`Fouche
`
`(10) Patent NO.: US 6,751,529 ~1
`Jun. 15,2004
`(45) Date of Patent:
`
`(54) SYSTEM AND METHOD FOR
`CONTROLLING MODEL AIRCRAFT
`
`(75)
`
`Inventor: J. Michael Fouche, Huntsville, AL
`(US)
`
`(73) Assignee: Neural Robotics, Inc., Huntsville, AL
`(US)
`
`( * ) Notice:
`
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 0 days.
`
`(21) Appl. No.: 101449,372
`May 30, 2003
`
`(22) Filed:
`
`Related U.S. Application Data
`(60) Provisional application No. 601385,315, filed on Jun. 3,
`2002.
`(51) Int. CL7 ................................................ B64C 11/34
`(52) U.S. C1. .......................... 70113; 24413.21; 2441164;
`2441171; 342129; 3401967
`(58) Field of Search .............................. 70113, 4, 7, 48;
`24413.21, 164, 171, 181, 183, 158 R, 177,
`179; 342129, 30; 3401967, 975, 978
`
`(56)
`
`References Cited
`U.S. PATENT DOCUMENTS
`5,553,812 A * 911996 Gold et al. ............... 244176 R
`5,797,105 A * 811998 Nakaya et al. ................. 70117
`5,841,537 A * 1111998 Doty .......................... 3561484
`712000 Calise et al.
`6,092,919 A
`6,473,676 B2 * 1012002 Katz et al. ..................... 70114
`
`OTHER PUBLICATIONS
`
`Buschek, H.; Calise, A.J., "p Controllers: Mixed and Fixed",
`A M J. of Guidance, Control, and Dynamics, vol. 20, No.
`1, pp. 34-41, 1997.
`Robinson, Rick, "Better than Human Flight Control Sys-
`tems", Research Horizons, Winter 2000, found at http://
`gtresearchnews.gatech.edulreshorlrh-win0Olflight.html.
`
`Sugeno, Michio, "Demostration of Unmanned Helicopter
`with Fuzzy Control", 1993, found at http:llm.cs.arizo-
`na.eduljapan/wwwlatiplpubliclatip.reports.951
`atip95.13r.html.
`Sugeno, Michio, "Fuzzy Logic Controller in an Intelligent,
`Unmanned Helicopter", 1995, found at http:llm.cs.arizo-
`na.eduljapan/wwwlatiplpubliclatip.reports.941
`sugeno.94.html.
`Bluck, John, "NASA Testing New Aircraft Safety Flight
`Control Software", Release 99-21AR, Apr. 14, 1999, found
`http:llamesnews.arc.nasa.govlreleasesll999199
`21
`at
`AR.htm1.
`
`(List continued on next page.)
`
`Primary Examiner-Thomas G Black
`Assistant Examiner-Tuan C To
`(74) Attorney, Agent, or FirmAanier Ford Shaver &
`Payne P.C.
`
`(57)
`
`ABSTRACT
`In one embodiment, a method for controlling an aircraft
`comprises providing an attitude error as a first input into a
`neural controller and an attitude rate as a second input into
`the neural controller. The attitude error is calculated from a
`commanded attitude and a current measured attitude, and the
`attitude rate is derived from the current measured attitude.
`The method also comprises processing the first input and the
`second input to generate a commanded servo actuator rate as
`an output of the neural controller. The method further
`comprises generating a commanded actuator position from
`the commanded servo actuator rate and a current servo
`position, and inputting the commanded actuator position to
`a servo motor configured to drive an attitude actuator to the
`commanded actuator position. The neural controller is
`developed from a neural network, wherein the neural net-
`work is designed without using conventional control laws,
`and the neural network is trained to eliminate the attitude
`error.
`
`28 Claims, 16 Drawing Sheets
`
`226?-rT-
`
`Motor
`
`204
`
`Roll Attitude
`Neural Controller
`
`j iL!v
`
`Differentiator
`
`Roll Actuator
`
`Sensor w
`
`Parrot Ex. 1008
`
`
`
`US 6,751,529 B1
`Page 2
`
`OTHER PUBLICATIONS
`
`Sacks, Richard, LoFlyte information found at http://--
`w.accurate-automation.com/Technology/Loflyte/loflyte-
`.html.
`Rolf, Rysdyk T.; Calise, A.J., "Nonlinear Adaptive Flight
`Control Using Neural Networks", ZEEE Controls Systems
`Magazine, vol. 18, No. 6, Dec. 1998.
`
`Leitner, Jesse; Calise, Anthony J.; Prasad, J.V.R., "Analysis
`of Adaptive Neural Networks for Helicopter Flight Con-
`trols", A M Journal of Guidance, Control, and Dynamics,
`vol. 20, No. 5, p. 972-979, Sep.-Oct. 1997.
`
`* cited by examiner
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 1 of 16
`
`(Front View)
`
`FIG. I A
`
`(Side View)
`
`FIG. 1B
`
`
`
`Motor 2266
`
`204
`
`21 8
`
`222
`
`202 --)
`Roll Attitude
`Neural Controller
`
`216
`
`f - 212
`
`Differentiator r
`
`Helicopter Cyclic
`Roll Actuator
`
`Sensor
`
`220
`
`FIG. 2
`
`
`
`326 7
`
`r 3 0 4
`
`servo
`Motor
`
`f - 306
`Helicopter Cyclic
`Pitch Actuator
`
`f- 208
`
`Helicopter
`Dynamics
`
`Sensor
`
`-
`
`r
`
`318 "7
`
`316
`
`322
`
`Pitch Attitude
`
`302 -,
`f r-. Neural Controller
`-
`r 312
`
`Differentiator
`
`4
`
`320
`
`FIG. 3
`
`
`
`8 - 7
`
`42 2
`402 7
`*
`Yaw Attitude
`Neural Controller
`
`,r- 416
`
`f - 412
`
`Differentiator 4
`
`Helicopter Cyclic
`Yaw Actuator
`
`Sensor
`
`420
`
`FIG. 4
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 5 of 16
`
`US 6,751,529 BI
`
`middle layer of neurons
`
`FIG. 5
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 6 of 16
`
`+
`Servo Motor
`
`7
`Rotor
`Actuator
`
`I
`Helicopter
`Dynamics
`
`7
`Attitude
`Sensor
`
`Current Servo Position
`
`Commanded
`Servo Rate
`
`Neural
`
`Attitude Rate
`
`Attitude Error
`
`FIG. 6
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 7 of 16
`
`I Mount RC Model Helicopter to Test Stand
`
`I
`
`Lock Down RC Model Helicopter to Prohibit
`Movement Except in One Axis
`
`706
`
`Provide Open-Loop Stimulus to Servo Motor
`
`+
`
`/- 708
`Generate Servo Motor Command Profile and
`Attitude Profile Data
`
`I
`
`Select a Training Region
`
`Train Neural Network Using Open-Loop
`Stimulus Data Set
`
`f- 712
`
`I
`
`Tune the Neural Network
`
`I
`r 716
`1
`/ Calculate Attitude Error Input Bias and Add
`I
`Attitude Error lnput Bias to the Attitude Error
`lnput Neuron of the Neural Network
`I Model Helicopter Using Neural Controller
`
`Generate Neural Controller and Flight Test RC
`
`FIG. 7
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 8 of 16
`
`Time (sec)
`
`FIG. 8
`
`
`
`U.S. Patent
`US. Patent
`
`Jun. 15,2004
`Jun. 15, 2004
`
`Sheet 9 of 16
`Sheet 9 0f 16
`
`US 6,751,529 B1
`
`
`
`FIG. 9
`FIG. 9
`
`
`
`U.S. Patent
`US. Patent
`
`Jun. 15,2004
`Jun. 15, 2004
`
`Sheet 10 of 16
`Sheet 10 0f 16
`
`US 6,751,529 BI
`US 6,751,529 B1
`
`FIG.10
`
`
`
`U.S. Patent
`US. Patent
`
`Jun. 15,2004
`Jun. 15, 2004
`
`Sheet 11 of 16
`Sheet 11 0f 16
`
`US 6,751,529 BI
`US 6,751,529 B1
`
`ยง
`
`8
`
`AtfiubengleaIiSachmmfiValm 8
`
`.5
`
`8 .nn...
`
`
`
`Tminirg reg'mto be cave-fitted
`and scaledfor mm]M
`
`
`
`
`Tum(mits)
`
`FIG. 11
`FIG. 11
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 12 of 16
`
`US 6,751,529 BI
`
`140
`
`120
`
`100
`
`- ................................ i ............................ .-.-." .................................... ; ................................. ; .......................................................
`
`- .................................. ....................
`.." ............
`..................................... .....................................
`i ................................... j
`
`............................
`
`- .................................. i .....................................
`i
`
`fJ0
`
`.......
`
`.i
`
`.....
`i
`
`60 - ................................. i ...........................
`
`.............................................................
`
`-
`
`roll ''em" (shaded in gray)
`
`Tm (units)
`
`FIG. 12
`
`
`
`U.S. Patent
`US. Patent
`
`Jun. 15,2004
`Jun. 15, 2004
`
`Sheet 13 of 16
`Sheet 13 0f 16
`
`US 6,751,529 B1
`
`
`
`Time (see)
`
`FIG. 13
`FIG. 13
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 14 of 16
`
`US 6,751,529 BI
`
`Current Servo Position
`
`Attitude
`Sensor
`
`-C
`
`+C
`f
`
`Attitude Rate
`
`Attitude Error
`
`Perforrnance-
`Shaping Constants
`
`FIG. 14
`
`
`
`/--------L----- I1 ____..-. 9
`
`!
`
`~
`
`0
`
`0
`
`1
`
`
`OOCJO
`
`I
`
`I
`!
`
`I
`
`--
`
`-
`
`j
`--
`i
`SOao
`m(r0
`SfXJo-
`I
`t o a t i.__ 1000-
`1 0 ~ 0
`----- m0'0
`LOO@
`i
`WOO
`EL00
`C100-
`!
`!
`
`1 .
`
`
`
`I
`
`- 5 1
`16
`16
`06
`
`!
`
`J
`I
`
`
`
`U.S. Patent
`
`Jun. 15,2004
`
`Sheet 16 of 16
`
`. - .. - ........
`. . . .....
`[ - ~ i t e ~ t i ~ n s ' n s ' ? i ~ ~ ~ r b o u n ~ ~ ~ ~ r '
`' b ~ i ~
`i neural network i
`bias
`i
`1
`j
`output
`i
`
`--'1
`
`I
`
`I
`
`FIG. 16
`
`
`
`US 6,751,529 B1
`
`1
`SYSTEM AND METHOD FOR
`CONTROLLING MODEL AIRCRAFT
`
`2
`Even for RC fixed-wing aircraft, the user needs to know how
`to glide the aircraft to the ground.
`
`BACKGROUND
`
`RELATED APPLICATION
`The application claims the benefit of priority under 35 s
`U,S,C, 5 119(c) of U,S, ProvisionalApplication No, 601385,
`315 filed on Jun. 3, 2002, the entirety of which is incorpo-
`rated herein by reference.
`
`SUMMARY
`In one embodiment, a method for controlling an aircraft
`providing an attitude error as a first
`into a
`neural controller, the attitude error calculated from a com-
`manded attitude and a current measured attitude, providing
`an attitude rate as a second invut into a neural controller. the
`10 attitude rate derived from the current measured attitude,
`processing the first input and the second input to generate a
`1. Field
`commanded servo actuator rate as an output of the neural
`controller, generating a commanded actuator position from
`The present invention generally relates to aircraft control
`the commanded servo actuator rate and a current servo
`techniques and, in particular, to a system and method for
`controlling an aircraft via the use of a neural network IS position, and inputting the commanded actuator position to
`controller.
`a servo motor configured to drive an attitude actuator to the
`commanded actuator position, wherein, the neural controller
`2. Description of the Related Art
`Aircraft generally have three ranges or axes of motion
`is
`the
`a
`designed without using conventional control laws, the neural
`(roll, pitch, and yaw), and it is necessary to actively control
`the aircraft,s motion about each of the three axes of motion 20 network trained to eliminate the attitude error.
`1" another embodiment, an apparatus for controlling an
`via one or more aerodynamic actuators. In general, for
`aircraft comprises an attitude sensor operable to ~rovide a
`fixed-wing aircraft (e.g., airplanes), roll, pitch, and yaw are
`current attitude, a differentiator operable to receive as input
`primarily controlled via the aircraft's ailerons, horizontal
`the current attitude and derive an attitude rate, a neural
`stabilizer, and vertical stabilizer, respectively. For rotary-
`wing aircraft (e.g., helicopters), roll and pitch are generally 25 controller operable to receive a plurality of inputs com~ris-
`controlled via the aircraft's main or horizontal rotor, and
`ing an attitude error and the attitude rate, the attitude error
`attitude and the mrrent
`a
`yaw is generally controlled via the aircraft's tail or vertical
`rotor. However, it is common for a particular actuator to
`attitude, the neural controller also operable to generate a
`commanded servo rate from the plurality of inputs, the
`contribute to more than one axis of motion, and it is possible
`for other types of actuators to be employed in addition to 30 commanded servo rate applied to a current actuator position
`andlor in lieu of the aforementioned actuators.
`to generate a commanded actuator position, and a servo
`Properly controlling an aircraft,s motion can be a difficult motor operable receive the commanded actuator position,
`the servo motor further operable to drive an attitude actuator
`task, particularly
`in environmental conditions (e,g,,
`position, wherein the
`the
`turbulence) that cause the aircraft to behave in an unpre-
`dictable manner, Indeed, most pilots spend an enormous 35 controller is developed from a neural network designed
`laws.
`without
`amount of time and effort in learning how to properly control
`These and other embodiments of the present invention
`their aircraft.
`will also become readily apparent to those skilled in the art
`Control of model aircraft (i,e,, miniature, unmanned
`from the following detailed description of the embodiments
`aircraft) adds an additional layer of difficulty since there is
`no on-board pilot that can apply the appropriate inputs for 40 having reference to the attached figures, the invention not
`being limited to any particular embodiment(s) disclosed.
`properly controlling the aircraft, A "pilot on the ground"
`cannot sense nuances in the aircraft movement and, thus, can
`BRIEF DESCRIPTION OF THE DRAWINGS
`become disoriented very quickly. For example, if a helicop-
`The following drawings incorporated in and forming a
`ter is facing away from a pilot (i,e,, helicopter nose points in
`then the pilot,s left is the 45 part of the specification illustrate, and together with the
`same direction as pilot,s
`detailed description serve to explain various aspects of the
`helicopter,s left, But, if the helicopter yaws 180 degrees and
`embodiment(s) of the invention
`faces the pilot, then the pilot has to change his/her orienta-
`the invention
`and "Ot
`tion and method of thinking because "left is right" and "right
`FIGS. 1 A and 1B illustrate an exemplary RC model
`is left." A pilot on board will never face this problem.
`helicopter mounted to a 3-axis test stand suitable for control
`Rotary-wing model aircraft are inherently unstable in that
`by a neural controller, according to one embodiment.
`they lack positive dynamic stability, With fixed-wing
`FIG. 2 illustrates a block diagram of one embodiment of
`aircraft, their actuators can sometimes be positioned or
`a neural network roll attitude control, according to the
`configured such that the fixed-wing aircraft generally main-
`tains stable flight without additional input from the actuators 55 Present
`FIG. 3 illustrates a block diagram of one embodiment of
`(also called trimmed flight), However, most rotary-wing
`a neural network pitch attitude control, according to the
`aircraft fly in an unstable manner unless control inputs for
`present invention.
`the actuators are continuously provided. One drawback is
`FIG. 4 illustrates a block diagram of one embodiment of
`the resulting difficulty of controlling a remote-controlled
`60 a neural network yaw attitude control, according to the
`(RC) aircraft in flight.
`Present invention.
`For example, in order for a user to successfully fly and
`FIG. 5 illustrates an exemplary neural network for learn-
`control a RC helicopter either for fun or business, the user
`ing 3-dimensional relationships.
`has to be an expert pilot. In addition to having to know how
`FIG. 6 illustrates a block diagram of one embodiment of
`to fly, the user also needs to know how to autorotate the RC
`helicopter in the event that the RC helicopter engine quits or 65 an exemplary closed-loop process for a neural network
`stalls in mid air. The skills required to autorotate a helicopter
`helicopter attitude control, according to the present inven-
`is very different from the skills required to fly the helicopter.
`tion.
`
`
`
`US 6,751,529 B1
`
`a
`
`3
`4
`FIG. 7 illustrates a flow chart of one embodiment of a
`Thus, a module may include, by way of example,
`components, such as, software components, processes,
`method by which a neural controller is developed, according
`functions, subroutines, procedures, attributes, class
`to the present invention.
`ComPonents, task components, object-oriented software
`FIG. 8 illustrates an example of an operator-induced
`s components, segments of program code, drivers, firmware,
`decavinn sinusoidal wave stimulus.
`, a
`microcode, circuitry, data, and the like.
`FIG. 9 illustrates a RC model helicopter mounted on a test
`The program logic can be maintained or stored on a
`stand and canted a positive a degrees from a roll neutral
`computer-readable storage medium. The term "computer-
`vosition.
`- FIG, 10 illustrates an exemplary depiction of the effect of
`medium" refers
`any medium that par-
`an exponentially decaying sinusoidal waveform on a RC 10 ticipates in providing the symbolic representations of opera-
`tions to a processor for execution. Such media may take
`model helicopter mounted to a test stand.
`many forms, including, without limitation, volatile memory,
`l1
`nonvolatile memory, flash memory, electronic transmission
`an
`graphical
`training region, according to the present invention.
`media, and the like. Volatile memory includes, for example,
`FIG. 12 illustrates an exemplary graphical depiction of a IS dynamic memory and cache memory normally present in
`training region comprising two regions of overshoot,
`computers. Nonvolatile memory includes, for example, opti-
`according to the present invention.
`cal or magnetic disks.
`FIG. 13 illustrates an exemplary graphical depiction of an
`It should also be understood that the programs, modules,
`upper ~erformance-sha~ing line and a lower performance-
`processes, methods, and the like, described herein are but
`shaping line about a transient response curve, according to 20 exemplary implementations and are not related, or limited,
`the present invention.
`to any particular computer, apparatus, or computer language.
`FIG. 14 illustrates a block diagram of one embodiment of
`Rather, various types of general-purpose computing
`an exemplary closed-loop process for a neural network
`machines or devices may be used with programs constructed
`having a neural controller tuning concept, according to the
`in accordance with the teachings described herein. Similarly,
`present invention.
`2s it may prove advantageous to construct a specialized appa-
`FIG. 15 is a table illustrating an exemplary mapping
`ratus to perform some or all of the method steps described
`herein by way of dedicated computer systems with hard-
`between a plurality of input training sets for a RC model
`wired logic or programs stored in non-volatile memory, such
`helicopter roll attitude and its corresponding commanded
`as, by way of example, read-only memory (ROM).
`servo rate, according to the present invention.
`FIG. 16 is a table illustrating exemplary chronological
`The present disclosure is generally directed to a system
`results of an iterative roll attitude error input bias calculation
`and corresponding methods that facilitate the control of
`for a RC model helicopter, according to the present inven-
`aircraft in flight. In accordance with one embodiment of the
`-
`tion.
`vresent invention, a servo motor for moving one of the
`3s aircraft's actuators is given an open-loop stimulus (e.g., a
`DETAILED DESCRIPTION
`sinusoidal control signal with exponentially decreasing
`amplitude) that causes the servo motor to move the actuator
`The various embodiments of the present invention and
`such that the aircraft oscillates at least once about one of the
`their advantages are best understood by referring to FIGS, 1
`aircraft's axis of movement. During the oscillation, data
`through 16 of the drawings, The elements of the drawings
`are not necessarily to scale, emphasis instead being placed 40 indicative of the aircraft's response to the open-loop stimu-
`upon clearly illustrating the principles of the invention,
`lus is captured. This data is then utilized to train a neural
`network used for controlling the aircraft's actuator.
`Throughout the drawings, like numerals are used for like and
`corresponding parts of the various drawings.
`More specifically, the data is utilized to train the neural
`network to control the actuator such that the actuator tends
`Turning first to the nomenclature of the specification, at
`least one embodiment described in the detailed description 45 to return the aircraft to an equilibrium position when dis-
`placed from the equilibrium position. In other words, the
`that follows is presented largely in terms of processes and
`neural network is trained to "zero-out" an attitude error (i.e.,
`symbolic representations of operations performed by
`a displacement from the equilibrium position). Once the
`computers, including computer components. A computer
`neural network is trained, it is implemented as a neural
`may be any microprocessor or processor (hereinafter
`referred to as processor) controlled device capable of so controller that is used to control the actuator during actual or
`test flight conditions. Based on the aircraft's flight
`enabling or performing the processes and functionality set
`performance, the neural controller is tuned by adjusting
`forth herein. The computer may possess input devices such
`inputs to the neural controller such that the neural controller
`as, by way of example, a keyboard, a keypad, a mouse, a
`properly maintains the stability of the aircraft.
`microphone, or a touch screen, and output devices such as a
`computer screen, printer, or a speaker. Additionally, the ss
`Even though the principles of the various embodiments of
`computer includes memory such as, without limitation, a
`the invention described herein are suitable for controlling
`memory storage device or an addressable storage medium.
`aircraft in general, for ease and clarity of explanation, the
`The computer, and the computer memory, may advanta-
`invention will be further disclosed in the context of control-
`geously contain program logic or other substrate configu-
`ling remote controlled (RC) aircraft. More particularly, a
`ration representing data and instructions, which cause the 60 neural network and resulting neural controller suitable for
`computer to operate in a specific and predefined manner as,
`controlling a RC model rotary-wing aircraft such as, by way
`described herein. The program logic may advantageously be
`of example, a helicopter, will be disclosed. It is appreciated
`implemented as one or more modules. The modules may
`that the principles of the invention disclosed herein in
`advantageously be configured to reside on the computer
`conjunction with the control of RC model helicopters may
`memory and execute on the one or more processors (i.e., 65 be utilized to control RC model fixed-wing aircraft as well
`as actual rotary and fixed-wing aircraft (i.e., non-model
`computers). The modules include, but are not limited to,
`software or hardware components that perform certain tasks.
`aircraft).
`
`30
`
`
`
`US 6,751,529 B1
`
`6
`5
`troller 202 is a software implementation of a plurality of
`FIGS. 1A and 1B illustrate an exemplary RC model
`helicopter 10 mounted to a 3-axis test stand 20 suitable for
`equations that define a neural network that is taught to
`reduce the roll attitude error to zero. Stated another way, the
`control by a neural controller, according to one embodiment.
`As depicted, FIG. 1A illustrates a front view of RC model
`neural network is trained to control the roll actuator of RC
`helicopter 10 mounted to test stand 20 (i.e., a holding s model helicopter 10 such that the angular displacements of
`mechanism) and FIG. 1B illustrates a side view of RC model
`RC model helicopter 10 about the roll axis are "zeroed out."
`helicopter 10 mounted to test stand 20. RC model helicopter
`Designing and teaching a neural network suitable for use in
`10 comprises a fuselage 102, a rotor 104 coupled to fuselage
`designing roll attitude neural controller 202 is further dis-
`102, a tail boom 106 coupled to fuselage 102, and a tail rotor
`cussed below.
`108 coupled to tail boom 106 substantially at a distal end 10
`Roll attitude neural controller 202 receives as input a roll
`opposite fuselage 102. As used herein, the terms
`attitude error 214 and a roll attitude rate 216. Roll attitude
`error 214 is the difference between a commanded roll
`"connected," "coupled," or any variant thereof, means any
`attitude 218 and a measured (actual) roll attitude 220, and
`connection or coupling, either direct or indirect, between
`roll attitude rate 216 is the derivative of measured roll
`two or more elements; the coupling or connection between
`the elements can be physical, logical, or a combination 1s attitude 220. Roll attitude neural controller 202 processes
`thereof.
`the inputs and generates a servo actuator rate command 222,
`As depicted, rotor 104 generally functions to control roll
`which is an incremental delta position (negative or positive)
`that is applied to a current actuator position 224 to generate
`(i.e., motion about the z-axis)-and pitch (i.e., motion about
`the x-axis), and tail rotor 108 generally functions to control
`a commanded actuator position 226 to servo motor 204.
`yaw (i.e., motion about the y-axis). Rotor 104 is coupled to 20
`one embodiment, servo actuator rate command 222 is
`a rotor actuator (not shown) and tail rotor 108 is coupled to
`multiplied by a delta-time value (rate*delta-time), and the
`a tail rotor actuator (not shown). Each rotor is mechanically,
`resulting value is added to current actuator position 224 (old
`electrically, or hydraulically coupled or linked to its respec-
`position) to generate commanded actuator position 226 (a
`tive actuator. A change in actuator position directly causes a
`position), stated another way, a new position=old
`change in the lateral position of its coupled rotor, which, in 25 position+(rate*~e~ta-time), ~h~ new position command is
`turn, affects the roll, pitch, andlor yaw attitude (i.e., the
`then sent to servo motor 204,
`lo. The
`Servo motor 204 is coupled to helicopter cyclic roll
`of RC
`y,
`and z-axes are indicated in FIGS. 1A and 1B by "dashed"
`actuator 206 and generally functions to drive helicopter
`lines and are not part of RC model helicopter 10.
`cyclic roll actuator 206. Servo motor 204 receives as input
`Test stand 20 generally functions to hold RC model 30 commanded actuator position 226 and, based on this input,
`helicopter 10 and, more particularly, functions to enable RC
`drives or controls helicopter cyclic roll actuator 206 to
`model helicopter 10 to move about a single axis while
`accordingly change position in response to commanded
`prohibiting movement about the other two axes. Stated
`actuator position 226. In this instance, helicopter cyclic roll
`differently, RC model helicopter 10 can be mounted to test
`actuator 206 is coupled to rotor 104, and a change in
`stand 20 and configured such that RC model helicopter 10 is 35 helicopter cyclic roll actuator 206 position directly causes an
`attitude change about the longitudinal axis of rotor 104. In
`free to move about a single axis of motion (e.g., about the
`particular, servo motor 204 drives helicopter cyclic roll
`x-axis) and unable to move about the other two axes of
`motion (e.g., about the y-axis and z-axis).
`actuator 206, which in turn drives the control surface (i.e.,
`As depicted, test stand 20 comprises an arm 110 coupled 40 rotor 104).
`to a base 112. Arm 110 generally extends from base 112 and
`The aerodynamic forces that result from the control
`functions to couple to RC model helicopter 10 at a distal end
`surface movement cause RC model helicopter 10 to change
`opposite the distal end coupled to base 112. In one
`attitude. In particular, the attitude change about the longi-
`embodiment, arm 110 is coupled to RC model helicopter 10
`tudinal axis of rotor 104 affects helicopter dynamics 208
`at the helicopter's center of gravity such that no movement 45 and, in particular, the roll attitude of RC model helicopter
`is created by virtue of RC model helicopter 10 being coupled
`10. Attitude sensor 210 generally functions to measure an
`to test stand 20.
`attitude change and output a new or measured attitude. In
`this instance, attitude sensor 210 measures the roll attitude
`For example, to design a neural controller to control the
`z-axis (roll axis) of RC model helicopter 10, a user mounts
`change and outputs measured roll attitude 220, which is used
`RC model helicopter 10 to test stand 20 and enables move-
`to generate roll attitude error 214 and roll attitude rate 216.
`ment only in the z-axis while locking-down or preventing
`Differentiator 212 generally functions to generate an
`movement about the x-axis (pitch axis) and the y-axis (yaw
`attitude rate from an input attitude measurement. For
`axis). The axis of interest (i.e., roll axis) is the only "free"
`example and in this instance, differentiator 212 receives as
`axis, and the other two axes (i.e., the pitch and Yaw axes) are
`input measured roll attitude 220 from attitude sensor 210
`locked-down.
`55 and generates roll attitude rate 216 by calculating the
`FIG. 2 illustrates a block diagram of one embodiment of
`derivative of measured roll attitude 220. Roll attitude rate
`216 is then provided as one input to roll attitude neural
`a neural network roll attitude control, according to the
`controller 202.
`present invention. The neural network roll attitude control
`generally functions to control the roll axis of RC model
`In one embodiment, servo motor 204, helicopter cyclic
`helicopter 10. As depicted, the neural network roll attitude 60 roll actuator 206, and attitude sensor 210 are housed within
`or as part of RC model helicopter 10 and roll attitude neural
`control block diagram comprises a roll attitude neural con-
`troller 202, a servo motor 204, a helicopter cyclic roll
`controller 202 is located external to RC model helicopter 10.
`actuator 206, a helicopter dynamics 208, an attitude sensor
`For example, roll attitude neural controller 202 may be
`210, and a differentiator 212.
`housed and execute within a computer. In this embodiment,
`Roll attitude neural controller 202 generally functions to 65 roll attitude neural controller 202 can communicate with the
`control or maintain RC model helicopter 10 in a commanded
`components housed within RC model helicopter 10 either
`rollattitude.Inoneembodiment,rollattitudeneuralcon- t h r o u g h w i r e l e s s c o m m u n i c a t i o n ( e . g . , r a d i o
`
`
`
`US 6,751,529 B1
`
`8
`7
`FIG. 4 illustrates a block diagram of one embodiment of
`communication, etc.) or via a physical connection. In
`another embodiment, roll attitude neural controller 202 is
`a neural network yaw attitude control, according to the
`present invention. As depicted, the neural network yaw
`housed within or as part of RC model helicopter 10.
`attitude control comprises a yaw attitude neural controller
`It is appreciated that the aforementioned components
`406,
`yaw
`404, a
`depicted in the neural network roll attitude control are only 5 402, a
`helicopter dynamics 208, attitude sensor 210, and a differ-
`illustrative and the neural network roll attitude control may
`entiator 412.
`comprise other components and modules not depicted. The
`The neural network roll attitude control generally func-
`depicted components and modules may communicate with
`tions to control the yaw axis of RC model helicopter 10 in
`each other and other components comprising the
`a manner similar to that of the neural network roll attitude
`network roll attitude control through mechanisms such as, 10 control and the neural network pitch attitude control dis-
`by way of
`direct
`access, interprocess
`closed above. In particular, yaw attitude neural controller
`402 generally functions to control or maintain RC model
`communication, procedure and function calls, application
`helicopter 10 in a commanded yaw attitude,
`program interfaces, other various program interfaces, and
`one
`embodiment, yaw attitude neural controller 402 is asoftware
`and
`protocols.
`Furthermore, the functionality provided for in the compo- 15 implementation of a plurality of equations that define a
`nents and
`may be
`fewer
`neural network that is taught to reduce the yaw attitude error
`or modules or further separated into additional components
`to zero,
`or modules.
`Yaw attitude neural controller 402 receives as input a yaw
`FIG. 3 illustrates a block diagram of one embodiment of
`attitude error 414 and a yaw attitude rate 416. Yaw attitude
`a neural network pitch attitude control, according to the 20 error 414 is the difference between a commanded yaw
`attitude 418 and a measured (actual) yaw attitude 420, and
`Present invention. As depicted, the neural network pitch
`yaw attitude rate 416 is the derivative of measured yaw
`attitude control comprises a pitch attitude neural controller
`302, a servo motor 304, a helicopter cyclic pitch actuator
`attitude 420. Yaw attitude neural controller 402 processes
`306, helicopter dynamics 208, attitude sensor 210, and a 25 the inputs and generates a servo actuator rate command 422,
`differentiator 312.
`which is an incremental delta position (negative or positive)
`that is applied to a current actuator position 424 to generate
`The neural network pitch attitude control generally func-
`tions to control the pitch axis of RC model helicopter 10 in
`a commanded actuator position 426 to servo motor 204.
`a manner similar to that of the neural network roll attitude
`Servo motor 404 is coupled to helicopter cyclic yaw
`control disclosed above. In particular, pitch attitude neural 30 actuator 406 and generally functions to drive helicopter
`controller 302 generally functions to control or maintain RC
`cyclic yaw actuator 406, Se