC01-0050-9 damage to the aircraft nor put the aircraft in an unrecoverable condition. This envelope is bounded by a Figure 1. The NASA F/A-18 AFF research aircraft in maximum airspeed of 250-kn calibrated airspeed and an formation flight.
altitude range from 20,000 to 32,000 ft. A single design condition has been selected within this envelope: Mach 0.56 and an altitude of 25,000 ft. This point is in the Formation Flight Systems Overview middle of the Class B altitude range and is at a reasonably high dynamic pressure for good aircraft The formation flight systems were located on the controllability. The single-point design has been trailing airplane and integrated an air-to-air telemetry evaluated in the simulation and cleared for operation receiving system and a research flight control system.
within an altitude band of 20,000 30,000 ft at a constant The telemetry system was used for intership indicated airspeed. communication, and the control system computed American Institute of Aeronantics and Astronautics cards. 6 These cards applied a 40-Hz low-pass filter to formation station-keeping commands. These commands replaced the pilot stick commands in the F/A-18 each command and multiplexed them into four identical inner-loop control system. Other than the signals. The signals interfaced with the four channels of quadruply-redundant F/A-18 flight control computers, the production support flight control computers the experimental systems were single-string. Figure 2 (PSFCCs) and to a pilot display. These PSFCCs are shows the formation flight systems. specially modified F/A-18 flight control computers 7' 8 with software that allows selection between the pilot's pitch and roll stick commands and the external ARTS commands. A pushbutton display unit located in the aft cockpit of the trailing airplane and connected to the Air-to-air _,.: Analog telemetry multiplex ARTS allowed the flight test engineer to monitor the receiver _: and filter status of the ARTS and change various parameters of the = cards GPS _= ARTS formation autopilot, such as spacing commands and feedback gains.
i Analog ," Push /LJ_ multiplex PSFCCs button and filter I-i,/I (quadruply IIII Formation Autopilot display _z-v I cards VI redundant)_Jl A precision formation autopilot (fig. 3) has been
unit
Aircra Actuator data ----J commands designed to control the lateral and vertical separation of 020158 the trailing airplane with respect to the leading airplane in the formation. This control is accomplished by Figure 2. The trailing airplane formation flight systems.
replacing the pilot control stick commands to the F/A-18 inner-loop control laws with the formation autopilot commands. The separation distance between the nose of the trailing airplane and the tail of the Inertial state information and GPS position data telemetered from the leading airplane were received and leading one is regulated by the pilot of the trailing one through throttle adjustments.
decoded by the Airborne Research Test System (ARTS) computer onboard the trailing airplane. A real-time embedded system, the ARTS also contained the A leader-follower guidance and control approach is formation autopilot software that calculated pitch and used. In this approach, the leading airplane has no roll stick commands to maintain the desired position of responsibility for maintaining the formation and is the trailing airplane within the formation. The merely required to pursue an independent flightpath.
single-string analog pitch and roll commands were sent The trailing airplane is solely responsible for making all from the ARTS to a set of analog multiplex and filter of the required maneuvers to achieve the desired Controller :l I Plant II
"
+ K"_ PERR :l r-- p ; _ JI F/A-18 CMD Jl , "W , ,otu ,o, _- l I_i _ _ / ll-I control - ,_0"_-i ;,i;U_'ff / _ _,. _ /- Ll' • / i i I system uommanos / i _1 -v I_1-' I ? / I I I_ ........................ J I i ' ' / L .o.on ro, s, ic.
;-b -_ ................. J NzJ, Relative velocity Formation Inertial velocity guidance Relative position and GPS position i_ GPS position navigation Inertial velocity 020159 Figure 3. Formation autopilot.
American Institute of Aeronantics and Astronautics position behind the leading airplane; formation control A second relative-position measurement between the is applied only to the trailing one. For design purposes, two aircraft is obtained by differencing and then the operational concept for the formation autopilot is integrating their inertial navigation system (INS) restricted to straight-and-level flight; however, some velocities at 20 Hz. The primary errors in this flight tests have been performed in dynamic conditions measurement are low-frequency drift caused by biases to fully evaluate the limitations of the system. in the INS velocities. The inertial position measurement therefore can be represented as Formation Guidance and Navigation x2(t ) = z(t) + n2(t ) (3) Relative positions are calculated in a formation reference frame (fig. 4) that is fixed in alignment with where x2(t ) is the INS-measured position, z(t) is the true the heading selected by the flight test engineer. This position, and n2(t ) is the low-frequency error caused by approach minimizes lateral spacing errors caused by integration drift.
fluctuations in one or both of the aircraft headings, a problem that can be magnified by increasingly large The two independent relative-position measurements nose-to-tail separations. are combined together using a complementary filter 11 to provide formation navigation information to the control algorithms (fig. 5). By subtracting one measurement Real-time relative positions between the two aircraft from the other (x 2 x 1), the error (e n 2 n 1) is are derived at 20 Hz by time-correlating and then obtained. A first-order low-pass filter, differencing their twice/sec GPS position measurements. 9 With common satellites and close 0.25g proximity between the aircraft, common mode errors G(s) (4) s + 0.25_ such as ephemeris data errors, satellite clock errors, and atmospheric distortion effects are very nearly eliminated, l° The remaining dominant error is caused tuned to remove n 1 from e, provides n 2. Subtracting this by the 2 samples/sec data being processed at 20 Hz. The result from the original inertial measurement (z + n 2 ) resulting GPS relative-position measurement therefore produces an accurate relative-position measurement.
can be represented as The complementary-filtered position estimate is differentiated to provide a velocity estimate with a xl(t ) = z(t) +nl(t ) (2) reliable low-frequency component. This estimate is used to isolate and correct the inertial relative velocity measurement through a second complementary filter where Xl(t ) is the GPS-measured position, z(t) is the true position, and n l(t ) is the measurement error. employing an identical first-order low-pass filter.
- _ North Trailing airplane :_ TiTiiiiiiiiiiiiiiiii; Leading _N_ iiii_!!!!i airplane Lateral axis Vertical axis Longitudinal axis Selected heading 020160 Figure 4. Formation reference frame.
American Instituteof Aeronanticsand Astronautics VINS (20 Hz) PINS + PFILT (20 Hz) m Ini PGps(2"z) ', 020161 Figure 5. INS/GPS complementary filter.
Control feedback errors are calculated as nonzero F/A-18 model that contains the nonlinear aircraft relative velocities and deviations in relative position inner-loop control system and a linear model of the from the commanded formation spacing. These errors dynamics of the trailing airplane. The design approach are provided to the formation autopilot control system has been to optimize the controller response to a l-it and also are used to drive the pilot's display.
step command within specified constraints for rise time, Relative-position errors are displayed to the pilot as percent of overshoot, settling time, and degree of deviations of the instrument landing system needles.
steady-state error. Each axis has been separately tuned, This implementation helps minimize engagement transients by assisting the pilot in "getting on condition" with constraints placed on the allowable stability before activating the formation control system. During margins. The leading airplane has been excluded from the experiment, the pilot could use the instrument the design process by virtue of the leader-follower landing system needles to monitor the performance of approach and by considering its states as time invariant.
the formation autopilot.
Because of differences between the quasilinear design Position Control Algorithms models and the nonlinear, hardware-in-the-loop F/A-18 simulation, small adjustments have been made to the A classical control design employing proportional plus derivative relative-position error feedback (fig. 3), gains provided by CONDUIT to achieve the desired is used with additional state feedback for improved combinations of robustness and performance described damping characteristics. An integral term is added to in table 2. Stability margins have been analyzed by ensure zero steady-state position error. Acceptable breaking the loop at the equivalent stick command, and damping is achieved in the vertical axis through normal in all cases met or exceeded their design requirements.
acceleration feedback, and in the lateral axis through All four gain sets have been implemented in the bank angle feedback. The controller uses the software and flight-tested.
complementary-filtered position and velocity estimates as inputs, in addition to normal acceleration and bank Table 2. Controller gain sets.
angle from the local aircraft INS computer.
Gain set Description Four gain sets have been designed for each axis of the controller. Three of the gain sets (A, B, and C) A Highly robust and low performance correspond to the three levels of stability robustness B Medium robustness and performance requirements shown in table 1. The fourth gain set (D) is designed for the medium stability robustness category, C Low robustness and high performance but includes a nonzero position error integral gain. The D Medium robustness and performance gain sets have been tuned using a feasible sequential with position error integral quadratic programming tool, CONDUIT, ]2 and an AmericanInstituteof Aeronantics andAstronautics Air-to-Air Navigation Accuracy Flight Test Results The GPS-based air-to-air navigation system relied on The leading airplane used production F/A-18 common satellites and close proximity between the autopilot modes to maintain constant heading, altitude, aircraft. Close proximity was assured through the and airspeed. Desired lateral, vertical, and nose-to-tail selection of test conditions. During the final six flights, separation distances between the aircraft were selected the two aircraft observed different satellite sets only by the flight test engineer through the pushbutton twice while in formation flight. Large GPS position display. When the research system was engaged, the errors occurred in these situations. A change in the trailing airplane automatically made all of the formation heading resulted in the reacquisition of adjustments necessary to correct for lateral and vertical common satellites. Although close formation flight relative-position errors from the specified location aided in ensuring common satellites between the within the formation. Nose-to-tail separation was aircraft, it also tended to give rise to telemetry system regulated by the pilot through throttle adjustments.
communication dropouts caused by multipath interference.
All testing was performed in formations that allowed the pilot of the trailing airplane good visibility of the The accuracy of the air-to-air navigation system was leading airplane. A nose-to-tail separation limit between evaluated by comparing the real-time recorded the two aircraft of 56 ft was imposed to mitigate the measurements with postflight-corrected GPS data.
possibility that errant pitch or roll inputs from the Postflight-corrected GPS data measurements have been research system would result in a collision between the shown through experimentation at NASA Dryden to be accurate to within 1 ft. Figure 6 shows a plot of the two aircraft while in formation flight. This limit equates difference between the real-time calculated and to one aircraft body length, and ensured that adequate postflight-corrected GPS relative positions for each axis maneuvering room was available for the trailing during the entire time at altitude for one of the AFF test airplane after the pilot disengaged the research control flights. Figure 6 also shows the normalized error system.
distribution for each axis. The error distributions show that the uncertainty in relative GPS positions was The standard test block developed for the experiment generally less than ±4 ft.
reflects a buildup approach. A full evaluation of the stability and performance of the most robust gain set Steady-State Tracking Performance was accomplished before proceeding to the next gain set. Six types of maneuvers were defined that had During all of the flights, the pilots reported turbulence increasing levels of aggressiveness. A steady-state levels ranging from "no turbulence" to "light chop." A tracking test was accomplished first, followed by slight degradation in tracking performance was position step command response tests independently observed with increased turbulence levels, although performed in each axis. A position step then was control accuracy was always better than the design goal commanded by the flight test engineer in the back seat of ±9 ft. Figure 7 shows a 3-min tracking performance of the trailing airplane simultaneously in both axes to plot of the "A" gain formation autopilot. The pilot of the evaluate coupling between the axes. Additional dynamic leading airplane performed all of the steady-state performance of the formation autopilot was evaluated tracking tests in a consistent manner, flying within 0.25 ° by having the leading airplane perform maneuvers in of the formation reference frame heading and with less first the lateral and then the vertical axis.
than 0.25 ° of bank angle. Additionally, rudder trim deflection was less than 1° with the pilot flying "feet on the floor."
A total of 167 test points were completed in 11 research flights over a span of 2 months. All of the primary project goals were accomplished; and several Occasionally, significant biases existed between the additional research objectives were achieved, including INS velocities of the two aircraft because they had tracking while in the wingtip vortex of the leading slightly different INS systems, and preflight operations airplane. The flight test results are presented herein in often resulted in unequal alignment qualities prior to five groups: air-to-air navigation, steady-state tracking, takeoff. Figure 8 shows a comparison of the normalized step command response, maneuvering flight, and flight heading alignment and INS velocity errors between the within the vortex.
two aircraft for the entirety of flight 715.
American Institute of Aeronantics and Astronautics 10 1.0 Longitudinal error, .5 ft .....
-5 -10 0 10 1.0 Lateral error, .5 ft
....... .... t
-5 -10 10 1.0 i .......................... i....................................
Vertical 0 .5 error, ft ................................................... J ..........
-5 i I I I ; -10 0 500 1000 1500 2000 -10 -5 0 5 10 Normalized error Time, sec distribution, ft 020162 Figure 6. Relative GPS errors from flight 713.
I I : _ D_esign goal I.
o .2 = 0 Q.
(,1 > -5 -10 -5 0 5 10 Lateral position error, ft 020 i 63 Figure 7. Flight 714 gain set "A" steady-state tracking performance.
American Institute of Aeronantics and Astronautics 1.0 t , o - East t-- i i ' x- North
, ?, t
: 1-
.8 U O .......... r ......... r i i I r i i g.4 "l ..............
i
,/
I - I i .......... L .... _----L _L ........ _ _ _
'iY-
.e-f- '-" ,, , i i z 0 -1.0 -.5 0 .5 1.0 - 10 -5 0 5 10 Heading alignment error distribution, deg INS velocity error distribution, ft/sec 020164 Figure 8. INS alignment errors from flight 715.
These heading alignment errors resulted in a biased effects of the lateral velocity error bias. When the filtered output was selected, the tracking performance of relative-velocity calculation. The complementary filter the formation autopilot improved significantly.
combined the derivative of the GPS-corrected position data with the INS velocities, effectively eliminating this Table 3 shows a summary of the steady-state tracking bias. The relative-velocity feedback signal to the performance of the four gain sets. No GPS errors are controller was selectable between filtered and unfiltered included and only results with the velocity filter values by the flight test engineer using the pushbutton engaged are considered.
display unit. Figure 9 shows a portion of the gain set "B" steady-state tracking test from flight 715 during Step Command Response which the velocity filter was activated. During the time in which the unfiltered velocity error was being sent to Step response characteristics of all four gain sets were the controller, the linear controller developed a large, crisp and predictable. Pilot comments indicate that gain lateral position offset while attempting to counteract the set "C" is a little too aggressive, although acceptable.
I I I s,.
i .......... L ........... L ..... _.___ z ........... ¢ .......................
2 10 i G> I I i- O ..... Velocity filter off - i-_--- *" 5 --_*' - Velocity filter on ...... _ ...........
O , ,, = i I i .......... L ........... L ..... ...... ........ i........... !...........
o ,, ,, ¢0 ,, ,, .J i i -5 I I I ] I i , Filtered output i "---*'r_- --'¢_-'_ - - L - . ._ - - _L - --L -_ -"=.
_- 0 ............ L ........... L .....
2"2": II_ ______.... ..... -: ........... :.......
..... Nonfiltered output ......
._ -4 I ..J ",-,"¢' 1 1 I I -6 0 10 20 30 40 50 60 Time, sec 020165 Figure 9. Velocity filter effects on gain set "B" tracking from flight 715.
American Institute of Aeronantics and Astronautics between the axes. Figure 11 shows the response of gain Table 3. Controller tracking performance summary.
set "C" for simultaneous vertical and lateral 20-ft position command changes. During the combined steps Lateral axis Vertical axis for each of the gain sets, no significant coupling was Gain Standard Absolute Standard Absolute observed between the axes, and pitch response generally set deviation, ft mean, ft deviation, ft mean, ft was faster than roll response.
A 0.9 0.8 0.8 0.3 Maneuvering Flight B 0.5 0.0 0.3 0.2 Dynamic response of the lateral position controller C 0.3 0.3 0.5 0.1 was tested by having the leading airplane fly heading D 0.8 0.2 0.6 0.1 sweeps of increasing magnitude. Although some phase lag, approximately 45 °, was evident in the lateral Note: No GPS errors included; results with velocity filter on.
position response of the trailing airplane (fig. 12), no instabilities were observed. The tracking performance of the system deteriorated to worse than the design goal of The pilot's description of the combined step response of ±9 ft for all four gain sets, partly because of the effect of gain set "B" is "[I] couldn't have done it any better varying aircraft heading while measuring errors relative myself." Figure 10 shows a comparison between the to a fixed formation heading. The stable dynamic gain set "B" design model response predictions and six response of the system is acceptable, however, because repeated flight test results for independent 30-ft position it was designed strictly for nonmaneuvering flight.
command changes for both axes. A high level of correlation exists between the model predictions and Dynamic response of the vertical position controller results from flight for all four gain sets. Variations was observed during a series of slowly increasing between the flight responses, especially in the lateral altitude sweeps performed by the leading airplane axis, primarily are a result of a nonsteady-state initial (fig. 13). Significantly less phase lag was present in the condition.
vertical axis than in the lateral axis. The vertical axis commands pitch stick, which in the F/A-18 aircraft Lateral and vertical position steps simultaneously essentially is a normal acceleration command, were commanded to identify any cross-coupling analogous to the second derivative of the vertical 4O I I I I I I I I I ¢_ 30 d O '_ 20 U) O O.
lO ....... !...... ' ..... ! ....... ?...... i...... !...... !....... !....... !.....
o ' - -- -," ....... !....... i....... ,_ ...... i ...... _ ..... O' Linear model .J ', ' ', ', ', ', ', - Flight data -10 i I i i i _ I ¢= 30 .............. ; ............... _...... :___ ,, _ ___ O .B *_ 20 O O.
"_ 10 rj _iiiiiiiiiiiiiiiiiiiiiiiiill II...... ' ..... ', ....... I-...... I-...... _...... 4 ....... ', ....... _......
o ...... -'-i ..... o Linear model _ :> , ; .... - Flight data -10 0 2 4 6 8 10 12 14 16 18 20 Time, sec 02016Ei Figure 10. Gain set "B" individual step responses.
American Institute of Aeronantics and Astronautics 3O The tracking performance of the vertical controller gain set "D" was evaluated in a steady dive, initiated by the pilot of the leading airplane, from an altitude of 20 ,'_ ;_O-ft step ! i 22,500 ft at a rate of descent of approximately \ ;_\,,,._up and out i i 2000 ft/min. The heading and airspeed were held constant. As figure 14 shows, the controller was able to o 10 Response _ _"_ : : : very accurately maintain the desired position within the .9 formation during the descent.
0 .....
Flight Within the Vortex -10 Response Although beyond the scope of the original objectives, --20 .............................
the Phase 0 formation autopilot was tested in the presence of the leading airplane wingtip vortex to learn whether the current controller structure is adequate to i i i '_ -30 -20 -10 0 10 20 30 achieve the Phase 1 objective of autonomous formation Lateralpositionerror, ft drag reduction. A model of the vortex was incorporated 020167 into the NASA Dryden F/A-18 hardware-in-the-loop Figure 11. Gain set "C" combined step error response simulation for this purpose. The most significant vortex from flight 717.
disturbance when positioned for maximum drag reduction is a strong rolling moment effect. 13 Gain set position error. The lateral stick command controls roll "D" was chosen for evaluation because of the ability of rate, which is analogous to the third derivative of the its position error integral term to generate a lateral lateral position error. The additional integral from command bias to offset the vortex rolling moment. The command to response in the lateral axis results in simulation showed that the formation autopilot can greater phase lag. The vertical tracking error for this test stabilize within the outer regions of the vortex where the also exceeded the ±9-ft design requirement; however, moment effects are not the strongest, but the system vertical tracking is stable and adequately tracks the becomes unstable when flown near the core of the leading airplane, considering it was designed for only vortex.
straight-and-level flight.
2 L k Leading airplane _ i [ ........... _..........
Heading deviation, deg -2 lo ........... !.......... -! ........... ! ........ -!.......... ,< .......
Bank angle, deg
o i .... ' : : ...... : _ :/r_ 1
-l°r ........... _ ......... r ; --- ; ;- --_ i
i i i i / ........................ L ........... l ........... l ........... ± ..........
Lateral position 0 __._...._ . ........ ¢-!-%- .... _,_ .....
error, ft -10 0 10 20 30 40 50 60 Time, sec 020168 Figure 12. Gain set "C" lateral dynamic response from flight 715.
American Instituteof Aeronauticsand Astronautics 29.2 I I I I I I I i I A Leading airplane 29.1 O Trailing Altitude, kft 29.0 airplane 28.9 28.8
199. +,'+ .: :" ! ! ! _. ! |
10 ...... +...... -,....... ; ....... ;- .... . ...... " .... : ....... :-- "," - :- .....
Pitch angle, deg 0_5 +...... ' • ---+' +++,+'-' .... : '+ ....
291 t + + ! + t - + !
Vertical position :::::::::::::::::::::::::::::::::::::::: error, ft _,ot::_i ....... '-+/-+- ..... , j-i ...... .....
-2ol i i i ; i i i i ; 0 10 20 30 40 50 60 70 80 90 100 Time, sec 020169 Figure 13. Gain set "C" vertical dynamic response from flight 717.
A Leading airplane O Trailing Altitude, airplane kft l + i + i + I + s + i + , + , Vertical velocity, ft/sec ....... i. ....... i. ....... L ....... L ....... J. ....... .L ...........
i i i i i ! i
Vertical position error, ff
i , i i i i i i
-10 0 10 20 30 40 50 60 70 80 90 Time, sec 020170 Figure 14. Gain set "D" vertical dive response from flight 717.
American Institute of Aeronantics and Astronautics The simulation vortex model had not been confirmed control requirements associated with the vortex rolling moment then exceeded the bandwidth of the formation to accurately model the characteristics of real F/A-18 autopilot and caused the trailing airplane to exit the wingtip vortices, so the response of the Phase 0 vortex. Because of large fluctuations in formation autopilot system to vortex effects was also pilot-commanded throttle position to maintain tested in flight. The results, shown in figure 15, were nose-to-tail separation, no appreciable fuel flow similar to the simulation study. With the wingtips of the reduction was measured.
two aircraft spaced approximately 10-ft apart and the leading airplane 15 ft above, the trailing airplane was The response similarities between simulation and commanded closer to the vortex core in 5-ft vertical flight improved confidence in the vortex model. Further increments. Upon reaching the same altitude as the simulation studies have shown that the removal of leading airplane, the trailing airplane appeared to stick-path nonlinearities and the addition of roll rate stabilize, with approximately 0.1 in. of roll trim, at feedback in the lateral axis significantly improve the 150sec. At this point, the trailing airplane was vortex disturbance rejection capability of the formation commanded 5-ft laterally closer to the vortex core. The autopilot.14, 15 -3O X Command I ', l l O Position : Lateral -40 relative ........................... 2;.2:,2.
position, ft -5O -60 Vertical 10 relative position, ft 0 -10 .6 Roll command, in.
.,/............iilXilXiilli i
o : --.2 0 50 100 150 200 250 Time, sec 020171 Figure 15. Gain set "D" vortex immersion flight test results from flight 718.
American Institute of Aeronantics and Astronautics 5Sitz, Joel R., F-18 Systems Research Aircraft Concluding Remarks Facility, NASA TM-4433, 1992.
The Autonomous Formation Flight project successfully designed, implemented, and flight-tested 6Carter, John F. and R C. Stoliker, Flying Quality an F/A-18 formation autopilot system that combines Analysis of a JAS 39 Gripen Ministick Controller in an inertial and global positioning system measurements to F/A-18 Aircraft, NASA TM-2000-209024, 2000.
maintain precision station-keeping with another aircraft in formation. Four formation autopilot gain sets were tested, and each exceeded project design requirements 7Carter, John F., Production Support Flight Control for steady-state tracking accuracy. Steady-state tracking Computers': Research Capability for F/A-18 Aircraft at performance exceeded the goal of ±10 ft and typically Dryden Flight Research Center, NASA TM-97-206233, showed values less than ±5 ft. The complementary filter 1997.
improved tracking results by eliminating inertial system initialization and drift errors. Step responses were 8Carter, John and Mark Stephenson, Initial Flight Test well-damped, and simultaneous maneuvering showed no cross-coupling issues. Pilot comments indicate that of the Production Support Flight Control Computers at the system demonstrated an acceptable level of NASA Dryden Flight Research Center, NASA performance for a formation flight autopilot.
TM-1999-206581, 1999.
The formation autopilot also showed acceptable 9Ryan, Jack, et al., Data Synchronization behavior beyond the scope of its design concept during Discrepancies in a Formation Flight Control System, maneuvering flight. More phase lag was observed in the lateral axis than in the vertical, although both axes NASA TM-2001-210720, 2001.
remained stable. Tracking response during descending flight was excellent, exceeding the steady-state design l°Parkinson, Bradford W., "GPS Error Analysis," requirement of±10 ft.
Global Positioning System: Theory and Applications, vol. 1, American Institute of Aeronautics and The strong level of agreement between simulation predictions and aircraft response in flight has provided Astronautics, Inc., Washington, D.C., 1996, confidence in the design tools used in development. The pp. 469 483.
performance of the formation autopilot demonstrated the feasibility of this type of tracking system for more 11Brown, Robert Grover and Patrick Y. C. Hwang, advanced applications.
Introduction to Random Signals' and Applied Kalman References Filtering, John Wiley & Sons, New York, 1997.
1Beukenberg, Markus and Dietrich Hummel, 12Tischler, Mark B., et al., "CONDUI_A New "Aerodynamics, Performance and Control of Airplanes Multidisciplinary Integration Environment for Flight in Formation Flight," ICAS Proceedings' 1990: 17th Control Development," AIAA-97-3773, 1997.
Congress of the International Council of the Aeronautical Sciences, Sept. 1990, pp. 1777 1794.
13Vachon, M. Jake, Ronald J. Ray, Kevin R. Walsh, 2Hummel, Dietrich, "Formation Flight as an Kimberly A. Elmix, "Measured Performance Benefits Energy-Saving Mechanism," Israel Journal of Zoology, During the Autonomous Formation Flight Program," vol. 41, 1995, pp. 261 278.
AIAA-2002-4491, Aug. 2002.
3Military Specification, Flight Control 14Lavretsky, Eugene, "F/A- 18 Autonomous Systems General Specifications for Design, Installation and Test ofPiloted Aircrafi, MIL-F-9490D, Formation Flight Control System Design," 1975.
AIAA-2002-4757, Aug. 2002.
4Gera, Joseph and John T. Bosworth, Dynamic 15Misovec, Kathleen, ''Applied Adaptive Techniques Stability and Handling Qualities Tests on a for F/A-18 Formation Flight," AIAA-2002-4550, Highly Augmented, Statically Unstable Airplane, NASA TM-88297, 1987. Aug. 2002.
AmericanInstituteof Aeronautics andAstronautics
REPORT DOCUMENTATION PAGE Form Approved
OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Re _orts, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.
1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORTTYPE AND DATES COVERED August 2002 Technical Memorandum 4. TITLE AND SUBTITLE 5. FUNDING NUMBERS An Overview of Flight Test Results for a Formation Flight Autopilot WU 706 35 00 E8 28 00 AFF 6. AUTHOR(S) Curtis E. Hanson, Jack Ryan, Michael J. Allen, and Steven R. Jacobson 8. PERFORMING ORGANIZATION 7. PERFORMING ORGANIZATION NAME(S)AND ADDRESS(ES) REPORT NUMBER NASA Dryden Flight Research Center RO. Box 273 H-2499 Edwards, California 93523-0273 10. SPONSORING/MONITORING 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) AGENCY REPORTNUMBER National Aeronautics and Space Administration NASA/TM-2002-210729 Washington, DC 20546-0001 11.SUPPLEMENTARYNOTES Also presented _ the AIAA Guidance, Navig_ion and Control, AIAA Atmospheric Flight Mechanics, AIAA Modeling and Simulation Technologies, and AIAA/AAS Astrodynamics Conferences, Monterey, California, AIAA-2002-4755.
12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE Unclassified--Unlimited Subject Category --08 This report is available at http://www.dfrc.nasa.gov/DTRS/ 13. ABSTRACT (Maximum 200 words) The first flight test phase of the NASA Dryden Flight Research Center Autonomous Formation Flight project has successfully demonstrated precision autonomous station-keeping of an F/A-18 research airplane with a second F/A-18 airplane. Blended inertial navigation system (INS) and global positioning system (GPS) measurements have been communicated across an air-to-air telemetry link and used to compute relative-position estimates. A precision research formation autopilot onboard the trailing airplane controls lateral and vertical spacing while the leading airplane operates under production autopilot control. Four research autopilot gain sets have been designed and flight-tested, and each exceeds the project design requirement of steady-state tracking accuracy within 1 standard deviation of 10 ft. Performance also has been demonstrated using single- and multiple-axis inputs such as step commands and frequency sweeps. This report briefly describes the experimental formation flight systems employed and discusses the navigation, guidance, and control algorithms that have been flight-tested. An overview of the flight test results of the formation autopilot during steady-state tracking and maneuvering flight is presented.
14. SUBJECTTERMS 15. NUMBER OF PAGES Formation autopilot, Formation flight, Station keeping 16. PRICE CODE A03 17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT OF REPORT OF THIS PAGE OF ABSTRACT Unclassified Unclassified Unclassified Unlimited NSN 7840-01-280-8800 Standard Form 298 (Rev. 2-89) Prescribed byANSI Std Z39 18 298 102