GB2399894A

GB2399894A - An adaptive control system that avoids signals to which the system is not to adapt

Info

Publication number: GB2399894A
Application number: GB0409249A
Authority: GB
Inventors: Eric Norman Johnson; Anthony J Calise
Original assignee: Georgia Tech Research Institute; Georgia Tech Research Corp
Current assignee: Georgia Tech Research Institute; Georgia Tech Research Corp
Priority date: 2000-04-25
Filing date: 2001-04-20
Publication date: 2004-09-29
Anticipated expiration: 2021-04-20
Also published as: GB2399894B; GB0409249D0

Abstract

An adaptive control system 10 for controlling a piece of plant 12 (for example a vehicle, aircraft or space craft) comprises a 'hedge' unit 20. The hedge unit 20 receives an input signal, a command control signal and a signal indicative of the plant state. The unit generates a hedge signal based on the received signals and a hedge model to modify the command control signal. The hedge model enables a characteristic of the plant (for example a manual intervention in an automatic system) that may have a deleterious effect on the control, to be avoided and thus not influence the subsequent adaptive control. The command control signal is modified by the hedge signal via a reference model 22. The modified signal is used to generate an error signal with the plant state signal. The error signal being passed to an adaptive law unit 26 (for example a neural network) that generates an adapted control signal for the system controller 14. The system operator may be a human or an automatic operator.

Description

if-- 2399894 ADAPT CONTROL SYSTEM HAVING HEl)GE Unlike ANI) RELATED

APPARATUS AND METHODS Statement of U.. Government Rights In the Invention This uvendon was made with V.S. Government fimd under compact no. NAG81638 awarded by National Aeronautics and Space Adiniration (NASA) Mars} all Space Plight Center and contract no. F49620-98-1-0437 awarded by the Air Force Office of Scientific Research (AFOSR). The U.S. Government has certain rights in the invention Cross-Rcference to Related Applicat on This supplication claims priority benefits of U. S. provisional application no. 60/199,615 filed April 24, 2000 naming Eric Nonnan Johnson and Anthony 1. Calise as inventors.

Fieldf the Invention

The invention is directed to an adaptive control system and related method.

More particularly, tile invention is directed to art adaptive control system tenth the capability to prevent or reduce undesired adaptadon of a control system due to selected characteristic(s) of die plant or control system.

Bacl4round of the Invention 20: Adaptive control systems have the capability to adapt control response to changing conditions within the plant being controlled or the plant's operating environment:. Adaptation to changing plant or environmental conditions provides enticed cc.ntrol response for many kinds of plants, asked may be required for certain types of plants that cannot be controlled win static control systems. I-Inever, conversional adaptive control systems have a tendency to adapt to plant or control system characteristics to which they should not. The control response of A - five control systems can be greatly drninisbed when subjected to cestaln charactcrishc(s) of the plant or system, md On be rendered unstable in attempting to adapt to these characteristics. An example of a characteristic Mat could adversely affect an adaptive conned system is a control or suthonty limit unposed on the control elements of an adaptive control system. If an operator generates a command signal that exceeds He ability of die control system or plant to respond, adaptation of the conch system can render He 'control system wstable. It is desirable to reduce or prevent the impact of such cactenstic(s) from adversely affecting the adaptive control system's perfomanc,e.

Summary of the Indirection

The methods, apparatus, and system of the invasion overcome the disadvantages noted above.

A first method of the prevention comprises generating a hedge signal to avoid adaptation to a characteristic of at least one of an adaptive control system and a plant controlled by such system. Me first method can comprise modifying a commanded state signal faith the hedge signal. The first method can also comprise generating a reference model state signal using the commanded state signal modified by the hedge signal. The first method also can comprise generating a tracking error signal based on the reference model state signal and a plant state signal, and generating an adaptive control signal to adapt control response of the adaptive control system. Through compensatiott for Me characteristic in the Backing error signal, the adaptive control signal can be generated so as not to significantly adapt to the characteristic. Hence a characterisdc of the plant or control system that would impair or be detrimental to control syste m's performance and/or stability can be hedged out of the adaptive portion of the control system to prevent adverse impact on control of Me plant. Lee hedge signal can be generated based on a difference between a first signal derived from a plant model not having the characteristic to be hedged, and a second signal derived from a plant model having the characteristic. The first signal can generated based on an input condo]. signal and a plant state signal in addition to the plant model not having the characteristic. Me second signal can be generated usmg a command control signal and a plant state signal, addition to the plant models Ada the charactenstic. The input control si:1 can be generated based on a conunanded state signal, a plant state signal, and an adaptive control signal, and Me command control signal can be generated using the input command signal modified by a control allocation and a control charactenstic imposed by a controller, The input control signal and command control signal can be used to hedge a cteristic of the control system to which adaptation is not to be performed. The method can also include generating a display based on the input control Age al. An operator can use the display to generate a command control signal.

ID this aspect of the invention, the operator's control and response can be hedged.

A s' cond method of the invention is executed by an adaptive control system.

The second method comprises generadug an input control signal based on a commanded state signal, a plant state signal, and an adaptive control signal. The second remend also comprises generadog a command control signal based on a A: commanded suite signal, a plant state sisal, an adaptive control signal, and fiber based on cone ol allocation and a control characteris:6c of a controller Id to generate be command control signal. The second method farther comprises supply the S conmapd control signal to an actuator, controlling a state of the plant based on the commapd control signal, sensing a State of the plant, and generating a plant state signal based on the sensing of the plant. The second method comprises generating a first signal based on the input control signal, the plant state signal, and a plant model without a plant characteristic for which Me adaptive control system is not to adapt. The second method also comprises generating a second signal based on the command control signal, the plant state signal, and a plant model with the plant characteristic for which the adaptive control system is to adapt. The second method fimher comprises generadng a hedge signal by differencug the first and second signals, and generating a reference model state signal by rnodifjg the commanded state signal with the hedge signal to mc] ude the effect of the control allocation and control characteristic on plant state from die reference model state signal. The second method BJrtller comprises compg tle plant state signal and We reference model state signal, generating a tracing error signal based on the comparing step, and generating Me adaptive control signal based on the tracking error signal. Ille second method can comprise generating a reference model signal based on the commanded state signal, the hedge signal, and a reference model signal derived from a reference model representing the target response of die plant, Me reference model signal tO generate the input control signal, The second method car, also comprise generating a reference model signal based on Me commanded state signal, the hedge signal and a reference model signal derived from a reference model representing Me target response of Me plant, the reference model signal, to generate the command control signal. lithe second method can also compuse generetg. linear control signal based on Me tracking error signal, generating a reference model signal based on the commanded sate signal, the hedge signal, and a reference model, and generating a pseudocontrol signal bred on the linear control signed, Me reference model signal, and the adaptive control signal, and Me pseudo- control signal. The adaptive control signal can be generated mth Me plant state signal.

The adaptive control signal can be generated with a neural network having connection weights adjusted based on Me backing error signal and the pseudo control signal. Me neural network maps the plant state signal to the adaptive control signal. The plant state sift can also be used to generate the adaptive control signal. The second method can compose generadug the commanded state signal based on a control action born an operator. The operator can be human, and Me method can comprise generating a display based on the plant state signal. The display can be used by the operator to generate the cornnanded state signal. The second method can comprise generating the s commanded state signal based on a signal generated by an operator that is a computer.

The second method can also compuse generating a display for an operator based on the input control signal so that the operator can generate the command control signal based on the display.

An apparatus of the invention can be used in an adaptive control system for lo controlling a plant. The apparatus is a hedge unit coupled to receive at least one control signal md a plastic state signal. The hedge mat generates a hedge signal based on the control signal, the plant state signal, and a hedge model including a first model not having a characteristic to which the adaptive control system is not to adapt, and a second model having the characteristic to which the adaptive control system is not to adapt. The hedge signal can be used in the adaptive control system to remove the characteristic Tom a signal supplied lo an adaptation law unit of the adaptive control system so that the adaptive control system does not adapt to the characteristic in controlling the plant. The characters to be hedged by the hedge unit can be a time delay between generation of the commanded state signal by the controller at a time, and receipt by the controller of the so plant state signal resulting from the commanded state signal generated at the time. Also, the chaactensti^ can be a time delay between generation of a state by the plant and sensing of the state of the plant by the sensor to generate the plant state signal.

Alternatively, tire characteristic can pertain to a control limit of the actuator used to control the plant. The control limit can be due to actuator end points, actuator dynamics, a as rate limit of the actuator, or quantization effects associated with the actuator, for example.

An adaptive control system of We invention is coupled to receive a command state signal indicative of a target state of a plant controlled by die adaptive control SysTeru. The adaptive condor system comprises a controller coupled to receive the cornnanded state signal, a plant state signal, and an adaptive control signal. The controller generates an so input command signal based on me commanded state signal, the plant state signal, the adaptive control signal, and a control model. Ibe controller generates a command control signal based on based on the commanded state signal, die plant state signal, Me adaptive control signal, the control model, control allocation of the controller, and at least one control characteristic of Me controller. The controller is coupled to supply the corrunand control signal to the plant to control the plaut's state.

The actuator is coupled to receive the command control signal, and affects physical control of the plant's state using the command control signal. The adaptive control system can comprise a sensor coupled to sense the plant state, that generates a plant state signal based-on the sensed plant state. The adaptive control system also comprises a hedge unit coupled to receive the input control signal, He command control signal, and the plant state signal. The hedge unit generates a hedge signal to modify the cotrunand state signal based on the input control signal, the conunand control signal, the plant state signal, and a hedge model indicative of a characteristic of at least one of the plant and the adaptive control system, to remove the effect of the characteristic on a tracking error sisal. We adaptive control system also comprises a reference model urut coupled to receive the command state signal and the hedge signal. The reference model unit generates a reference model state signal based on the commanded state signal and a hedge signal. The adaptive control mat also comprises a comparator unit coupled to receive the reference model state signal and the plant state signal. Ihe comparator unit generates a tracking error signal based on a difference between the plant state signal and the reference model state signal. The adaptive control system also includes an adaptation law unit coupled to receive Me tracking error signal. The adaptive control system generates Me adapveconolsig;nal based on the- tracking error signal. The adaptation law unit is coupled to supply the adaptive control signal to the controller. The controller can generate Me input control signal and the command control signal further based on the reference model state signal. Lee characteristic to be hedged fly the adaptive control system can be time delay between generation of the comrnandecl state signal by the controller at a particular time, and receipt by Me controller of the plant state signal resulting Tom the commanded state signal generated at Me particular time. Alternatively, the characteristic can be a time delay between generation of a state by the plant us response to Me command control signal, and sensing of the state of Me plant resulting from the command control signal. Purger, Me characteristic can be a control limit of the actuator used to control the plant. The control limit can be due to Dictator end pouts, actuator dynamics, a rate limit of the actuator, or quarutzation effects of me actuator, for example. The commanded state signal can be generated by an operator, and the adaptive control system can cornpuse art operator interface unit coupled to receive the plant state signal. The operator interface mat relaying Me plant state to We operator, The commsDd unit can be used by Me operator to generate the command state signal based OR the operator's control action. The operator interface can be a display generated based on the plant state S signal. The operator can be a human being that generates Me control action to the command unit to generate the commanded state signal. The eonnanded state signal is -- generated by a machine operator based on Me plant state signal. The adaptation law Unit can comprise a neural network having connection weights determined by Me tracking erro!. signal. The neural network can map the plant state signal to the adaptive control signal based on the connection weights to generate the adaptive control signal.

The controller can generate a pseudoontrol signal based on the commanded state signal and OR plant state sisal. Me controller c" be coupled to supply Me pseudo control signal to the neural network to adjust the connection weights of the neural network. Lee controller can comprise a dynamic inversion unit to generate the command control signal.

These together with other objects arid advantages, which will become subsequently apparent, reside in the details of construction and operation of the invented methods, apparatus, and ample as more fully hereinafter described and claimed, reference being made to the accompanyg drawings, forming a part hereof, wherein like numerals refer to like pares throughout the several views.

Brief Description of" the Drawings

Fig. 1 is a general block diagram of the adaptive control system of Me invention that uses a hedge signal to hedge against a charactenshc to which the adaptive control system is n>t to adapt; Fig. 2 is a block diagram of a controller of the adaptive control system that generates a command control signal TO control an Donator affecting a state of a plant; Fig. 3 is a block diagram of a hedge unit of file adaptive control system; Fig. 4 is a block diagram of a reference model urut of Me adaptive control system; Fig 5 is a view of a processor-based system for implementing the adaptive control system of Figs. 14; Fig. 6 is a flowchart of a method of Me Aversion used to generate a hedge signal to prevent or reduce adaptation of the adaptive control system of Figs. 14 to Me characteristic Mat is to be hedged; Fig. 7 is a a block diagram of an adaptive control system that includes a nel network for generating an adaptive control signal based on a tracking eTor signal arid a pseudoon1 signal; Fig. '} is a relatively detailed block diagram of a hedge unit used with the adaptive cotton system of Fig. 7; _ _ Fig. is a relatively_detailed block diagram of a reference model mat used lo implement a second-order reference model to generate a target reference model state response to a cozrunanded state signal; Fig. 10 is a relatively detailed view of a neural network Mat can be used to map a tracking error signal and a plant state signal to an adaptive control signal using connection Sleights set adaptively based on a tracking error signal, a pseudo control signal, and the plant state signal; and Fig. I 1 is a block diagram of a control and allocation and characteristic unit a special case in which art operator is included in the unit.

Detailed Description of the Invention

As used herein, the following terms have the following defiridonm: "Actuator" can be Ally any device capable of affecting the state of a plant to control a degree of freedom Hereof. Such actuator can be a motor, motoriven screw, a hydraulic cylinder, a pump or valve controlling a stream of air, a thermal heater, a compressor or suction generator, or other device. - - "Adaptive control system" means a control system having the capability to adapt to changes a controlled plant or its environmert over time.

"Charactenstic" is a property of a plant or control system that has an effect for which adaptation of the control system is not to be performed. The characteristic can be a time delay between generation of a command signal and sensing and report of the plant state r exulting Tom the command signal to Me control system. The characteristic can also be a control InTut such as actuator end points, e.., extreme positions, temperatures, pressures, etc. obtainable by the actuator, actuator discs, rate limits, quanitizatkn effects, and possibly others. The characteristic can also be a feature of a sensor, for example, the time delay from change of a plant state to sensmg of that changed plant state by He sensor. The characteristic can also be an operators control or response.

Control limit" is a limit on the capability of a control system to control a plant.

4 control limit can also be imposed by llmtations in flee actuators used to control the plant These limitations nn include actuator end points, e. g., extreme positions, temperatures, pressures, etc. obtainable by the actuator, actuator dynamics, rate limits, quanitizatio effects, and possibly others. Control lets could also be imposed intentionally for a variety of reasons. Adaptive control systems are sensitive to control limits That can cause the adaptive control systems to lose stability. The invention - - provides--the. capability to compensate for control limits to permit stable control of the plant with the adaptive control system.

"Hedge" means to reduce or prevent adaptation of an adaptive control system lo a characTenstic. - "Hedge model" is a model of one or more elements of the system or plant with and without a characteristic Mat is to be hedged A hedge model may be a model of a plant, a control system, e g., an actuator or serlsor, an operator, or any over feature of We control system or plant to which the control system is not to adapt.

"Operator" can be a human or computer, for example, 1bat serges a plant state using a plant state signal, and generates a commanded state signal to control Me plant.

"Plant" refers To a system controlled by a control system. For example, the plant can be an aircraft, spacecraft space-launch vehicle, satellite, missile, Added notion automobile, or other vehicle. The plant can also be a robot, or a pointing or orientation system such as a satellite orientation system to orient power-generation panels, a transceiver, orb dockingmechahisrn, Such plant can also be a braking system, an engine, a transmission, or an active suspension, or other vehicle subsystem.

The plant could be a manufacturing facility or a power generation facility. The plant could also be virtually any controllable system.

"Sensor" can be virtually any device(s) for sensing a degree of Deedom of a plarst's state, whether alone or in combination with one or more over sensors, to generate a measurement or estimate of plant state. The sensor can be virtually any device suitable for sensing information regarding a plant's stance. For example, the sensor could be a gyroscope for detecting orientation of a vehicle such as an aircraft, i.e., pitch or roll attitudes or side slip. The sensor can also be a temperature or pressure sensor, a position, velocity, or inertial sensor.

"(I)" means one or more of He thing meant by the word preceding "(s)". Thus, characteristicts) means one or more characteristics.

11e adaptive control system 10 of Fig. 1 operates in a control cycle that is repeatedly executed to control the plant 12 on an ongoing basis, at least 'until control is terminated. The basic Unction of the system 10 is to generate a commanded control signal dead using a conunanded state signal x,, a plant state signal up, an adaptive control signal vad' and optionally a reference mode state signal x,,#. In the case of the first execution of the control cycle, when previous values of the plant state signal up and doe adaptive control signal vad are employed, they are assigned predetermined values, respectively/. The commanded state signal or is used by Me controller 14 along with We plants state Signal Up and an adaptive control signal va, to generate at least one condor signal d. lf adaptation to one or more characteristics of Me controller 14 is to be avoided, the controller 14 is implemented to generate an input control signal d,,, that is the control signal d generated by the controller 14 before imposition of the characteristic(s) on such control signal, asked a command control signal dank that is the input control signal d,,, modified by the controller's characteristic. Ibe controller 14 is coupled to supply the command control signal dam to the acinator 16 that controls the state' of the plant 12. The plant sate is sensed by the sensor(s) 18 to generate a plsut state' signal up. The hedge unit 20 Is coupled to receive the input control signal din and the command control signal dead for the control cycle under execution fiom the controller X4. In addition, the hedge mat 20 is coupled to receive Me plant state signal xp, and optionally Me actuator state signal x. The hedge unit 20 curses a plant model to predict the plant states resulting from the input command signal din and the command control signal do and generates a hedge signal vh based on a difference between these twopredicted plant states. Therefore, the hedge urut 20 generates the hedge signal v', to isolate the contribution to plant state that results from the presence of the control characten.tic(s) the command control signal de,,,. In addition, the hedge unit 20 can impose a plant model without one or more plant cbaractenstics) for which adaptation is not to lx performed by the system 10, and a plant raodel Mat includes Me plant charactenstic(s). By using Me plant model without me plant characteristic to modify the input Control signal din to generate a first signal and by use the pIant model with thelcharactenstic(s) to modify the command control signal dam to generate a second signal, the hedge unit 20 can generate the hedge signal Vh to isolate the contribution to plant state that results from the presence of the plant characteristic, The hedge Wit 20 is coupled to supply the hedge signal ah to the reference model unit 22. The reference model mat 22 is also coupled to receive the commanded sate signal xr. The reference model urn; t 22 uses the hedge signal vh and a reference model indicating the target state <,f the plant based on the commanded state, to modify the eomms'nded state signal or to include the contnbution to plant state caused by Me characteristic(s) to be hedged. The comparator 24 is coupled to receive the reference model state signal X,7,' and the plant state signal Up, and generates a tracking error signal e. The comparator 24 can generate Me Tracking castor signal e by differencing the plant state signal xp and Me reference model state signal xr,n to generate die Tacking error signal e. The comparator 24 is coupled to supply the tracking error signal e to Me adaptation law unit 26. Because the charactensti.c(s) for which no adaptation of control response is to be made has been removed, a, least partially, from the tracking error signal e, the adaptation law unit 26 will necess;mly not adapt to such characteristic(s). Accordingly, the presence of the characteristic(s) has no impact on performance of the adaptation law unit 26. The adaptation law unit 26 generates an adaptive control signal vat based on the Racking error signal e. The adaptation law unit 26 can optionally be coupled to receive Me plant state signal up as indicated by the broker line Fig. 1, for use in generating the adaptive control signal v. In addition, the adaptation law unit 14 can be coupled to receive a pseudo-control signal vpc generated by Me controller 14 Or use by the adaptation law unit 26 in generating the adaptive control signal vie.

In Fig. 2, the controller 14 includes a control model 140 and a control allocation and characteristic urut 142. lye control model 140 and the control allocation and characteristic unit 142 can be implemented as software modules within the hedge unit 14. The control model 140 is basically a software implementation of die control law to be implemented by Me adaptive control system 12 to control the plant 12. The control model 14' is coupled to receive either the commanded state signal x, or the reference model state signal arm, and is also coupled receive Me plant state signal x', a'd the adaptive control signal Ad. The control model 140 maps these signals to the input cool signal dint Those of ordinary swill in the act will understand how to generate a control law appropriate for a plant controlled by Me adaptive control system 12. For example, the control model 140 can be a linear proportional plus derivative, or proportional plus integral control law unplemerted in a software module or fimetion.

The control model 140 is coupled to supply the input control signal d,,, to the control allocation and characteristic unit 142. The unit 142 maps the input control signal do to the contrc'1 signal dead. The urut 142 can map the input control signal din to the control signal dead so as to allocate control responsibility for Me plart's controlled degree of freedom based on a predetermined scheme. For example, the plant 12 could be an aircraft c:onfigured so Me ailerons on both wings can be controlled to achieve a commanded roll attitude. The unit 142 can serve to allocate the amount of aileron defection to Me two actuators that control the wing ailerons so as to influence We roll attitude commanded by the input control signal ah,,. The Unit 142 can include the on- line identification of a model of the plant and the optimal or nearoptinal allocation of muliplely-r.edundant control effecters based on the solution of an optimization scheme that employs either Me identified plant model or a stored model of the plant. The urut 142' can also impart control characterstic(s) to the input control signal d,, to generate the control signal do,,,. Such control characterishc(s) could be control lungs such as actuator end points that cannot be exceeded due to limitations of Me actuator or associated control linkages. Altematively, the control characteristic(s) could be coryservadve authority limits placed on the control signal d,, to insure Me actuator end points are never encountered, For exernple, the actuator may be capable of movmg a control surface by i 21 degrees of angle. If the input control signal din designates 25 degrees of angle, the unit 142 m11 clip the input control limit d., to produce a command control signal of 20 degrees of angle. Another control lit may pert to an actuator's rate limit. It is possible that the input command signal d,, may command the actuator 16 to respond more rapidly than it is able Accordlagly, the unit 142 can be programmed to generate the command control signal d,7,d to move Me actor 16 more gradually as compared to Meinput control signal din. As another example, the actuator 16 may be able to move only ir-quan-tized-steps.- The unit 142 can be used to map the control signal d,,, to the command control signal demo so that the command control signal is quantized. In addition, Me unit 142 can be used to impose authority limits on the actuator 16. Thus, although an actuator 16 can be capable of actuation to the point otendgering the plow Me mat 142 cm be used to impose a control Emit on the actuator 16. Therefore, for example, if the operator 30 is an auto pilot monitored by a human pilot to take control in emergency situations, the auto-pilof can be Ignited to control the aircraft plant 12 to limits set by Me unit 142. Control limits set by the unit 142 can also be imposed by Me sensor 18. For example, if the actuator 16 can change the state of the plant 12 faster than the sensor 18 can sense the resulting changes, the urut 142 can limit Me input control signal din to generate Me compared control signal dad to change the plant's state in a manner the sensor 18 can accurately sense. Yet another example of a control ferret is the finite tone required of a processor to process the venous input signals to the controller and to generate the controller's output signals.

iI Turning now to Fig. 3, the hedge urut 20 includes a hedge model 200 having a plant model 201 without the plant characteristic that is to be hedged and a plant model I' 202 with the plant characteristic(s) that is/are to be hedged, and a comparator 204.

These elements of the hedge unit 20 An be implemented as one or more software modules or functions within the hedge unit 20 The plant model 201 without the _characteistic(s) is coupled to receive the lupus control signal din and the plant state signal up. The plant model 202 with the charactenstic(s) generates a first signal that indicates the predicted plant response to the input control signal do given the plant state signal xp. The plant model 202 with the plant characteristic(s) that is to be hedged is coupled to receive the command control signal dó - and tle plant state signal xp. Based or these signals, the plant model 202 with the characeenstic(s) to be hedged is used to generate a. second signal. The comparator 204 is coupled to receive the first signal Font the Plant model 201 not having the plant characteristic to be hedged, and the plant model 20:, having the charactenstic(s) to be hedged. The comparator 204 differences the first and second signals to generate the hedge signal vh. The hedge signal vh in effect isolates the plant andlor system characteristic(s) that are to be hedged.

In Fig. 4, details of a possible implementation of Me reference model unit 22 are shown. The reference model unit 22 includes a reference model 220, a comparator 222, and a state computation unit 224. The reference model 220 receives the commanded state signal x, and the r-efe-re-nce - model -state- si-gnal-x,,,, from the state computation Unit 224. Th,: reference model 220 generates a signal or' based on the commanded state signal x, and the reference model state signal x,,. The reference model 220 is software module or fimction that maps the commanded state signal or to the reference model state signal x,,,. The reference model state signal xrm represents the target plant sate corespondg to the commanded state signal x,. The sigroal x,' Mom the reference model 2,!0 is supplied to the compamtor 222, The comparator 222 also receives Me hedge signal vh from the hedge unit 20. Ibe comparator 222 subtracts the hedge signal vh Dom Ale signal x,' to generate the signal xrt'. The signal x," is supplied to the state compute don unit 224. The state computation unit 224 computes the reference model state signal x,,,, Mom the signal xr". More specifically, Me state computation unit 224 computes scalar, derivative and/or integral values of the signal x," according to Me form or order of the reference model. Accordingly, Me state computation mat 224 generates the reference model state signal x,,,, as a vector with scalar, integral andlor denvati,e terms using the commanded state signal x, as modified in tbe unit 22 bar Me ! \ hedge signal Vh, and the reference model 222. The state computation urut 224 supplies He resulting reference model state signal x,,,, or predetermined terrus thereof as a feedback signal to the reference model 220. The state computation usut 224 also supplies the reference model state signal x,,,,, to the comparator 24, The reference model state signal I, includes Me reference model response to the characteristic to be hedged so-tbat-it Cain be used in the comparator 24 to extract from the plant state signal x', the contbutio.n to plant state resulting Tom the presence of the charactenstic(s).

Accordingly, the elect of We characteristic is eliminated or at least reduced Tom the tracking error signal e so that Me adaptation law unit 26 mill not adapt to the characteristic's impact on the system 10 or the plant 12. Of course, it may not be possible to remove all impact of the characteristic from the Tacking error signal e.

Flo\vever, enough of the impact of the characteristic should be removed from die tracking error signal e so Mat system control of the plant will not be compromised' Satisfaction of this objective depends upon prespecificabons for the system baked orr control performance objectives, control stability, and Me nature of the control system and plant 12.

Fits. 5 is a possible Implementation of the system 10, The actuator 16, the sensor 18, Me operator interface unit 28, the command unit 32, a processor 34, and a memos 76 are coupled to a bus 38. The processor 34 can be a microprocessor or a microcontroller, for example. For exernple, the processor 28 coul*be a microprocessor with 64-bit word size operating at a l.OGHz instruction execution cycle. The processor 28 can be a Pentium III microprocessor commercially-available Tom Intel ConoratiDn, Santa Clara, California, or an Athlon microprocessor Mom Advanced Micro Devices, Inc., Sunnyvale, Califomia lye memory 36 stores a control program 360 and data 362. The control program 360 is executed by Me processor 34 in Me petfornace of a control cycle of the adaptive control system 10. lithe control program 360 includes the son ware modules used to implement the controller 14, the hedge unit 20, the reference model unit 22, the comparator 24, and the adaptation law unit 26. The data 362 includes data used by Me processor 34 in execudug Me software m odules of the control program 360 or temporary data generated by We processor 34 asit executes the control program 360. In the operation of the system 10 of Fig. 5, Me seDsor 18 teas sensed the state of plant 12 to generate the plant state signal xp. The plant state signal up teas balm stored as data 362 stored in the m e m ory 36 via the bus 38. Also, signals required for use in the next control cycle, which could Include the adaptive control signal vad and the plant state signal up and possibly over signals as well, are stored as data 362 ilk the memory 36. Alternatively, the adaptive control signal vie can be generated in the control cycle under execution using a fixed-point solution. The command unit 32 writes the commanded state signal x, for the current control cycle as data 362 scored in the memory 36. In executing its control program 360 over Me control cycle under execution, Me processor 34 reads the corurnanded state signal xr Mom the memory 36. The processor 34 also reads Me plant state signal xp and the adaptive control signal vat and generates the input control signal din and the corrunand control signal dc,,,. Alternatively, the plant state signal up or a construction thereof from the sensor signal derived horn a time after execunon of the previous control cycle can be stored in memory for use in Me next control cycle. The processor 34 supplies the corned control signal dC,,,d to the actuator 16 Mat controls the state of the plant 12 based thereon. Me processor 34 also executes the control program 360 to generate the hedge signal Vh based on the input control signal do, Me command control signal d,,nd, the plant state signal x', and optionally also ore the actuator state signal xa. Processor 34 farther executes the control proven 360 to generate the reference model state signal X7. The Processor 34 subtracts me reference model state signal x,, from Me plant state signal xp to generate the tracking error signal e. The processor 36 uses Me tracking error signal e and optionally also the plant state signal up and a pseudo-command signal vp, to Reiterate the adaptive controlsigual The processor 34 can store any signals needed for the next control cycle in the memory 36.

Tile method of Fig. 6 follows operation of the adaptive control system 10 of Figs. 1-5 over a control cycle. In step Sl of Fig. 6, the method begins. Ln step S2 of Fig. 6, the command mat 32 generates the commanded state signal x, based on Me operator's action andlor operator-generated signal. step S3 the controller generates ó input control signal do based on the commanded state signal x,, We plant state signal xp from sensor 18 and the adaptive control signal via generated by the adaptation lay unit 26. Alternatively, the commanded state signal xr is first used by Me reference model unit 22 to generate the reference model state signal Or=. The controller 14 generate:; the input control signal din based on the reference model state signal or,,,, the plant stare signal up Mom die sensor 18, and the adaptive control signal void generated by the adap ration law unit 26. Ire step S4 the controller 14 generates the control signal d=,d based Ott the commanded state signal Or, and the plant state signal up and adaptive as control signal van using the control allocation and characteristic unit 142 of the \l controller 4. Lo optional step S5 the controller 14 generates a pseudo-conol signal vpc based on the commanded state signal x,, and the plant state signal xp and the reference model state signal X,7,,. In step S6, the controller 14 supplies Me control signal dC'ndto the actuator 16 to control the plant 12. In step S7 the sensor 18 senses the plant state, In step S9 the controller 14 generates We plant state signal xp based on Me sensed plant sta--In step S10-the processor 34 receives the plant smte signal up. In step S11 the hedge unit 20 generates the hedge signal vh based on Me input control signal d,,,, the command control signal dC7,,d optionally on the actuator state signal xa, the plant model win the characteristic whose elect is to be removed *tom the tracking error signal e supplied to the adaptation law Bit 26, and the plater model without the characteristic whose effect is to be removed Mom the tracking error signal. In step S12 Me reference model unit 22 generates Me reference model state signal x,,,, based on the commanded state signal x,, the hedge signal Vh, and the reference model. In step S13 the comparator IS 24 generates the tracking error signal e based on the plant state signal xp and Me reference model state signal arm. En step S14 the adaptive law unit generates the adaptive control signal van based on the tracking error signal e and/or the pseudo coptrol sigma up, generated by ':he controller 14. In step S15, the processor 34 stores any signals required for the next control cycle In the memory 36. Such sisals might include the plant state signal xp arid the adaptive control signal van. In step S16 the method of Fig. 6 ends. - - -I -- - - -- - - Figs. 1-13 are views of an exemplary embodiment of the adaptive control system 10. In Fig. 7 the controller 14 composes the control model 140 and Me control allocation and characteristic urut 142, The control model 140 includes a linear control module 144, a summing unit 146, and an approximate dynamic inversion module 148.

Tlie linear control module 144 generates a linear control signal via based on the tracking error signal e. More specifically,, the linear control module 144 applies a linear control law to map Me tracking error signal e to the linear control signal via. The reference model unfit 22 generates the reference model signal v,m Mat is a subset of the vector of the reference model state signal xr-. The linear control module 144 and the reference model unit 22 are designed to control the plant 12 for target system response and stability using design techniques well-known to hose of ordinary skill in the art. In the exemplary embodiment of Fig. 7, Me adaptation control Iloit 26 includes a neural nehvork 260 The neural network 260 receives as inputs Me plant state signal up and a pseudo-conol signal vp, and the tracking error signal e. The neural network 260 maps the plant slate signal x,, to the adaptive control signal van using connection weights adaptively:et each control cycle by the tracking error signal e and tile pseudo control signal vpc. By updating the connection weights of the neural network 260 untie successive control cycles, the system 10 is adaptive to changes over time in the plant 12 as Well as me system 10. The psendoontrol signal vpc use in part to adapt the _connection weights of the neural network 260 is generated in the control model 140. In the control model 140, the adaptive control signal v is subtracted from the sum of the reference model signal Vrm and Me linear control signal pad in the summing unit 146 to generate pseudo-control signal vp, The pseudocontrol signal vp, is supplied to the hedge unit 20 for optional use in generating the hedge signal Vh. The pseudo-control signal vp is also supplied to a dye inversion unit 148 of Me control model 140.

Me dynastic inversion urut 148 inverts Me pseudoontrol signal v,,c based on an inversion Unction representing the plant control respor.'se. The inversion fimction is a Emotion of the plant state signal X' and Me pseudo-conbol signal - . The inversion function maps these signals to the input command signal au,. The remainder of the adaptive control system 10 is similar in function and configuration to previously described embodiments.

Figure is an exemplary embodiment of the hedge unit 20 of Fig. 7. In this embodiment, the first signal dr supplied to the comparator 204 can be generated in one of two ways. More specifically, the plant statei-gnal xp and Me input cosnmard signal d,, can tic supplied to the plant model 210 that includes an inversion Function to generate the first signal d' based on these signals. Altematively, the psendo-conol signal X! can be supplied as the first signal dr directly to the comparator 204. Me command control signal dC,4 is supplied to Me plant model 202 that includes an actuator 'nodal 206 for one or more characteristics of the acntawr 16 to be hedged, and a plant model 208 with one or more characteristics of tibe plant 12 to be hedged. The command control signal d,,,d is fed to the actuator model 206 to generate command signal S. Me command signal is supplied to We plant model 208 along win Me plant st Ate signal x' for use in generating We second signal d2 supplied to the comparator 204.

Me comparator 204 generates the hedge signal vh by subtracting the first and second signals ah, do to isolate the effect of We charactensUcs to be hedged.

In Fig. 9 a relatively detailed example of the reference model unit 22 Is shown.

The reference model unit 22 in dais case is second-order, and has constant, denvive, as and double derivative terms. Me reference model unit 22 includes a multiplier 226, a summing unit 228, integrators 230, 232, and multipliers 234, 236. The multiplier 226 multiplies die commanded state signal x, by a predetermined constant en2 to generate a modified signal supplied to the summing unit 228. The summing unit 228 subtracts signals Font the multipliers 234, 236 from the modified signal from the multiplier 228 to generate the reference model signal v,=. The summing Bunt 228 supplies the Preference model signal v,,, to Me controller 14. The sunning unit 228 also supplies We reference model signal vrm to the comparator 220. The comparator 220 substracts the hedge signal Ah to the reference model signal vrm to generate a signal supplied to the integrator:230. The integrator 230 integrates the signal from comparator 220 to generate integrated signal arm. Ike integrated signal or,,, is supplied to the integrator 232 to generate the reference model state signal x,,,, that in this case has a inconstant" term, a derivative tern, and a seconderivative term. The reference model state signal x,, is supplied to the comparator 24. The integrated signal By,, from the unit 230 is also supplied to the multiplier 234 that multiplies this integrated signal by the constant 25mn in which ú and On are constants, and supplies the resulting signal to the summing trait 228, The "constant" term from the reference model state signal lo,,' is also supplied to the multiplier 236 that multiplies this signal by the constant one to generate a signal supplied t, the sag unit 228. The signals from multipliers 234, 236 are subtracted from the signal from multiplier 226 to generate the reference model signal V,7,,.

Fig. 10 is a diagram of a neural network 260. Ibe neural networl; 260 includes an input layer 262, a hidden layer 264, and an output layer 266. The input layer 262 has Nit nodes receiving elements of the plant state signal x and Me pseudo-control signal vie, N being a positive integer. The North nodes Of the input layer 262 are multiplied by respective connection weights V to generate the Input signals to Me N2 nodes of the hidden layer 264, 2 fig a positive integer. The weighty input sits to Me hidden layer 264 are supplied as input signals to the signmoidal activation fimction or (z) of the form: ( ) 1 e= (l) in which a is a predetermined constant and z represents the V-weighted input signals from the Lnput layer 262. The outputs from the hidden layer 264 are weighted by the connection eights W. and are supplied as input signals to respective nodes 1 - N3 of the output lryer 266. These nodes add respective input sisals to generate the adaptive C02l01 signal -vale The mapping of the plant state signal up lo We adaptive control signal van pe,rfomed by Me neural network 260 can. be expressed as: Mad Y i = [ w <' ( v j' X k + b V, ) w] (2) i = 1,2, .. , 1\r 3 where No, N2, N3 are We number of nodes the input, hidden, and output layers 262, 264, 266, respectively, referenced by correspond indexes k, J. i. The connection weights vat and wig are set adaptively by the states of the tracking error signal e and Me N2 inputs to the neural network input layer 262 The constants by md be' are predetermined. Irk matrix form equation (2) can be expressed as: V Àd = Y = W a (V X) (3) in which:c is the neural network input signal, LIT iS Me transpose of the connection weight vector V, c, is the sigmoidal activation cdor, WT is Me transpose of the _. . connection weight vector W. and y c va' is the adaptive control signal. The signal VG can either be multiplied by "-1 " of mapped by the neural network 260 to generate the signal -va' for supply to the controller 14.

Tile manner in which the hacking error signal e and the input layer signal are used to adapt the connection weights V and W is now described. The pseudoontrol signal vpc is generated using the reference model signal v,,,,,.the linear control signal vie, and the adaptive control signal va, 7 as follows: Air = V All V is - ad (4) The pseudoontrol signal vpc is related to the acceleration term of the reference model state. We pseudocontrol signal can be fiercer augmented by terms may be required To SUppO31 proof of Loudness. example of a ten cooy refeed to the robustifyg term and is well know to those of ordinary skill Me art. Dyc inversion is used to reduce the control design problem to that of a control design for a linear, time-invariant plant. However, as is well-known to those of ordinary skill in this technology, use of an imperfect model In the dynamic inversion process can corrupt the desired relationship between acceleration of the plant state vector and the pseudo- control by an amount often referred to as an inversion error. This relationship bee acceleration of the plant state, pseudo-conol, and the inversion error is defined in Equation (5).

x p = v, (5) The derivative of the Lacking error signal e can be expressed as: e - A e + b (1\ v 'd (6) in which A is Hurwitz. Me output of the neural network vad is used to approximate die nversior error, so that the error dynamics of Equation (6) will remain bounded, and Lacking error is mined. The constant is defined by the equation: = X T Pb (a) in which!T is the transpose of Me tracking error signal e, b is a predestined matrix constant from Equation (6), and P is the solution of a Lyapunov equation (8).

A T p + PA = - Q (a) in which Q is a positive definite matrix. The adaptation law for updating the neural network weights and implemented by Me adaptation law mat 26 be expressed as: - - LAW rat'+ As I I rv (g) W = -rwL(o - evr-xy + I hi Wl (10) in which is the derivative of the connection weight vector V of the mural network 26O, W is Me derivative ofthe W weight vector of the neural nerworlc 260, '(a) is Me partial deri,ative of signmoidal function (z) v qth respect to, v, kw, (, Ev. w are predetenruried vectors, and x is the plant state signal.

For the case of a second-order reference model (Figs. 7-10) the remaining Xp = f (xp,xp,dane (din)) (11) Xrm V rm V h frm (Xrn Xrm Xr) h (12) vC =Kp(X,m Xp)+ K D(XrR' Xp) (13) V p<, = V m + Ic V ad (14) V =Vpc- f(X,Xp, d(dm)) (15) d in = f (X p, X p,V po) (16) characteristics are: wbich.Kp and 1 are predeennined consants. }3quation (16) correspords to e dync i:nversion ut 148 of Fig. 7. Tbe tracking error signal e can be expressed as :he vector of dierences: e= f rm XP1 (17) LX,m -Xf7,] It cal1 be shown that whether = din a dr,'d or S d,,, ó dC,z,d, the following equation holds: e - Ae + biA(xp, xp, dcm)+ Vad (18) Hence, dle to the hedge signa1 vh generated by the hedge unit 20, e adapove control system 13 is bounded with respect to tracking error and neural network weights. The plant will track the desired response as close as is possible wiin the its of dcmd Withoul: the hedge unit 20, it can be shown at the system 10 would not be stable for cases in,ch din d.

h Fg. 11 a special case of the cont:rol allocation and charactenstic unit 142 is shown. More specifically, the urut 142 includes an operator interface urut 144, and operator 146, and a commad uDit 148. The operator interface unit 144 receives e input co.mmand signal dn and e plant state signal xp and generates a sigral or display based or, these sigrs. The operator 146 can be a buman operator, a computer, or orher machine, for example. In the case of a human operator, the operator interface unit gererates a display based on the input command signal din and Me plant state signal up.

The operator 146 uses the display *tom the operator interface unit 144 to control the s cormnand mat 148 to generate the command control signal dC,, ,d. The command control signal darid is supplied to the actuator 16 to control We state of the plant. The configuration of the unit 142 in Fig. 1 1 is resell in numerous contexts. For example, in an aircraft, it may be desirable to have an auto-pilot whose control of Me aircraft is limited. Ire situations in which it is not desirable for the auto-pilot to control the aircraft, such as take-off or landing in which emergency maneuvers are more likely to be required, the aircrah's flight control system can be implemented to switch the operator 146 into the control loop with aunts 144 and 148 as shown in Fig. 11. In this configuration, control and response characteristic(s) generated by the operator are hedged by the hedge tout 20. One barrier to implementation of adaptive control systems in contexts such as aircraft is the stringent testing and certification required of adaptive control systems. Certification is made difficult by the fact fbat it may be exceedingly difficult or Possible to subject the adaptive control system to all plant states it is likely to encounter. The corfiguTation of Fig. 11 provides the advantage of permitting the use of an adaptive control system in which the pilot operator can control the aircraft without causing Me control system to adapt to the pilot's control and response. Accordingly, the configuration of Fig. 121 should facilitate tesg and certification of an adaptive control system incoporanug the features of the unit 142 in Fig. 1 1.

The adaptive control system 10 be used numerous applications. POT example, the plant 12 can be a manned or unmanned vehicle, Such vehicle can be an aircraft spacecraft missile, or guided orifice. general, the adaptive control system 1 () is assigned to control one degree of freedom of We plant 12. Ike actuator 16,, Me sensor 18, the operator interface unit 28, the operator 30, and Me command unit 32 depend upon the nature of the plant 12 and *be degree of freedom thereof to be controlled by the adaptive control system 1 0. For example, if the plant 12 is a Added vehicle such as an aircraft, spacecraft, missile or other guided ordinance, Me actuator 16 could be a motor, a motordriven screw, a hydraulic cylinder or over device attached to a control surface such as an aileron, rudder, or stabilizer. Altematively, the actuator 16 could be a pump or valve that generates air fetes) to charge he flow of air as over Me guided vehicle's surface, or a frame actuator that changes He shape of the guided vehiole's surface. In addinon, the actuator 16 could be Trust controllers to control the direction ofthrust generated by a power plant of the aircraft. Such actuators can be used to control Me degree of freedom (e.g. , pitch, roll, or yaw) that is controlled by the adaptive control system 10. In the guided vehicle context, the sensor 18 can be a gyroscope or other device to measure the degree of freedom controlled by the actuator 16. to the case of a manned vehicle, We operator 30 can be a human, the operator interface unit 28 a display, and the command unit 32 a control sticlc and/or flight control system, for example If the plant 12 is an automobile, the actuator 16 can be a valve for a Mel injection post, a hydraulic cylinder to move a braking element into contact wit. a brake drum, a transmission or other element. In this case, the sensor lS can be a slxedometer, a pressure sensor in an engine cylinder, an inertial sensor, or over elements. The plant 12 could also be a satellite, and the actuator 16 could be a thruster to orient arid position the satellite in orbit. The satellite's actuator 16 could be a motor-driven electromecical device to position a solar panel or transceiver unit in a desired direction. In the satellite context, We sensor 18 could be a gyroscope, for example As another example, the operator 30 can be a combination of art auto-pilot and human operator to take control of the plant in circumstances in which the auto- pilot is nest to control the plant Such implementation can be used in aircraft, for example The command unit 32 can be programmed to switch control between a machine and human operator to control the aircraft plant 12. The hedge unit 20 can generate Ale hedge signal to hedge characteristics of the human control of the command unit 32 for stable control of the aircraft plant. It should be understood that the use of a vehicle context In the foregoing description is exemplary only, and is not intended to limit the scope or context in which the invented adaptive control system 10 Carl be used.

1-hose of ordinary skill in the art should under that the system lo can be used in numerous other contexts and environments, such as maufactunog plants, power generation stations, and numerous other types of plants.

Any trademarks listed herein are the property of their respective owners, and reference herein to such ademarlts is intended only to indicate Me source of a particular product or service.

The many features and advantages of the present invention are apparem from the derailed specifc:ion and it is intended by the appended claim to cover all such feanres and advantages of the described methods and apparatus which follow Me true scope of the uven6on. Furler, since numerous modifications and changes mill ! readily occur to those of ordinary skill in the aft, it is not desired to limit the invention to the exacttrnplementation and operation illustrated and descnbed. Accordingly, all suntable modifications and equivalents may be resorted to as falling within the scope of the mventio'.

Industrial Applicability

-. This invention can be applied to Me control of a plant such as an aircraft, spacecraft, space-launch vehicle, satellite, missile, guided munition, automobile, or other vehicle, a robot, or a pointing or orientation system such as a satellite orientation system to orient powergeneration panels, a transceiver, or a docking mechanism, a braking system, an engine, a transmission, or an active suspension, or other vehicle subsystem, a manufacturing facility, a power generation facility, or virtually any controllable system. The invention thus has widespread applicability numerous industries

Claims

Claims 1. A method executed by an adaptive control system, the method

comprising the steps of: s a) generating an input control signal based on at least one of a reference model state signal, a commanded state signal, a plant state signal, and an adaptive control signal; b) generating a command control signal based on at least one of a commanded state signal, a plant state signal, an adaptive control signal, and further based on control allocation and a control characteristic of a controller used to generate the command control signal; c) supplying the command control signal to an actuator; d) controlling a state of the plant based on the command control signal; d) sensing a state of the plant; e) generating a plant state signal based on the sensing of the step (d); f) generating a first signal based on the input control signal, the plant state signal, and a plant model without a plant characteristic for which the adaptive control system is not to adapt; g) generating a second signal based on the command control signal, the plant state signal, and a plant model with the plant characteristic for which the adaptive control system is to adapt; h) generating a hedge signal by differencing the first and second 2s signals; i) generating a reference model state signal by modifying the commanded state signal with the hedge signal to include the effect of the control allocation and control characteristic on plant state from the reference model state signal; j) comparing the plant state signal and the reference model state signal; k) generating a tracking error signal based on the comparing of the step (I); and 1) generating the adaptive control signal based on the tracking error signal.
2. A method as claimed in claim 1 further comprising the step of: m) generating the reference model signal based on the s commanded state signal, the hedge signal, and a reference model representing the target response of the plant, the reference model signal used in the step (a) to generate the input control signal.
3. A method as claimed in claim 1 further comprising the step of: m) generating the reference model signal based on the commanded state signal, the hedge signal, and a reference model representing the target response of the plant, the reference model signal used in the step (b) to generate the command control signal.
4. A method as claimed in claim 1 further comprising the step of: m) generating a linear control signal based on the tracking error signal; n) generating a reference model signal based on the commanded state signal, the hedge signal, and a reference model; and o) generating a pseudocontrol signal based on the linear control signal, the reference model signal, and the adaptive control signal, the pseudo-control signal used in the generation of the adaptive control signal in the step (1).
5. A method as claimed in claim 4 wherein the adaptive control signal is generated in the step (1) based on the plant state signal, the step (1) performed by a neural network having connection weights adjusted based on the tracking error signal and the pseudo-control signal, the neural network mapping the plant state signal to the adaptive control signal in the performance of the step (1).
6. A method as claimed in claim 4 wherein the plant state signal is used in the step (1) to generate the adaptive control signal.
7. A method as claimed in claim 1 further comprising the step of: m) generating the commanded state signal based on a control action from an operator.
8. A method as claimed in claim 7 wherein the operator is human, the method further comprising the step of: n) generating a display based on the plant state signal, the display used by the operator to generate the commanded state signal in the step (m).
9. A method as claimed in claim I further comprising the step of: m) generating the commanded state signal based on a signal s generated by an operator that is a computer.
10. A method as claimed in claim 1 further comprising the step of: m) generating a display for an operator based on the input control signal, the operator generating the command control signal based on the display.

0
11. A method as claimed in claim 1 wherein the plant is an aircraft and/or spacecraft.
12. A method as claimed in claim 1 wherein the plant is an automobile.
13. A method as claimed in claim 1 wherein the plant is an unmanned vehicle.
14. An adaptive control system coupled to receive a command state signal indicative of a target state of a plant controlled by the adaptive control system, the adaptive control system comprising: a controller coupled to receive the commanded state signal, a plant state signal, and an adaptive control signal, the controller generating an input command signal based on the commanded state signal, the plant state signal, the adaptive control signal, and a control model, and the controller generating a command control signal based on the commanded state signal, the plant state signal, the adaptive control signal, the control model, control allocation of the controller, and at least one control characteristic of the controller, the controller coupled to supply the command control signal to the plant to control the plant's state; an actuator coupled to receive the command control signal, and affecting physical control of the plant's state based on the command control signal; a sensor coupled to sense the plant state, and generating a plant state signal based on the sensed plant state; a hedge unit coupled to receive the input control signal, the command control signal, and the plant state signal, and generating a hedge signal to modify the command state signal based on the input control signal, the command control signal, the plant state signal, and a hedge model indicative of a characteristic of at least one of the plant and the adaptive control system, to remove the effect of the characteristic on a tracking error signal; a reference model unit coupled to receive the command state signal and the hedge signal, the reference model unit generating a reference model s state signal based on the commanded state signal and a hedge signal; a comparator unit coupled to receive the reference model state signal and the plant state signal, and generating a tracking error signal based on a difference between the plant state signal and the reference model state signal, and an adaptation law unit coupled to receive the tracking error signal, and generating the adaptive control signal based on the tracking error signal, the adaptation law unit coupled to supply the adaptive control signal to the controller.
15. An adaptive control system as claimed in claim 14 wherein the controller generates the input control signal and the command control signal further based on the reference model state signal.
16. An adaptive control system as claimed in claim 14 wherein the characteristic is a time delay between generation of the commanded state signal by the controller at a time, and receipt by the controller of the plant state signal resulting from the commanded state signal generated at the time.
17. An adaptive control system as claimed in claim 14 wherein the characteristic is a time delay between generation of a state by the plant in response to the command control signal, and sensing of the state of the plant resulting from the command control signal.
18. An adaptive control system as claimed in claim 14 wherein the characteristic pertains to a control limit of the actuator used to control the plant.
19. An adaptive control system as claimed in claim 18 wherein the control limit pertains to actuator end points.
20. An adaptive control system as claimed in claim 18 wherein the control limit pertains to actuator dynamics.
21. An adaptive control system as claimed in claim 18 wherein the control limit pertains to a rate limit of the actuator.
22. An adaptive control system as claimed in claim 18 wherein the control limit pertains to quanitzation effects of the actuator.
23. An adaptive control system as claimed in claim 14 wherein the commanded state signal is generated by an operator, the adaptive control system s further comprising: an operator interface unit coupled to receive the plant state signal, the operator interface unit relaying the plant state to the operator; and a command unit operable by the operator, and generating the command state signal based on the operator's control action.
24. An adaptive control system as claimed in claim 4 wherein the operator interface is a display generated based on the plant state signal, and the operator is a human being that generates the control action to the command unit to generate the commanded state signal.
25. An adaptive control system as claimed in claim 14 wherein the commanded state signal is generated by a machine operator based on the plant state signal.
26. An adaptive control system as claimed in claim 14 wherein the adaptation law unit characterizes a neural network having connection weights determined by the tracking error signal, the neural network mapping the plant state signal to the adaptive control signal based on the connection weights to generate the adaptive control signal.
27. An adaptive control system as claimed in claim 26 wherein the controller generates a pseudo-control signal based on the commanded state signal and the plant state signal, the controller coupled to supply the pseudo-control signal to the neural network to adjust the connection weights of the neural network.
28. An adaptive control system as claimed in claim 14 wherein the controller includes a dynamic inversion unit to generate the command control signal.
29. An adaptive control system as claimed in claim 14 wherein the input control signal is used to generate a display, and the operator generates a command control signal based on the display.
30. An adaptive control system as claimed in claim 14 wherein the plant is an aircraft and/or spacecraft.
31. An adaptive control system as claimed in claim 14 wherein the plant is an automobile.
32. An adaptive control system as claimed in claim 14 wherein the plant is an unmanned vehicle positioned remotely from the operator.