CN105137763B - Supersonic motor robustness recursion neutral net Variable Structure Control system and method - Google Patents
Supersonic motor robustness recursion neutral net Variable Structure Control system and method Download PDFInfo
- Publication number
- CN105137763B CN105137763B CN201510651079.9A CN201510651079A CN105137763B CN 105137763 B CN105137763 B CN 105137763B CN 201510651079 A CN201510651079 A CN 201510651079A CN 105137763 B CN105137763 B CN 105137763B
- Authority
- CN
- China
- Prior art keywords
- control
- recursion
- represent
- supersonic motor
- neutral net
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Feedback Control In General (AREA)
Abstract
The present invention relates to a kind of supersonic motor robustness recursion neutral net Variable Structure Control system and method, the system includes pedestal and the supersonic motor being provided thereon, supersonic motor side output shaft is connected with photoelectric encoder, opposite side output shaft is connected with flywheel inertia load or direct current generator, the output shaft of flywheel inertia load or direct current generator is connected through shaft coupling with torque sensor, and photoelectric encoder, the signal output part of torque sensor are respectively connected to control system.The control system is made up of robustness recursion neutral net sliding mode control and supersonic motor, the system of whole controller is established on sliding mode, also using smooth face as its Tuning function in the design of robust controller, so as to obtain more preferable controlled efficiency.Not only control accuracy is high and simple in construction, compact by the present invention, and using effect is good.
Description
Technical field
The present invention relates to Motor Control Field, particularly a kind of supersonic motor robustness recursion neutral net sliding die
State control system and method.
Background technology
Total collection indeterminate is considered in the design of existing supersonic motor recursion nerve network control system, and it is total
Collection indeterminate contains the cross-linked disturbance occurred in drive system.In order to improve the control effect followed, Wo Menshe
The equivalent control that robustness recursion neutral net Variable Structure Control system is come in approximate Variable Structure Control system is counted.From
In the experimental result that a variety of tracks follow, it has been found that system has a significant improvement in motion tracking effect, and parameter
The factors such as variation, noise, cross-linked interference and frictional force can not almost impact for kinematic system effect, therefore robust
Property recursion neutral net Variable Structure Control system can effectively promote the controlled efficiency of system, and further reduce system pair
Preferable dynamic characteristic can be obtained in the Position And Velocity control of probabilistic influence degree, therefore motor.
The content of the invention
The purpose of the present invention is to propose to a kind of supersonic motor robustness recursion neutral net Variable Structure Control system
And method, not only control accuracy is high, and simple in construction, compact, and using effect is good.
The system of the present invention is realized using following scheme：A kind of supersonic motor robustness recursion neutral net sliding die
State control system, including pedestal and the supersonic motor on pedestal, the side output shaft and photoelectricity of the supersonic motor
Encoder is connected, and the opposite side output shaft of the supersonic motor is connected with flywheel inertia load or direct current generator, described
The output shaft of flywheel inertia load or direct current generator is connected through yielding coupling with torque sensor；The photoelectric encoder
Signal output part, the signal output part of the torque sensor are respectively connected to control system.
Further, the control system includes supersonic motor drive control circuit, the supersonic motor driving control
Circuit processed includes control chip circuit and driving chip circuit, the signal output part of the photoelectric encoder and the control chip
The respective input of circuit is connected, the output end of the control chip circuit and the respective input of the driving chip circuit
It is connected, to drive the driving chip circuit；The driving frequency Regulate signal output end of the driving chip circuit and driving
Respective input of the half-bridge circuit Regulate signal output end respectively with the supersonic motor is connected.
The method of the present invention is realized using following scheme：It is a kind of based on supersonic motor robustness recursion described above
The method of neutral net Variable Structure Control system, recursion neutral net sliding mode control is located at the control chip
In circuit, the recursion neutral net sliding mode control is established on sliding mode, and adjusted using smooth face as it
Function, to obtain more preferable controlled efficiency.
Further, the dynamic equation of the recursion neutral net sliding mode control can be represented as follows：
Wherein α1, α2, α3, α4And α5It is all positive number；It is indeterminate H estimated value；Ap=-B/J, BP=J/Kt＞
0,CP=-1/J；B is damped coefficient, and J is rotary inertia, KtFor current factor, U (t) is the output torque of motor, AnFor ApIt
Standard value, BnFor BPStandard value, S (t) is smooth face, and W is nonlinear function, and u (t) is the control input of an auxiliary, UrIt is
Robust controller, d, v and r are the parameters in neutral net, F ∈ R1×KFor the adjustable power from hidden layer to output layer
Weight vector.
It is preferred that the principle of the present invention is further as follows：
The dynamical equation of supersonic motor drive system can be written as：
Wherein Ap=-B/J, BP=J/Kt＞ 0, CP=-1/J；B is damped coefficient, and J is rotary inertia, KtFor current factor,
Tf(v) it is frictional resistance torque, TLFor load torque, U (t) is the output torque of motor, θr(t) it is to be surveyed by photoelectric encoder
The position signalling measured.
The parameter for first assuming system now is all known, and External force interference, cross-couplings interference and frictional force are all not deposit
, then the master pattern of motor is shown in following formula：
Wherein AnFor ApStandard value, BnFor BPStandard value.
If producing indeterminate, (such as system parameter values deviate from standard value or External force interference occurs in system, intersect
Coupled interference and friction torque etc.), now the dynamical equation of control system is modified as：
Wherein CnFor CPStandard value, Δ A, Δ B, Δ C represent change, and D (t) is total collection indeterminate, definition
For：
The border of total collection indeterminate is assumed to be, it is known that such as by we herein | D (t) |≤ρ, ρ be one it is given just
Constant term.In order to avoid occurring not expected indeterminate in motor, we are slided using robustness recursion neutral net
Modal control system is controlled to system.
In order to reach the purpose of control, exactly it is that finding a control law causes state variable θr(t) can follow
Upper reference command θm(t)。
Define tracking error e (t)=θm(t)-θr(t) (5)
Wherein θm(t) motion control commands of motor are represented.
Defining smooth face is：
Wherein λ is positive constant value.By S (t) to t differential, utilize (3), can obtain：
When designing Variable Structure Control system, it is necessary first to obtain equivalent control power of the system on smooth face.These
Effect controling power can be obtained by following formula：
Formula (7) is brought into formula (8), can be obtained
(9) formula of solution a, wherein solution is as follows：
SinceThen the dynamic characteristic of system sliding mode represents as follows in t >=0：
After selecting appropriate λ value, dynamic characteristic such as rise time, the amount of surmounting and stabilization time etc. required by system all may be used
With simple designs into a second-order system.If the parameter of system determines that then formula (11) will be invalid, the stability of such system
It will be destroyed.In order to ensure the stability of system in the above cases, the Shandong based on control design case is carried out below
Rod recursion neutral net sliding mode control designs.
From (6), (7) and (8), preferable equivalent control rule (9) can be changed into：
Wherein W is nonlinear function, and it is defined as follows：
In order to want approximate ideal equivalent control rule, designed as follows：
Ueq(t)=W-u (t) (14)
Wherein u (t) is the control input of an auxiliary.
(14) formula is substituted into (12) formula, then closed loop system becomes
In actually controlling, u (t) can be PID controller, and its design rule is as follows：
Wherein KS, KPAnd KIIt is control gain.We can choose KPAnd KIIt is as follows：
KP=KS×2λ；KI=KS×λ2 (17)
Formula (17) is substituted into formula (18), can be obtained
U (t)=- KSS(t) (18)
By formula (18), it is as follows that new closed loop control system can be retrieved：
It is as follows that we define Liapunov function：
By formula (20) to substituting into formula (19) after time diffusion, can obtain：
Due toThereforeTo bear semidefinite, i.e. V1(S(t))≤V1(S (0)), wherein S (t) is bounded
's.
Assuming that functionWith integral function Γ1(t) it is all time variable,
V1(S (0)) bounded and V1(S (t)) is the nonincreasing function of a bounded, therefore can obtain following result：
Because Γ1And bounded, according to Barbara lemma,Therefore work as S (t) → 0 item t → ∞,
Therefore it is stable for can determine that control design case, therefore the tracking error of control system converges to 0 in S (t) → 0.
Further, the design of robustness recursion neural network is carried out：
In formula (13), it is contemplated that many probabilistic influences of nonlinear function W, it is outside such as the variation of mechanical parameter
Noise, the cross coupling effect and frictional force of axle and between centers etc..Because the variation of systematic parameter is not easy acquisition and noise, intersection
The influence of coupling and frictional force also can not all obtain a definite numerical value, so in the application of reality, these indeterminates
It is difficult to learn in advance to be all, therefore formula (14) is almost what can not be realized.Therefore, it is proposed that controller such as formula (24) is used for closely
Like nonlinear function W：
WhereinFor intelligent controller, learning of nonlinear functions W is can be used to, it is defined as follows：
WhereinIt is that recursion neutral net exports, UrIt is robust controller.Recursion neural networkCan be with
For learning nonlinear equation.Due to the uncertainty of system, we devise robust control UrCome compensate W andBetween
Difference.
Further, recursion neural network design is carried out：
One three layers of recursion neutral net contains input layer, hidden layer and output layer, and using Gaussian function as it
Function is triggered, with following formula subrepresentation：
Y=WRNN(x,d,v,r,F)≡F (26)
Wherein y is the recursion neutral net of single output；F∈R1×KIt is adjustable from hidden layer to output layer for one
Weight vector；K is the number of nodes of hidden layer；T∈RK×1It is the output vector of hidden layer；It is recurrence
The input vector of formula neutral net；vikAnd dikIt is center and the width of Gaussian function respectively；rkIt is internal feedback oscillator；It is weighed
Weight values can represent as follows：
For the recursion neutral net of formula (26), can uniform approximate non-linear function, even time-varying
Equation.Due to its approximate characteristic, a preferable recursion nerve network controller can be usedIt is nonlinear to learn this
Function W, W can represent as follows：
Wherein ε is minimum reconstruction error；d*, v*And r*Optimized parameter d in recursion neural network respectively, v and
r.Therefore following formula can be obtained
WhereinWithAll it is the optimization parameter then estimated with adaptive algorithm by condition.Then will
(28) formula subtracts (29) formula, approximate errorIt is defined as follows：
WhereinWithWe are with a kind of method of linearisation by nonlinear recursion nerve net
Network function is converted into the form of partial linear, is obtained under Taylor seriesCurved-edge polygons：
WhereinT*It is T optimization parameter；It is T*Estimation parameter；
Onv∈Rj×1It is the vector of high order part.
Then formula (31) is substituted into formula (30)：
WhereinFor indeterminate.According to (12,15,18,24,30 and 32)
Equation, dynamic equation can represent as follows：
Wherein α1, α2, α3, α4And α5It is all positive number；It is indeterminate H estimated value.
Use Liapunov function：
Formula (40) further can be obtained into following formula to time diffusion and using formula (32)：
If formula (34-37) is the adaptation rule of recursion neutral net, robust Controller Design is formula (38), and it is estimated
Method of determining and calculating is formula (39), then (41) can remodify into following formula：
It is negative semidefinite, i.e.,
This demonstrate that S (t),
WithAll it is to have dividing value.Make functionIt can be obtained to time integral：
BecauseBe have dividing value andBeing one nonincremental has
Dividing value, so it is as follows to obtain result：
To there is dividing value.Proved by Barbara lemmaTherefore work as S (t) → 0 item t → ∞.
We carry out the anglec of rotation of controlled motor rotor using recursion neutral net sliding mode control.
Compared with prior art, the present invention uses supersonic motor robustness recursion neutral net Variable Structure Control system
System, system have variation, noise, cross-linked interference and the frictional force of significant improvement and parameter in motion tracking effect
It can not almost be impacted etc. factor for kinematic system effect, therefore robustness recursion neutral net Variable Structure Control system
The controlled efficiency of system can effectively be promoted, and further reduce system for probabilistic influence degree, improve control
Accuracy, preferable dynamic characteristic can be obtained.In addition, the device is reasonable in design, simple in construction, compact, manufacturing cost is low,
With very strong practicality and wide application prospect.
Brief description of the drawings
Fig. 1 is the structural representation of the embodiment of the present invention.
Fig. 2 is the control circuit schematic diagram of the embodiment of the present invention.
[primary clustering symbol description]
In figure：1 is photoelectric encoder, and 2 be photoelectric encoder fixed support, and 3 be supersonic motor output shaft, and 4 be ultrasound
Ripple motor, 5 be supersonic motor fixed support, and 6 be supersonic motor output shaft, and 7 be flywheel inertia load or direct current generator, 8
It is yielding coupling for flywheel inertia load or direct current generator output shaft, 9,10 be torque sensor, and 11 consolidate for torque sensor
Fixed rack, 12 be pedestal, and 13 be control chip circuit, and 14 be driving chip circuit, and 15,16,17 be photoelectric encoder output
A, B, Z phase signals, 18,19,20,21 be driving frequency Regulate signal caused by driving chip circuit, and 22 be driving chip circuit
Caused driving half-bridge circuit Regulate signal, 23,24,25,26,27,28 be driving chip circuit caused by control chip circuit
Signal, 29 be supersonic motor drive control circuit.
Embodiment
Below in conjunction with the accompanying drawings and embodiment the present invention will be further described.
As shown in figure 1, present embodiments provide a kind of supersonic motor robustness recursion neutral net sliding mode control
System processed, including pedestal 12 and the supersonic motor 4 on pedestal 12, the side output shaft 3 and light of the supersonic motor 4
Photoelectric coder 1 is connected, and the opposite side output shaft 6 of the supersonic motor is connected with flywheel inertia load or direct current generator 7,
The output shaft 8 of the flywheel inertia load or direct current generator 7 is connected through yielding coupling 9 with torque sensor 10；The light
The signal output part of photoelectric coder 1, the signal output part of the torque sensor 10 are respectively connected to control system.
Above-mentioned supersonic motor 4, photoelectric encoder 1, torque sensor 10 are respectively through supersonic motor fixed support 5, light
Photoelectric coder fixed support 2, torque sensor fixed support 11 are fixed on the pedestal 12.
As shown in Fig. 2 in the present embodiment, the control system includes supersonic motor drive control circuit 29, described
Supersonic motor drive control circuit 29 includes control chip circuit 13 and driving chip circuit 14, the photoelectric encoder 1
Signal output part is connected with the respective input of the control chip circuit 13, the output end of the control chip circuit 13 with
The respective input of the driving chip circuit 14 is connected, to drive the driving chip circuit 14；The driving chip electricity
The driving frequency Regulate signal output end on road 14 and driving half-bridge circuit Regulate signal output end respectively with the supersonic motor 4
Respective input be connected.The driving chip circuit 14 produces driving frequency Regulate signal and driving half-bridge circuit regulation letter
Number, frequency, phase and the break-make that A, B two phase PWM are exported to supersonic motor are controlled.By opening and turning off PWM ripples
Output controls the startup of supersonic motor and out of service；By the frequency of PWM ripples and the phase difference of two-phase that adjust output
Carry out the optimal operational condition of regulation motor.
The present embodiment additionally provides a kind of based on supersonic motor robustness recursion neutral net slip described above
The method of modal control system, recursion neutral net sliding mode control is located in the control chip circuit, by institute
Recursion neutral net sliding mode control is stated to establish on sliding mode, and using smooth face as its Tuning function, to obtain
Obtain more preferable controlled efficiency.
In the present embodiment, the dynamic equation of the recursion neutral net sliding mode control can be represented such as
Under：
Wherein α1, α2, α3, α4And α5It is all positive number；It is indeterminate H estimated value；Ap=-B/J, BP=J/Kt＞
0,CP=-1/J；B is damped coefficient, and J is rotary inertia, KtFor current factor, U (t) is the output torque of motor, AnFor ApIt
Standard value, BnFor BPStandard value, S (t) is smooth face, and W is nonlinear function, and u (t) is the control input of an auxiliary, UrIt is
Robust controller, d, v and r are the parameters in neutral net, F ∈ R1×KFor the adjustable power from hidden layer to output layer
Weight vector.
It is preferred that the principle of the present embodiment is further as follows：
The dynamical equation of supersonic motor drive system can be written as：
Wherein Ap=-B/J, BP=J/Kt＞ 0, CP=-1/J；B is damped coefficient, and J is rotary inertia, KtFor current factor,
Tf(v) it is frictional resistance torque, TLFor load torque, U (t) is the output torque of motor, θr(t) it is to be surveyed by photoelectric encoder
The position signalling measured.
The parameter for first assuming system now is all known, and External force interference, cross-couplings interference and frictional force are all not deposit
, then the master pattern of motor is shown in following formula：
Wherein AnFor ApStandard value, BnFor BPStandard value.
If producing indeterminate, (such as system parameter values deviate from standard value or External force interference occurs in system, intersect
Coupled interference and friction torque etc.), now the dynamical equation of control system is modified as：
Wherein CnFor CPStandard value, Δ A, Δ B, Δ C represent change, and D (t) is total collection indeterminate, definition
For：
The border of total collection indeterminate is assumed to be, it is known that such as by we herein | D (t) |≤ρ, ρ be one it is given just
Constant term.In order to avoid occurring not expected indeterminate in motor, we are slided using robustness recursion neutral net
Modal control system is controlled to system.
In order to reach the purpose of control, exactly it is that finding a control law causes state variable θr(t) can follow
Upper reference command θm(t)。
Define tracking error e (t)=θm(t)-θr(t) (5)
Wherein θm(t) motion command of motor is represented.
Defining smooth face is：
Wherein λ is positive constant value.By S (t) to t differential, utilize (3), can obtain：
When designing Variable Structure Control system, it is necessary first to obtain equivalent control power of the system on smooth face.These
Effect controling power can be obtained by following formula：
Formula (7) is brought into formula (8), can be obtained
(9) formula of solution a, wherein solution is as follows：
SinceThen the dynamic characteristic of system sliding mode represents as follows in t >=0：
After selecting appropriate λ value, dynamic characteristic such as rise time, the amount of surmounting and stabilization time etc. required by system all may be used
With simple designs into a second-order system.If the parameter of system determines that then formula (11) will be invalid, the stability of such system
It will be destroyed.In order to ensure the stability of system in the above cases, the Shandong based on control design case is carried out below
Rod recursion neutral net sliding mode control designs.
From (6), (7) and (8), preferable equivalent control rule (9) can be changed into：
Wherein W is nonlinear function, and it is defined as follows：
In order to want approximate ideal equivalent control rule, designed as follows：
Ueq(t)=W-u (t) (14)
Wherein u (t) is the control input of an auxiliary.
(14) formula is substituted into (12) formula, then closed loop system becomes
In actually controlling, u (t) can be PID controller, and its design rule is as follows：
Wherein KS, KPAnd KIIt is control gain.We can choose KPAnd KIIt is as follows：
KP=KS×2λ；KI=KS×λ2 (17)
Formula (17) is substituted into formula (18), can be obtained
U (t)=- KSS(t) (18)
By formula (18), it is as follows that new closed loop control system can be retrieved：
It is as follows that we define Liapunov function：
By formula (20) to substituting into formula (19) after time diffusion, can obtain：
Due toThereforeTo bear semidefinite, i.e. V1(S(t))≤V1(S (0)), wherein S (t) is bounded
's.
Assuming that functionWith integral function Γ1(t) it is all time variable,
V1(S (0)) bounded and V1(S (t)) is the nonincreasing function of a bounded, therefore can obtain following result：
Because Γ1And bounded, according to Barbara lemma,Therefore work as S (t) → 0 item t → ∞,
Therefore it is stable for can determine that control design case, therefore the tracking error of control system converges to 0 in S (t) → 0.
Further, the design of robustness recursion neural network is carried out：
In formula (13), it is contemplated that many probabilistic influences of nonlinear function W, it is outside such as the variation of mechanical parameter
Noise, the cross coupling effect and frictional force of axle and between centers etc..Because the variation of systematic parameter is not easy acquisition and noise, intersection
The influence of coupling and frictional force also can not all obtain a definite numerical value, so in the application of reality, these indeterminates
It is difficult to learn in advance to be all, therefore formula (14) is almost what can not be realized.Therefore, it is proposed that controller such as formula (24) is used for closely
Like nonlinear function W：
WhereinFor intelligent controller, learning of nonlinear functions W is can be used to, it is defined as follows：
WhereinIt is that recursion neutral net exports, UrIt is robust controller.Recursion neural networkCan
With for learning nonlinear equation.Due to the uncertainty of system, we devise robust control UrCome compensate W andIt
Between difference.
Further, recursion neural network design is carried out：
One three layers of recursion neutral net contains input layer, hidden layer and output layer, and using Gaussian function as it
Function is triggered, with following formula subrepresentation：
Y=WRNN(x,d,v,r,F)≡F (26)
Wherein y is the recursion neutral net of single output；F∈R1×KIt is adjustable from hidden layer to output layer for one
Weight vector；K is the number of nodes of hidden layer；T∈RK×1It is the output vector of hidden layer；It is recurrence
The input vector of formula neutral net；vikAnd dikIt is center and the width of Gaussian function respectively；rkIt is internal feedback oscillator；It is weighed
Weight values can represent as follows：
For the recursion neutral net of formula (26), can uniform approximate non-linear function, even time-varying
Equation.Due to its approximate characteristic, a preferable recursion nerve network controller can be usedIt is nonlinear to learn this
Function W, W can represent as follows：
Wherein ε is minimum reconstruction error；d*, v*And r*Optimized parameter d in recursion neural network respectively, v and
r.Therefore following formula can be obtained
WhereinWithAll it is the optimization parameter then estimated with adaptive algorithm by condition.Then will
(28) formula subtracts (29) formula, approximate errorIt is defined as follows：
WhereinWithWe are with a kind of method of linearisation by nonlinear recursion nerve net
Network function is converted into the form of partial linear, is obtained under Taylor seriesCurved-edge polygons：
WhereinT*It is T optimization parameter；It is T*Estimation parameter；
Onv∈Rj×1It is the vector of high order part.
Then formula (31) is substituted into formula (30)：
WhereinFor indeterminate.According to (12,15,18,24,30 and 32)
Equation, dynamic equation can represent as follows：
Wherein α1, α2, α3, α4And α5It is all positive number；It is indeterminate H estimated value.
Use Liapunov function：
Formula (40) further can be obtained into following formula to time diffusion and using formula (32)：
If formula (34-37) is the adaptation rule of recursion neutral net, robust Controller Design is formula (38), and it is estimated
Method of determining and calculating is formula (39), then (41) can remodify into following formula：
It is negative semidefinite, i.e.,
This demonstrate that S (t),
WithAll it is to have dividing value.Make functionIt can be obtained to time integral：
BecauseBe have dividing value andBeing one nonincremental has
Dividing value, so it is as follows to obtain result：
To there is dividing value.Proved by Barbara lemmaTherefore work as S (t) → 0 item t → ∞.
We carry out the anglec of rotation of controlled motor rotor using recursion neutral net sliding mode control.
The foregoing is only presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification, it should all belong to the covering scope of the present invention.
Claims (1)
1. a kind of supersonic motor robustness recursion neutral net Variable Structure Control system, including pedestal and on pedestal
Supersonic motor, it is characterised in that：The side output shaft of the supersonic motor is connected with photoelectric encoder, the ultrasound
The opposite side output shaft of ripple motor is connected with flywheel inertia load or direct current generator, the flywheel inertia load or direct current generator
Output shaft be connected through yielding coupling with torque sensor；The signal output part of the photoelectric encoder, the torque pass
The signal output part of sensor is respectively connected to control system；
Wherein, the control system includes supersonic motor drive control circuit, the supersonic motor drive control circuit bag
Include control chip circuit and driving chip circuit, the phase of the signal output part of the photoelectric encoder and the control chip circuit
Input is answered to be connected, the output end of the control chip circuit is connected with the respective input of the driving chip circuit,
To drive the driving chip circuit；The driving frequency Regulate signal output end and driving half-bridge circuit of the driving chip circuit
Respective input of the Regulate signal output end respectively with the supersonic motor is connected；
Wherein, recursion neutral net sliding mode control is located in the control chip circuit, by recursion god
Established through network sliding mode control on sliding mode, and using smooth face as its Tuning function, to obtain more preferable control
Efficiency processed；
Wherein, the dynamic equation of the recursion neutral net sliding mode control is represented as follows：
Wherein α1, α2, α3, α4And α5It is all positive number；It is indeterminate H estimated value；Ap=-B/J, BP=J/Kt＞ 0, CP
=-1/J；B is damped coefficient, and J is rotary inertia, KtFor current factor, U (t) is the output torque of motor, AnFor ApStandard
Value, BnFor BPStandard value, S (t) is smooth face, and W is nonlinear function, and u (t) is the control input of an auxiliary, UrIt is robust
Controller, d, v and r are the parameters in neutral net, F ∈ R1×KFor an adjustable weight arrow from hidden layer to output layer
Amount；
Wherein,Represent F actual value and the error of estimate；Represent F estimate, F ∈ R1×KFor one from hidden layer to
The adjustable weight vector of output layer；RepresentFirst derivative；Represent T estimate；TδRepresent Θ to the inclined of variable d
Derivative coefficient matrix,T=[Θ1Θ2…Θm], Θ1Θ2…ΘmFor in T it is each to
Amount； TvPartial derivative coefficient matrixes of the Θ to variable v is represented,TrRepresent Θ to the inclined of variable r
Derivative coefficient matrix, Representing matrixTransposition；RepresentFirst derivative；Table
Show d actual value and the error of estimate；Represent v actual value and the error of estimate；Represent r actual value and estimation
The error of value；RepresentFirst derivative；RepresentFirst derivative.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510651079.9A CN105137763B (en) | 2015-10-10 | 2015-10-10 | Supersonic motor robustness recursion neutral net Variable Structure Control system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510651079.9A CN105137763B (en) | 2015-10-10 | 2015-10-10 | Supersonic motor robustness recursion neutral net Variable Structure Control system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105137763A CN105137763A (en) | 2015-12-09 |
| CN105137763B true CN105137763B (en) | 2017-12-15 |
Family
ID=54723148
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510651079.9A Expired - Fee Related CN105137763B (en) | 2015-10-10 | 2015-10-10 | Supersonic motor robustness recursion neutral net Variable Structure Control system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105137763B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10603799B2 (en) * | 2017-06-30 | 2020-03-31 | L'oreal | System for use with encoded end effectors and related methods of use |
| CN107748809B (en) * | 2017-09-20 | 2021-01-26 | 苏州芯智瑞微电子有限公司 | Semiconductor device modeling method based on neural network technology |
| CN109240082A (en) * | 2018-08-25 | 2019-01-18 | 南京理工大学 | A kind of caterpillar mobile robot sliding formwork cloud model cross-coupling control method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336523A (en) * | 2013-06-19 | 2013-10-02 | 闽江学院 | Control characteristic testing device and control strategy of ultrasonic motor servocontrol system |
| CN203324452U (en) * | 2013-07-10 | 2013-12-04 | 闽江学院 | Ultrasonic motor transient characteristic testing device |
| CN103513181A (en) * | 2013-07-10 | 2014-01-15 | 闽江学院 | Ultrasonic wave motor transient characteristic testing device and control system thereof |
-
2015
- 2015-10-10 CN CN201510651079.9A patent/CN105137763B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336523A (en) * | 2013-06-19 | 2013-10-02 | 闽江学院 | Control characteristic testing device and control strategy of ultrasonic motor servocontrol system |
| CN203324452U (en) * | 2013-07-10 | 2013-12-04 | 闽江学院 | Ultrasonic motor transient characteristic testing device |
| CN103513181A (en) * | 2013-07-10 | 2014-01-15 | 闽江学院 | Ultrasonic wave motor transient characteristic testing device and control system thereof |
Non-Patent Citations (1)
| Title |
|---|
| 基于RBF神经网络的永磁同步直线电机全局滑模控制;马素君等;《组合机床与自动化加工技术》;20120731(第7期);第82-84、88页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105137763A (en) | 2015-12-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104065322B (en) | Method for controlling output feedback of motor position servo system | |
| CN105137763B (en) | Supersonic motor robustness recursion neutral net Variable Structure Control system and method | |
| CN104300863B (en) | A kind of adaptive sliding-mode observer method of varying load PMSM Speed | |
| Chi et al. | Fuzzy PID feedback control of piezoelectric actuator with feedforward compensation | |
| CN101986564A (en) | Backlash operator and neural network-based adaptive filter | |
| CN108880394A (en) | A kind of wavelet neural network position-less sensor of switched reluctance motor forecast Control Algorithm | |
| CN106329986B (en) | A kind of supersonic motor contragradience adaptive servo control method | |
| CN104155877A (en) | Brushless DC motor fuzzy control system based on genetic algorithm and control method thereof | |
| Dhanalakshmi et al. | Application of multi stage fuzzy logic control for load frequency control of an isolated wind diesel hybrid power system | |
| GB2383568A (en) | Steering control with magnetorheological damping | |
| CN106505909A (en) | Supersonic motor nonlinear dead-zone compensation control system and method based on generalized regression nerve networks | |
| Adeel et al. | Autonomous dual wheel self balancing robot based on microcontroller | |
| CN100433535C (en) | DC electric machine position control method | |
| CN105099263B (en) | Supersonic motor Position And Velocity control system based on generalized regression nerve networks identifier | |
| Danh et al. | Comparison of estimator-based compensation schemes for hydrostatic transmissions with uncertainties | |
| Cheng et al. | Fuzzy logic design of self-tuning switching power supply | |
| CN100593767C (en) | Control method of capacity transducer | |
| CN112039520A (en) | Digital phase-locked loop based on fuzzy RBF self-adaptive control | |
| CN106787940B (en) | A kind of improved supersonic motor contragradience adaptive servo control method | |
| CN106842957B (en) | Ultrasonic motor servo control method based on output feedback controller | |
| CN104238359A (en) | Control method of large electromechanical mixed inertia system | |
| Peng et al. | Adaptive recurrent cerebellar model articulation controller for linear ultrasonic motor with optimal learning rates | |
| Hwang et al. | Fuzzy MRAC controller design for vane-type air motor systems | |
| Tarik et al. | Ant colony optimization algorithm and fuzzy logic for switched reluctance generator control. | |
| Ma et al. | Parameters self-adjusting fuzzy PID control in magnetic levitation system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171215 Termination date: 20211010 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |