WO1999048017A2 - Verfahren und anordnung zur rechnergestützten ermittlung einer zugehörigkeitsfunktion - Google Patents
Verfahren und anordnung zur rechnergestützten ermittlung einer zugehörigkeitsfunktion Download PDFInfo
- Publication number
- WO1999048017A2 WO1999048017A2 PCT/DE1999/000525 DE9900525W WO9948017A2 WO 1999048017 A2 WO1999048017 A2 WO 1999048017A2 DE 9900525 W DE9900525 W DE 9900525W WO 9948017 A2 WO9948017 A2 WO 9948017A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- cluster center
- cluster
- center
- membership function
- determined
- 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.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B37/00—Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
Definitions
- the invention relates to the determination of a membership function.
- a membership function is used in the context of mapping an input variable to an output variable using a neural network or a fuzzy system.
- a radial basis function network is a special forward-looking neural network which describes a mapping by which an input variable is mapped to an output variable and which has only one layer of hidden neurons.
- a hidden neuron has a special, radial-symmetrical (radial) membership function (activation function).
- a membership function or activation function is a basic function of a subsystem (submodel) of a system of functions in a room (overall model).
- the overall model is used for a mapping which maps an input variable to an output variable in the room.
- a membership function or activation function describes a membership of a data point in the room to the corresponding partial model.
- the sub-model has a so-called center, which is a
- Membership function provides an absolute maximum. With increasing distance of a selected data point from the
- Center decreases the membership function value for the selected data point.
- R - »R be a function from R to R given by N interpolation points.
- the function system used for an approximation consists of radially symmetrical basic functions hj_, which are each assigned to a support point as the respective center. They are defined as follows:
- the basic function h has positive scalar values. It is only dependent on the distance of a vector X from the corresponding support point Xj_, which is assigned to a neuron i is, in any given standard. Usually a Euclidean distance between the vectors X and Xj_ in R is considered
- the radial basis function hj_ has different characteristics, such as, for example, a width or a variance of a Gaussian function, for an assigned center i.
- a so-called fuzzy clustering method for data analysis is known from [2] and [3].
- c clusters and corresponding affiliations of data vectors x ⁇ are determined in such a way that data vectors that are close to a cluster in a data space have the highest possible affiliation and data vectors x ⁇ that are far away from the cluster have the lowest possible affiliation to the respective cluster. This is done by minimizing a sum of the
- the clusters are described by a membership matrix U, which has c rows and n columns.
- Each element uj_ k of the membership matrix U has a value within the interval [0, 1] and describes a membership of the data vector X k to the i-th cluster.
- a cluster must contain at least one element, so that:
- the cost function J m of the membership values is formed according to the following rule:
- A denotes a predeterminable induced norm of the inner product according to regulation (4), which is usually given by the identity matrix (Euclidean distance).
- the cost function J m is minimized using a so-called Picard iteration.
- Membership values uj_ k and cluster centers y_i are formed one after the other according to the following regulations:
- the determination of the membership values uj_ and the cluster centers y_j_ is repeated until a specified number of iterations has been carried out or until a change in the membership values u and / or until a change in the cluster centers y_i is below a predetermined threshold value.
- fuzzy C means clustering
- the clusters are described by their cluster centers y_i. From [4] is a so-called Voronoi tiling
- a Voronoi cell is delimited by a convex polygon that defines an assigned cluster center ⁇ . encloses.
- An edge kj_ ⁇ of the polygon can be defined as a perpendicular to a neighboring cluster center ⁇ .
- n ⁇ . - ⁇ .
- equation (15) x data point of the edge kj_j, which limits the i-th Voronoi cell with respect to the neighboring j-th Voronoi cell.
- the edge k j _ j introduced for the two-dimensional entrance space is a plane.
- the convex polygon is a convex
- edge is replaced by the term hyper-plane and the term polygon by the term hyper-polyhedron.
- mapping that describes a system behavior of a technical system or process and uses a known membership function that determines the mapping has the disadvantage in many cases that the mapping does not realistically describe the system behavior in many cases.
- the invention is based on the problem of specifying a method and an arrangement with which a membership function can be determined, an image ' which maps an input variable to an output variable using the membership function being improved in such a way that a process which is carried out by the Figure is described, is described as realistically as possible.
- a method for the computer-aided determination of a membership function for a data point to a selected clusters in a given space, which membership function describes the membership of the data point to the selected cluster has the following steps: a) clustering is carried out using predetermined data points; b) at least three clusters are determined; c) an associated cluster center is determined for each cluster; d) the membership function is dependent on
- Distances from the selected cluster center to the cluster centers closest to the selected cluster center are determined.
- An arrangement for the computer-aided determination of a membership function for a data point to a selected cluster in a predetermined space, which membership function describes the membership of the data point to the selected cluster has a processor with which the following method steps can be carried out: a) it is clustering feasible using predetermined data points; b) at least three clusters can be determined; c) an associated cluster center can be determined for each cluster; d) the membership function can be determined as a function of distances from the selected cluster center to the cluster centers closest to the selected cluster center.
- the distance is preferably taken into account in such a way that an increase in the membership function increases with an increasing distance between the selected cluster center and a closest neighboring cluster center.
- the membership function is preferably used as an activation function.
- An improved description of a technical process by means of a neural network or a fuzzy system can be achieved in that the membership function has at least one first subfunction and at least one second subfunction.
- the membership function is preferably standardized.
- the cluster centers can preferably be determined using a minimal Euclidean norm.
- space is a multidimensional space.
- the data points are measured, the data points being working points of a technical system.
- the technical system is preferably a steel rolling mill.
- Cluster center determined with respect to a cluster center next to the cluster center.
- the degree of overlap ⁇ is preferably determined taking into account the distance of the selected cluster center to a cluster center which is closest to the cluster center.
- the degree of overlap ⁇ is formed for a two-dimensional space according to the following rule:
- Cluster center ⁇ i + ⁇ i limit of an overlap area for the i + lth cluster center with respect to the i-th
- the degree of overlap ⁇ is formed for a multidimensional space according to the following rule:
- the cluster is preferably a Voronoi cell.
- the second subfunction is formed according to the following rule:
- Cluster center with respect to the i + lth cluster center ⁇ i + i limit of an overlap area for the i + lth
- Further training / further training is / is preferably used for determining a basic function for a neural network and / or for determining an activation function for a fuzzy system.
- Training used in online learning of a neural network.
- a configuration for monitoring a steel rolling mill is preferably used.
- another Design for a control of a steel rolling mill can be used.
- Figure 1 Schematic representation of components of a steel rolling mill monitored and controlled by a neural network or a fuzzy system
- Network or fuzzy system Figure 3 Representation of an image by a neural network
- Figure 4 Representation of an image by a fuzzy system
- Figure 5 Representation of a clustered by Voronoi cells
- FIG. 1 components of a steel rolling system for steel processing monitored and controlled using a neural network are shown schematically.
- Figure 1 shows a system of the steel rolling mill 101 for steel processing (rolling process).
- the system of the steel rolling mill 101 is monitored and controlled using a neural network 102.
- a system behavior of the steel rolling mill 101 is described by the neural network 102, which is of the radial basic function network type.
- a fuzzy system can also be used.
- Suitable measuring means 103 for example sensors 103, are also shown, with which process variables which influence the process of steel processing are measured.
- the measuring means 103 are connected to a memory 105 via a bus 104.
- the process variables are measured at predefinable times and stored in the memory 105.
- the measured process variables are chemical process variables, such as a carbon concentration or a manganese concentration, an end strip temperature of a rolled steel strip, an end strip thickness of the rolled steel strip and a roller speed.
- a further measuring means 106 is shown with which a further process variable, a rolling force, which is a variable to be monitored and a control variable of the process or the system of the steel rolling mill 101, is measured.
- the further measuring means 106 is also connected to the memory 105 via the bus 104.
- the further process variable is also measured on the steel rolling mill system 101 at the predeterminable times and stored in the memory 105.
- the memory 105 is connected via a further bus 115 to a processing unit 107 which has a processor 108, for example a computer.
- the neural network 102 or the fuzzy system 110 is stored in the form of software in the processing unit 107.
- Processor 108 executes the software.
- the rolling force is an output variable of the neural network
- Imaging behavior is determined.
- Process variables selected process variables that significantly influence the process of steel processing.
- the influence of a process variable on the process is determined by a sensitivity analysis.
- the following input variables are selected:
- the input variables are applied to the neural network 102 or to the fuzzy system 110.
- the neural network 102 or the fuzzy system 110 determines the output quantity rolling force using the input quantities.
- the output variable is tapped at the neural network 102 or at the fuzzy system 110 by means 108 and to the system via a data line 109 which connects the neural network 102 or the fuzzy system 110 to the system of the steel rolling mill 101 transferred to the steel rolling mill 101.
- the system steel rolling mill 101 is monitored and controlled using suitable means 111, whereby a suitable one Setting the rolling force a predetermined final strip thickness of the
- the rolling force to be set is determined using the neural network 102 or the fuzzy system 110 for a predetermined end strip thickness of the steel strip.
- FIG. 2 shows method steps that are carried out as part of a training of the neural network, with which training a predefinable mapping behavior is determined.
- training data are determined in such a way that the selected process variables 201 and the rolling force 202 are measured on the steel rolling mill at predeterminable times, in each case summarized as a training data vector 203 and stored 204 in the memory.
- the training data vectors are applied 205 to the neural network.
- a so-called rolling force correction factor determined using the rolling force output variable is determined 206 as a target variable for training the neural network.
- the neural network is trained 207 using a known training method as described in [1].
- the training of the neural network is preferably carried out in the operation of the steel rolling mill, the training data in the
- the online training has the advantage that a change in the system behavior of the steel rolling mill is taken into account in the imaging behavior of the neural network without a great time delay. Such a change can be caused, for example, by wear of a component of the steel rolling mill.
- a mapping behavior of a neural network or a fuzzy system such as is used for monitoring and controlling a system of a technical system, for example the system of the steel rolling mill described above, is described in more detail below.
- Figure 3 shows a simplified representation of an image by a neural network.
- a one-dimensional input space (x) is used as an input space of the neural network with a one-dimensional input variable x.
- a dimension of the entrance space is determined by a number of input variables.
- the neural network for monitoring and controlling the steel rolling mill described above has, for example, a multi-dimensional entrance space.
- the output variable of the neural network is a one-dimensional output variable y.
- the neural network which describes a system behavior of a technical system, thus maps the input variable x to the output variable y in a state space (x, y) 301.
- Figure 3 shows the state space (x, y) 301, in which the
- the neural network which is of the type radial basic function network as described in [1], can be represented as follows:
- the weighted local partial model yi (x) 303 is a product of an i-th standardized membership function ⁇ (x) and an i-th
- the i-th local sub- model C ⁇ is a constant weighting factor which is assigned to the i-th center 304.
- Weighting factor for the i-th center 304 g non-standardized membership function (hereinafter abbreviated as membership function)
- a Gaussian function of the form becomes the membership function gi of the i-th local partial model
- Partial model is determined by the variance ⁇ i.
- the system behavior of the process can be described by the neural network.
- Center ⁇ i 304 and the next neighboring center ⁇ i + i 306 are assigned.
- ⁇ + ⁇ a *
- FIG. 4 shows an illustration of a neural network that describes the system behavior of the process in a simplified manner.
- FIG. 4 shows the simplified state space (x, y) 401, in which the system behavior of the process is described by so-called working points 402 of the process, which are shown in the form of a first function 412.
- the fuzzy system which maps an input variable x to an output variable y, can be described as follows:
- N i l
- ⁇ i (x) is a membership function 403 normalized to 1
- Center ⁇ i 404 is determined by a clustering method as described in [2] and [3].
- the first area 407 is in each case the center ⁇ i 404 of the i-th partial model.
- the first area 407 comprises the following interval:
- the second 410 and the third 411 area of the i-th center are arranged in such a way that the second 410 and the third 411 area each directly adjoin the first area 407 of the center ⁇ 404.
- the second region 410 is between the i-lth center 413 and the center ⁇ i 404 as well as the
- ⁇ ( ⁇ i, ⁇ + ⁇ ) degree of overlap of the ith cluster center 404 with respect to the i + lth cluster center 412 (for the third area 411).
- ⁇ i (x) gi, i-i for x e [ ⁇ i_ ⁇ , i; ⁇ i, i_ ⁇ ]
- 02 ) ⁇ i, i + ⁇ ⁇ i + a *
- FIGS. 5, 6 and 7 a, b describe a procedure for adapting the simplified fuzzy system shown above for a fuzzy system that has a two-dimensional input variable.
- the fuzzy system has a three-dimensional state space (x, y) which comprises a two-dimensional input space (x).
- the fuzzy system is represented as follows:
- FIG. 5 the two-dimensional entrance space () 501 with the cluster centers is ⁇ . 502 shown. Furthermore, FIG. 5 shows ⁇ delimited by polygons 504 and in each case one cluster center. 502 assigned effective areas 503.
- the effective areas 503, which are also referred to as areas of influence or Voronoi cells 503, are determined by a so-called Voronoi tiling, as described in [4].
- the Voronoi cell 503 of a cluster center ⁇ . 502 is through
- dminfe ⁇ i minU (x - ⁇ ⁇ * (x - ⁇ j, xe R n (14) with: d m i n (x, ⁇ .) minimum Euclidean distance one
- a Voronoi cell 503 is delimited by a convex polygon 504, which has a center ⁇ assigned to the Voronoi cell 503. 502 encloses. Edges kij 505 of the polygon 504 can be considered
- Cluster center ⁇ . 506 determined.
- FIG. 6 shows a first 601 and a second 602 area corresponding to the simplified fuzzy system for an i-th cluster center ⁇ . 604 regarding a jth
- Limits of a first 601 or second 602 area are linear polygons 607 which are parallel to an edge kij 605 of the Voronoi cell 606.
- the polygons 607 are determined in such a way that all points of a boundary line
- a form factor a has the value 0.5, for example.
- a degree of overlap ⁇ ( ⁇ ., ⁇ . J is determined.
- a normalized membership function ⁇ (x) of the i-th Voronoi cell 606 is corresponding to the normalized
- Affiliation function of an i-th sub-model, as described in the simplified fuzzy system described above, is determined.
- a first 701 special area is shown in FIG. 7a) and a second 702 special area is shown in FIG. 7b), for which a standardized membership function ⁇ sl (x) and ⁇ s2 ( ⁇ ) w i- e are determined as follows:
- Voronoi cell 705 and 704 respectively
- the clustering process can be carried out using a K-means process (KM) or a neural gas process (NG).
- KM K-means process
- NG neural gas process
- any partial function ci (x) for example a linear function, can also be used in the local partial model.
- ci (x) for example a linear function
- Embodiment is used for monitoring and control of a steel rolling mill, also for the
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
- Feedback Control In General (AREA)
- Complex Calculations (AREA)
- Control Of Metal Rolling (AREA)
- Stored Programmes (AREA)
Abstract
Description
Claims
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP99916761A EP1070291A2 (de) | 1998-03-18 | 1999-02-26 | Verfahren und anordnung zur rechnergestützten ermittlung einer zugehörigkeitsfunktion |
| JP2000537149A JP2002507793A (ja) | 1998-03-18 | 1999-02-26 | メンバーシップ関数を計算機支援により検出する方法及び装置 |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE19811906 | 1998-03-18 | ||
| DE19811906.2 | 1998-03-18 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| WO1999048017A2 true WO1999048017A2 (de) | 1999-09-23 |
| WO1999048017A3 WO1999048017A3 (de) | 1999-11-04 |
Family
ID=7861418
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/DE1999/000525 Ceased WO1999048017A2 (de) | 1998-03-18 | 1999-02-26 | Verfahren und anordnung zur rechnergestützten ermittlung einer zugehörigkeitsfunktion |
| PCT/DE1999/000524 Ceased WO1999048020A2 (de) | 1998-03-18 | 1999-02-26 | Verfahren und anordnung zur rechnergestützten ermittlung einer abbildungsvorschrift |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/DE1999/000524 Ceased WO1999048020A2 (de) | 1998-03-18 | 1999-02-26 | Verfahren und anordnung zur rechnergestützten ermittlung einer abbildungsvorschrift |
Country Status (4)
| Country | Link |
|---|---|
| EP (2) | EP1071999B1 (de) |
| JP (1) | JP2002507793A (de) |
| DE (1) | DE59901352D1 (de) |
| WO (2) | WO1999048017A2 (de) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10013509C1 (de) * | 2000-03-20 | 2001-10-18 | Harro Kiendl | Verfahren zur datenbasierten Konstruktion eines Moduls mit den Daten entsprechendem Übertragungsverhalten und Modul hierfür |
| US7962441B2 (en) | 2006-09-22 | 2011-06-14 | Denso Corporation | Air conditioner for vehicle and controlling method thereof |
| JP4360409B2 (ja) * | 2007-02-13 | 2009-11-11 | 株式会社デンソー | 車両用空調装置、車両用空調装置の制御方法および制御装置 |
| JP4990115B2 (ja) | 2007-12-06 | 2012-08-01 | 株式会社デンソー | 位置範囲設定装置、移動物体搭載装置の制御方法および制御装置、ならびに車両用空調装置の制御方法および制御装置 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW297107B (de) * | 1991-06-18 | 1997-02-01 | Meidensha Electric Mfg Co Ltd | |
| DE4416364B4 (de) * | 1993-05-17 | 2004-10-28 | Siemens Ag | Verfahren und Regeleinrichtung zur Regelung eines Prozesses |
-
1999
- 1999-02-26 EP EP99916760A patent/EP1071999B1/de not_active Expired - Lifetime
- 1999-02-26 EP EP99916761A patent/EP1070291A2/de not_active Withdrawn
- 1999-02-26 JP JP2000537149A patent/JP2002507793A/ja not_active Withdrawn
- 1999-02-26 WO PCT/DE1999/000525 patent/WO1999048017A2/de not_active Ceased
- 1999-02-26 WO PCT/DE1999/000524 patent/WO1999048020A2/de not_active Ceased
- 1999-02-26 DE DE59901352T patent/DE59901352D1/de not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| WO1999048020A3 (de) | 1999-11-04 |
| EP1070291A2 (de) | 2001-01-24 |
| JP2002507793A (ja) | 2002-03-12 |
| WO1999048017A3 (de) | 1999-11-04 |
| EP1071999A2 (de) | 2001-01-31 |
| DE59901352D1 (de) | 2002-06-06 |
| EP1071999B1 (de) | 2002-05-02 |
| WO1999048020A2 (de) | 1999-09-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| DE69324052T2 (de) | Neuronalnetzwerk-Lernsystem | |
| DE112023000011T5 (de) | Ein primäres Frequenzmodulationssteuersystem der thermischen Stromerzeugungseinheit mit kleinen Störungen | |
| EP2108139B1 (de) | Verfahren zur rechnergestützten regelung und/oder steuerung eines technischen systems, insbesondere einer gasturbine | |
| EP1626317B1 (de) | Verfahren zur Erstellung eines nichtlinearen, stationären oder dynamischen Modells einer Steuergrösse einer Arbeitsmaschine | |
| DE102018001028B4 (de) | Numerische Steuerung | |
| EP3785177A1 (de) | Verfahren und vorrichtung zum ermitteln einer netzkonfiguration eines neurona-len netzes | |
| DE102018221002A1 (de) | Steuereinrichtung zur Steuerung einer Fertigungsanlage sowie Fertigungsanlage und Verfahren | |
| DE102017220954A1 (de) | Verfahren, Vorrichtung und Computerprogramm zur Ermittlung einer Anomalie | |
| EP0997570B1 (de) | Verfahren und Anordnung zur rechnergestützten Ermittlung von Clustern zur Erkennung von Schaumbildung in einer Waschmaschine sowie Verfahren und Anordnung zur Erkennung von Schaumbildung in einer Waschmaschine | |
| WO2000008599A2 (de) | Anordnung miteinander verbundener rechenelemente, verfahren zur rechnergestützten ermittlung einer dynamik, die einem dynamischen prozess zugrunde liegt und verfahren zum rechnergestützten trainieren einer anordnung miteinander verbundener rechenelemente | |
| EP1055180B1 (de) | Verfahren und anordnung zum entwurf eines technischen systems | |
| EP0897155A1 (de) | Verfahren zur Steuerung von Prozessvorgängen | |
| DE102020213238A1 (de) | Erzeugung von vereinfachten computer-implementierten neuronalen netzwerken | |
| WO1999048017A2 (de) | Verfahren und anordnung zur rechnergestützten ermittlung einer zugehörigkeitsfunktion | |
| DE69802372T2 (de) | Klassifizierungssystem und -verfahren mit N-Tuple- oder RAM-basiertem neuronalem Netzwerk | |
| DE102022209080A1 (de) | Verfahren zum Kalibrieren eines Sensors, Recheneinheit und Sensorsystem | |
| DE112019007453T5 (de) | Sensorvorrichtung und Sensorvorrichtungssystem | |
| DE112023005124T5 (de) | Verfahren und system zur anomalie-erfassung in einem betrieblichen asset und verfahren zum reparieren eines betrieblichen assets | |
| DE102019214546A1 (de) | Computerimplementiertes Verfahren und Vorrichtung zur Optimierung einer Architektur eines künstlichen neuronalen Netzwerks | |
| EP0561792B1 (de) | Wissensbasierte steuer- und regeleinrichtung | |
| DE102013206274A1 (de) | Verfahren und Vorrichtung zum Anpassen eines nicht parametrischen Funktionsmodells | |
| DE102022200418A1 (de) | Vorrichtung, System und computerimplementiertes Verfahren zum Betreiben eines technischen Systems | |
| DE102020211596A1 (de) | Verfahren zum Generieren eines trainierten neuronalen Faltungs-Netzwerks mit invarianter Integrationsschicht zum Klassifizieren von Objekten | |
| EP1214631B2 (de) | Verfahren, anordnung und computerprogramm-erzeugnis zur simulation eines technischen systems | |
| DE102024204828A1 (de) | Verfahren und Vorrichtung zum Erstellen eines datenbasierten oder parametrischen Funktionsmodells für eine Steuerung oder Regelung eines technischen Systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A2 Designated state(s): JP US |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
| AK | Designated states |
Kind code of ref document: A3 Designated state(s): JP US |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A3 Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
| WWE | Wipo information: entry into national phase |
Ref document number: 1999916761 Country of ref document: EP |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 09646445 Country of ref document: US |
|
| WWP | Wipo information: published in national office |
Ref document number: 1999916761 Country of ref document: EP |
|
| WWW | Wipo information: withdrawn in national office |
Ref document number: 1999916761 Country of ref document: EP |

