CN111796519A - Automatic control method of multi-input multi-output system based on extreme learning machine - Google Patents

Automatic control method of multi-input multi-output system based on extreme learning machine Download PDF

Info

Publication number
CN111796519A
CN111796519A CN202010539521.XA CN202010539521A CN111796519A CN 111796519 A CN111796519 A CN 111796519A CN 202010539521 A CN202010539521 A CN 202010539521A CN 111796519 A CN111796519 A CN 111796519A
Authority
CN
China
Prior art keywords
network
elm
control
output
input
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.)
Granted
Application number
CN202010539521.XA
Other languages
Chinese (zh)
Other versions
CN111796519B (en
Inventor
唐若笠
徐华徽
林巧
张尚煜
张鹏
周锦翔
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN202010539521.XA priority Critical patent/CN111796519B/en
Publication of CN111796519A publication Critical patent/CN111796519A/en
Application granted granted Critical
Publication of CN111796519B publication Critical patent/CN111796519B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses an automatic control method of a multi-input multi-output system based on an extreme learning machine, which comprises the following steps: 1) analyzing the equipment to be controlled, and determining a control target of the equipment and control variables associated with the equipment and each control target; 2) for each controlled variable uiEstablishing an extreme learning machine network in a controller; 3) adjusting the existing latest sample data in the training sample set; 4) training the corresponding network according to an ELM algorithm to obtain each trained ELM network; 5) after each network is completely trained, acquiring a single prediction result of each control variable from each network output end; 6) and the ELM networks are mutually input and output to carry out sufficient iteration, so that the effective value of each control variable under a specific control target is obtained and acts on the control object equipment. The invention is suitable forThe method is used for a system with the complex characteristics of multiple input and multiple output, strong coupling and time variation, and is beneficial to improving the accuracy and effectiveness of a control result.

Description

Automatic control method of multi-input multi-output system based on extreme learning machine
Technical Field
The invention relates to an automatic control technology, in particular to an automatic control method of a multi-input multi-output system based on an extreme learning machine.
Background
In the field of automatic control technology, for a system (controlled object) with complex characteristics such as multiple inputs and multiple outputs, strong coupling and time-varying property, a model of the controlled object must be accurately established for effective control, and the control effect is often closely related to the accuracy of a mathematical model. Even if a reliable mathematical model is built, the automatic control of the complex system is still very complex, and the control effect is often unsatisfactory.
As a brand-new technical means, the machine learning technology learns the existing data samples based on a specific network model structure and algorithm rules, and deeply excavates the internal rule characteristics contained in a large number of data samples, so that effective prediction output can be given to new data input.
If the complex system is controlled by adopting a machine learning technology, the input and output rules contained in a large number of data samples are deeply mined based on the learning training of the running data of the controlled object, and the controlled object is controlled by utilizing the rules, so that the method has strong technical feasibility.
Disclosure of Invention
The invention aims to solve the technical problem of providing an automatic control method of a multi-input multi-output system based on an extreme learning machine aiming at the defects in the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows: an automatic control method of a multi-input multi-output system based on an extreme learning machine comprises the following steps:
1) analyzing the equipment to be controlled, and determining a control target of the equipment and control variables associated with the equipment and each control target; noting the number of control targetsM, the number of control variables is n; the ith control variable is uiI is 1,2, …, n, and the jth control target is vj，j＝1,2,…,m；
The control object device is a multi-autonomous system used in industrial production and application equipment;
2) according to the number of the control variables, for each control variable uiEstablishing an extreme learning machine network in a controller: ELMi；
The extreme learning machine network comprises (m + n-1) input layer nodes, a plurality of hidden layer nodes and 1 output layer node; to limit learning machine network ELMiInput layer node is vj，j＝1,2,…,m；utT ≠ 1,2, …, n, and t ≠ i; output layer node is ui；
3) A control loop is started. Adjusting the current latest K groups of sample data in the training sample set according to the network structure of the extreme learning machine, setting the current latest K groups of sample data in the training sample set, and regarding the kth group of sample data DkArranging the input layer nodes and the output layer nodes of the extreme learning machine network into the available form of each ELM network;
4) training the corresponding network by using the training sample data of each trimmed ELM network according to an ELM algorithm to obtain each trained ELM network;
5) based on each ELM network which is completely trained in the step 4), the current value v of the control target of the equipment is obtainedj *And the current value u of the control variablei *Sending the data into corresponding nodes of each ELM network input layer, and obtaining single prediction results of corresponding control variables, namely the control variable u, from nodes of each network output layeriCorresponding network ELMiWhen the training is complete, it will
Figure BDA0002538401550000031
Inbound network ELMiInputting layer nodes, and calculating network output to obtain control variable uiSingle prediction of result ui-p 1；
6) Each ELM network is used for performing iteration for mutual input and output; the method comprises the following specific steps:
subjecting each ELM in step 5)iOutput value u of network output layer node i-p1 respectively sending the signals to nodes corresponding to the control variables of other ELM network input layers;
then, each ELM network recalculates the network output based on the latest input to obtain the next generation of prediction output, and further completes one iteration;
7) continuously repeating iteration of each ELM network according to the mode of the step 6) until the preset iteration number N is reached, stopping iteration, and recording the ELM of each current networkiRespectively of ui-pN；
8) Carrying out N iterations on each ELM network in the step 7) to obtain a final output value
u1-pN,...,ui-pN,...,un-pNThe control variable value is used as an effective output value of the controller and directly acts on the corresponding position of the controlled equipment; completing automatic control of the control object equipment according to the control target;
9) and (3) acquiring the latest input and output data of the controlled equipment to update the training sample set, ending the current control cycle, and jumping to the step 3) to start the next control cycle.
According to the scheme, the current value of the control variable in the first iteration in the step 5) is the current actual value of each control variable, namely the effective value output by the controller in the last control loop, and the current value of the control variable in each iteration is the predicted output value of the previous iteration of the output layer node of the ELM network.
According to the scheme, the network training in the step 4) is completed at one time without an iteration process, the control variable prediction in the steps 5) to 8) is completed for multiple times, and the ELM networks corresponding to the control variables need to be input and output mutually to perform sufficient iteration.
According to the scheme, the multi-autonomous system in the step 1) is a system with multiple inputs and multiple outputs, strong coupling and time-varying property.
The invention has the following beneficial effects:
1. the method is suitable for a system with the complex characteristics of multiple input and multiple output, strong coupling and time-varying property, and can accurately establish the model of the controlled object. Based on the learning training of the self-operation data of the controlled object, the input and output rules contained in a large number of data samples are deeply mined, and the rules are utilized to implement control operation on the controlled object, so that the learning problem of the multilayer neural network is effectively solved.
2. The control process of the invention is the continuous repetition of the control cycle, and the new data acquisition work is also blended into the control cycle with a certain frequency, thereby leading the whole control process to be the reciprocating process of 'control-operation-data acquisition and sample updating-ELM network retraining-re-control'. Therefore, even if the controlled object is disturbed or the working characteristics of the controlled object change due to time, the data for network training also change with the same property, and the ELM network can learn the change from the data, so that the time-varying characteristics of the controlled object are restrained to a certain extent, and the accuracy and the effectiveness of the control result are improved.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method of an embodiment of the present invention;
FIG. 2 is a diagram of an ELM network for the ith control variable in accordance with an embodiment of the present invention;
fig. 3 is a schematic connection diagram of the respective ELM networks in an input-output iteration manner.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, an automatic control method for a multiple-input multiple-output system based on an extreme learning machine includes the following steps:
1) and input and output of the controlled object are defined. Determining control targets for a device and associating the device with each control targetControlling a variable; recording the number of control targets as m and the number of control variables as n; the ith control variable is uiI is 1,2, …, n, and the jth control target is vj，j＝1,2,…,m；
2) According to the number of the control variables, for each control variable uiEstablishing an extreme learning machine network in a controller: ELMi；
The extreme learning machine network comprises (m + n-1) input layer nodes, a plurality of hidden layer nodes and 1 output layer node; to limit learning machine network ELMiInput layer node is vj，j＝1,2,…,m；utT ≠ 1,2, …, n, and t ≠ i; output layer node is ui；
3) A control loop is started. The training samples are collected into a plurality of groups of the latest data samples, and are sorted according to the input and output structure of each ELM network, so that each ELM network can obtain a sufficient amount of training sample data;
4) training the corresponding network by using the training sample data of each trimmed ELM network according to an ELM algorithm to obtain each trained ELM network;
5) based on each ELM network which is completely trained in the step 4), the current value v of the control target of the equipment is obtainedj *And the current value u of the control variablei *Sending the data into corresponding nodes of each ELM network input layer, and obtaining single prediction results of corresponding control variables, namely the control variable u, from nodes of each network output layeriCorresponding network ELMiWhen the training is complete, it will
Figure BDA0002538401550000061
Inbound network ELMiInputting layer nodes, and calculating network output to obtain control variable uiSingle prediction of result ui-p1；
6) Each ELM network is used for performing iteration for mutual input and output; the method comprises the following specific steps:
subjecting each ELM in step 5)iOutput value u of network output layer node i-p1 respectively sending the signals to nodes corresponding to the control variables of other ELM network input layers;
then, each ELM network recalculates the network output based on the latest input to obtain the next generation of prediction output, and further completes one iteration;
7) continuously repeating iteration of each ELM network according to the mode of the step 6) until the preset iteration number N is reached, stopping iteration, and recording the ELM of each current networkiRespectively of ui-pN；
8) Carrying out N iterations on each ELM network in the step 7) to obtain a final output value u1-pN,...,ui-pN,...,un-pNThe control variable value is used as an effective output value of the controller and directly acts on the corresponding position of the controlled equipment; completing automatic control of the control object equipment according to the control target;
9) and (3) acquiring the latest input and output data of the controlled equipment to update the training sample set, ending the current control cycle, and jumping to the step 3) to start the next control cycle.
For automatic power control of a photovoltaic system of a large ship, the photovoltaic system is designed to contain 10 local photovoltaic arrays which can independently control power and are marked as PV1,PV2,…,PV10. As a part of the comprehensive energy system of the ship, the purpose of the automatic power control of the photovoltaic system is as follows: receiving a control instruction from a scheduling layer of the integrated energy system, namely the total output power P of the photovoltaic system under the current environmental parameters (illumination intensity and battery temperature)tTarget value P oft *The controller can immediately calculate the PV given the current environmental parameters1,PV2,…,PV10Respective operating voltage ui(i-1, 2, …,10) is a reasonable value. Subsequently, PV is put in1,PV2,…,PV10The working point is directly driven to the value position, so that the total output power P of the ship photovoltaic systemtCan quickly approach to the scheduling target P under the environmental parameter at one timet *。
The current environmental parameters can be measured by a sensor, and the total output power target value of the photovoltaic system comes from a control instruction of a comprehensive energy system scheduling layer. The automatic power control process of the photovoltaic system may include the steps of:
step S1: and input and output of the controlled object are defined. The control target of the automatic power control of the photovoltaic system is defined as follows: total output power PtAnd the illumination intensity and cell temperature S of each local photovoltaic arrayi、Ti(i ═ 1,2, …, 10); the control variables are defined as: operating voltage u of each local photovoltaic arrayi(i ═ 1,2, …, 10); the purpose of the control is as follows: at Si、TiUnder the environment described by the current value, the total output power P of the photovoltaic system is enabledtAchieving a particular scheduling objective Pt *。
Step S2: 10 extreme learning machine networks were built in the controller: ELMi(i ═ 1,2, …, 10). Each network contains 30 input layer nodes, several hidden layer nodes and 1 output layer node. The corresponding relationship between the input layer and the output layer nodes of each ELM network is shown in Table 1:
table 1 correspondence between nodes of input layer and output layer of each ELM network
Figure BDA0002538401550000081
Figure BDA0002538401550000091
Step S3: a control loop is started. And arranging a plurality of groups of latest data in the training sample set according to the input and output structures of each ELM network. Specifically, let the latest K sets of sample data D in the training sample setk＝[uk1,uk2,...,uk10|Pk,Sk1,Tk1,...,Sk10,Tk10](K1, 2.. K), then for the kth set of sample data DkThe useful form of the network is shown in table 2:
TABLE 2 sample data DkUse form for each ELM network
Figure BDA0002538401550000092
Step S4: training the corresponding network by using the training sample data of each ELM network obtained in the step S3 according to the ELM algorithm rule;
step S5: based on each ELM network trained completely in step S4, scheduling target value P of total output power of the current photovoltaic systemt *Current value of environmental parameter
Figure BDA0002538401550000101
And the current value of the working voltage of each local photovoltaic array
Figure BDA0002538401550000102
Sending the control variables into corresponding nodes of each ELM network input layer, and obtaining single prediction results of the corresponding control variables from the nodes of each network output layer: u. of1-p1，u2-p1，…，u10-p1；
Step S6: and each ELM network is used for performing iteration for mutual input and output. Specifically, the output value (i.e., u) of each ELM network output layer node in step S5 is determined1-p1，u2-p1，…，u10-p1) Respectively sending the signals to other nodes corresponding to the control variable of the input layer of the ELM network, namely: will u1-p1Are respectively sent into a network ELM2、…、ELM10Corresponding input layer nodes of (1); will u2-p1Are respectively sent into a network ELM1、ELM3、…、ELM10Corresponding input layer nodes of (1); will u3-p1Are respectively sent into a network ELM1、ELM2、ELM4、…、ELM10Corresponding input layer nodes of (1); will u4-p1Are respectively sent into a network ELM1、ELM2、ELM3、ELM5、…、ELM10Corresponding input layer nodes of (1); will u5-p1Are respectively sent into a network ELM1、ELM2、ELM3、ELM4、ELM6、…、ELM10Corresponding input layer nodes of (1); will u6-p1Are respectively sent into a network ELM1、…、ELM5、ELM7、…、ELM10Corresponding input layer nodes of (1); will be provided withu7-p1Are respectively sent into a network ELM1、…、ELM6、ELM8、ELM9、ELM10Corresponding input layer nodes of (1); will u8-p1Are respectively sent into a network ELM1、…、ELM7、ELM9、ELM10Corresponding input layer nodes of (1); will u9-p1Are respectively sent into a network ELM1、…、ELM8、ELM10Corresponding input layer nodes of (1); will u10-p1Are respectively sent into a network ELM1、…、ELM9Corresponding input layer node of (2).
Then, each ELM network recalculates the network output based on the latest input to obtain the next generation predicted output u1-p2，u2-p2，…，u10-p2And then one iteration is finished;
step S7: each ELM network continuously iterates in the manner described in step S6 until a stop condition is reached: and the percentage difference value of the adjacent two iteration outputs of each network is lower than a preset threshold eta or reaches a preset iteration number N. Stopping iteration, recording the current ELM of each networkiOutput layer node output values of (i ═ 1,2, …,10) are each u1-pN,u2-pN,...,u10-pN；
Step S8: the final output value u of each ELM network in the step S7 after N iterations1-pN,u2-pN,...,u10-pNAs the effective output of the controller under the current environmental parameters and directly acting on the controlled object, i.e. driving the working voltage of each local photovoltaic array to the controller output u1-pN,u2-pN,...,u10-pN. Measuring the total actual output power P of the photovoltaic system under the current illumination intensity S and the battery temperature TtAnd with its scheduling target value Pt *Calculating deviation, finely adjusting the working voltage of each local photovoltaic array according to a certain priority, and finally stabilizing the actual output power of the photovoltaic system at a scheduling target value P under the environmental parametert *；
Step S9: collecting the current working voltage u of each local photovoltaic array1-pN,...,ui-pN,...,un-pNCurrent environmental parameters
Figure BDA0002538401550000111
Total actual output power P of current photovoltaic systemtAnd storing the data in the latest position in the training sample set. The current control cycle is ended, and the process proceeds to step S3 to start the next control cycle.
The method is also suitable for a multi-input single-output system (namely a single control target), can be used as a special case of multi-input (namely a plurality of control variables) and multi-output (namely a plurality of control targets), and can be used for reducing corresponding nodes on the input layer of each control variable ELM network.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.

Claims (4)

1. An automatic control method of a multi-input multi-output system based on an extreme learning machine is characterized by comprising the following steps:
1) analyzing the equipment to be controlled, and determining a control target of the equipment and control variables associated with the equipment and each control target; recording the number of control targets as m and the number of control variables as n; the ith control variable is uiI is 1,2, …, n, and the jth control target is vj，j＝1,2,…,m；
The control object device is a multi-autonomous system used in industrial production and application equipment;
2) according to the number of the control variables, for each control variable uiEstablishing an extreme learning machine network in a controller: ELMi；
The extreme learning machine network comprises (m + n-1) input layer nodes, a plurality of hidden layer nodes and 1 output layer node; to limit learning machine network ELMiInput layer node is vj，j＝1,2,…,m；utT ≠ 1,2, …, n, and t ≠ i; output layer node is ui；
3) Starting a control loop, and training samples according to the network structure of the extreme learning machineThe existing latest K groups of sample data are concentrated for adjustment, the training sample is concentrated with the existing latest K groups of sample data, and the kth group of sample data DkArranging the input layer nodes and the output layer nodes of the extreme learning machine network into the available form of each ELM network;
4) training the corresponding network by using the training sample data of each trimmed ELM network according to an ELM algorithm to obtain each trained ELM network;
5) based on each ELM network which is completely trained in the step 4), the current value v of the control target of the equipment is obtainedj *And the current value u of the control variablei *Sending the data into corresponding nodes of each ELM network input layer, and obtaining single prediction results of corresponding control variables, namely the control variable u, from nodes of each network output layeriCorresponding network ELMiWhen the training is complete, it will
Figure FDA0002538401540000021
Inbound network ELMiInputting layer nodes, and calculating network output to obtain control variable uiThe single prediction result of (1);
6) each ELM network is used for performing iteration for mutual input and output; the method comprises the following specific steps:
subjecting each ELM in step 5)iThe predicted result output values of the network output layer nodes are respectively sent to the nodes of other ELM network input layers corresponding to the control variables;
then, each ELM network recalculates the network output based on the latest input to obtain the next generation of prediction output, and further completes one iteration;
7) continuously repeating iteration of each ELM network according to the mode of the step 6) until the preset iteration number N is reached, stopping iteration, and recording the ELM of each current networkiRespectively of ui-pN；
8) Carrying out N iterations on each ELM network in the step 7) to obtain a final output value u1-pN,...,ui-pN,...,un-pNAs effective output value of the controller, and directly acting the control variable value thereof on the controlled devicePreparing the corresponding position; completing automatic control of the control object equipment according to the control target;
9) and (3) acquiring the latest input and output data of the controlled equipment to update the training sample set, ending the current control cycle, and jumping to the step 3) to start the next control cycle.
2. The method according to claim 1, wherein the current value of the control variable in the first iteration of step 5) is the current actual value of each control variable, i.e. the effective value output by the controller in the previous control loop, and the current value of the control variable in each subsequent iteration is the predicted output value of the previous iteration of the output layer node of the ELM network.
3. The automatic control method based on the extreme learning machine as claimed in claim 1, wherein the network training in step 4) is completed in one time without an iterative process, and the prediction of the control variables in steps 5) to 8) is completed multiple times, which requires the ELM networks corresponding to the control variables to perform mutual input and output sufficient iteration.
4. The extreme learning machine-based automatic control method as claimed in claim 1, wherein the multi-autonomous system in step 1) is a system with multiple inputs and multiple outputs, strong coupling and time-varying property.
CN202010539521.XA 2020-06-14 2020-06-14 Automatic control method of multi-input multi-output system based on extreme learning machine Active CN111796519B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010539521.XA CN111796519B (en) 2020-06-14 2020-06-14 Automatic control method of multi-input multi-output system based on extreme learning machine

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010539521.XA CN111796519B (en) 2020-06-14 2020-06-14 Automatic control method of multi-input multi-output system based on extreme learning machine

Publications (2)

Publication Number Publication Date
CN111796519A true CN111796519A (en) 2020-10-20
CN111796519B CN111796519B (en) 2022-05-06

Family

ID=72803290

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010539521.XA Active CN111796519B (en) 2020-06-14 2020-06-14 Automatic control method of multi-input multi-output system based on extreme learning machine

Country Status (1)

Country Link
CN (1) CN111796519B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114895190A (en) * 2022-05-18 2022-08-12 武汉理工大学 Charge estimation method and device based on extreme learning and extended Kalman filtering

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103400190A (en) * 2013-08-13 2013-11-20 浙江大学 Integrated framework method for optimizing extremity learning machine by using genetic algorithm
CN104918030A (en) * 2015-06-05 2015-09-16 河海大学 Color space conversion method based on ELM extreme learning machine
CN107102727A (en) * 2017-03-17 2017-08-29 武汉理工大学 Dynamic gesture study and recognition methods based on ELM neutral nets
CN107168292A (en) * 2017-06-21 2017-09-15 武汉理工大学 Submarine navigation device circuit failure diagnosis method based on ELM algorithms
CN108021947A (en) * 2017-12-25 2018-05-11 北京航空航天大学 A kind of layering extreme learning machine target identification method of view-based access control model
CN108876054A (en) * 2018-07-06 2018-11-23 国网河南省电力公司郑州供电公司 Short-Term Load Forecasting Method based on improved adaptive GA-IAGA optimization extreme learning machine
CN108983849A (en) * 2018-07-12 2018-12-11 沈阳大学 It is a kind of to utilize compound extreme learning machine ANN Control greenhouse method
CN109034278A (en) * 2018-07-19 2018-12-18 工极（北京）智能科技有限公司 A kind of ELM-IN-ELM frame Ensemble Learning Algorithms based on extreme learning machine
CN109948640A (en) * 2018-12-26 2019-06-28 杭州电子科技大学 Electromyographic signal classification method based on two-parameter core Optimization-type extreme learning machine
CN110322077A (en) * 2019-07-10 2019-10-11 燕山大学 Cement raw material Vertical Mill raw material fineness index prediction technique based on convolutional neural networks

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103400190A (en) * 2013-08-13 2013-11-20 浙江大学 Integrated framework method for optimizing extremity learning machine by using genetic algorithm
CN104918030A (en) * 2015-06-05 2015-09-16 河海大学 Color space conversion method based on ELM extreme learning machine
CN107102727A (en) * 2017-03-17 2017-08-29 武汉理工大学 Dynamic gesture study and recognition methods based on ELM neutral nets
CN107168292A (en) * 2017-06-21 2017-09-15 武汉理工大学 Submarine navigation device circuit failure diagnosis method based on ELM algorithms
CN108021947A (en) * 2017-12-25 2018-05-11 北京航空航天大学 A kind of layering extreme learning machine target identification method of view-based access control model
CN108876054A (en) * 2018-07-06 2018-11-23 国网河南省电力公司郑州供电公司 Short-Term Load Forecasting Method based on improved adaptive GA-IAGA optimization extreme learning machine
CN108983849A (en) * 2018-07-12 2018-12-11 沈阳大学 It is a kind of to utilize compound extreme learning machine ANN Control greenhouse method
CN109034278A (en) * 2018-07-19 2018-12-18 工极（北京）智能科技有限公司 A kind of ELM-IN-ELM frame Ensemble Learning Algorithms based on extreme learning machine
CN109948640A (en) * 2018-12-26 2019-06-28 杭州电子科技大学 Electromyographic signal classification method based on two-parameter core Optimization-type extreme learning machine
CN110322077A (en) * 2019-07-10 2019-10-11 燕山大学 Cement raw material Vertical Mill raw material fineness index prediction technique based on convolutional neural networks

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CHENGBO LU等: "An improved weighted extreme learning machine for imbalanced data classification", 《MEMETIC COMPUTING》 *
李军等: "基于ELM的一类MIMO仿射非线性系统的鲁棒自适应控制", 《控制与决策》 *
罗仲梦: "ELM在机器人建模与控制中的应用研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114895190A (en) * 2022-05-18 2022-08-12 武汉理工大学 Charge estimation method and device based on extreme learning and extended Kalman filtering
CN114895190B (en) * 2022-05-18 2024-04-09 武汉理工大学 Method and equipment for estimating charge quantity based on extreme learning and extended Kalman filtering

Also Published As

Publication number Publication date
CN111796519B (en) 2022-05-06

Similar Documents

Publication Publication Date Title
CN110472778A (en) A kind of short-term load forecasting method based on Blending integrated study
CN109190670B (en) Charging pile fault prediction method based on extensible hoisting tree
CN109270842B (en) Bayesian network-based regional heat supply model prediction control system and method
CN107220734A (en) CNC Lathe Turning process Energy Consumption Prediction System based on decision tree
CN104217258B (en) A kind of electric load sigma-t Forecasting Methodology
CN111612125B (en) Novel HTM time pool method and system for online learning
CN110287540B (en) Photovoltaic cell parameter identification method based on elite crowd differential evolution algorithm
CN113762603B (en) Photovoltaic base station short-term photovoltaic power prediction method based on improved sparrow algorithm optimization
CN107728478A (en) Fuel cell excess oxygen coefficient network response surface method
CN113125960A (en) Vehicle-mounted lithium ion battery charge state prediction method based on random forest model
CN109325613B (en) Method for online predicting static voltage stability margin by using parallel sequential extreme learning machine
CN104656620A (en) Comprehensive evaluation system for remanufacturing of heavy-duty machine tool
CN108804800A (en) Lithium ion battery SOC on-line prediction methods based on echo state network
CN111796519B (en) Automatic control method of multi-input multi-output system based on extreme learning machine
CN115409263A (en) Method for predicting remaining life of lithium battery based on gating and attention mechanism
CN115456460A (en) Multi-quality index output prediction method and system for tobacco shred loosening and conditioning process
CN112649736A (en) SOC and SOH joint estimation method for retired battery
CN111459022A (en) Device parameter adjustment method, device control apparatus, and computer-readable storage medium
CN111537888A (en) Data-driven echelon battery SOC prediction method
CN115764870A (en) Multivariable photovoltaic power generation power prediction method and device based on automatic machine learning
CN117093033A (en) Resistance heating furnace temperature control system for optimizing PID parameters based on particle swarm optimization
CN105955921A (en) Robot hierarchical reinforcement learning initialization method based on automatic discovery of abstract action
CN110232432B (en) Lithium battery pack SOC prediction method based on artificial life model
CN108256688B (en) Steam flooding development effect prediction method for optimizing GRNN by improving firefly algorithm
CN113328435A (en) Active and reactive power combined control method for active power distribution network based on reinforcement learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant