WO2006001121A1 - データ圧縮装置及び方法,データ解析装置及び方法並びにデータ管理システム - Google Patents
データ圧縮装置及び方法,データ解析装置及び方法並びにデータ管理システム Download PDFInfo
- Publication number
- WO2006001121A1 WO2006001121A1 PCT/JP2005/008170 JP2005008170W WO2006001121A1 WO 2006001121 A1 WO2006001121 A1 WO 2006001121A1 JP 2005008170 W JP2005008170 W JP 2005008170W WO 2006001121 A1 WO2006001121 A1 WO 2006001121A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- euron
- neuron
- analysis
- data compression
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 210000002569 neuron Anatomy 0.000 claims abstract description 83
- 230000013016 learning Effects 0.000 claims abstract description 37
- 238000013144 data compression Methods 0.000 claims description 53
- 238000007405 data analysis Methods 0.000 claims description 39
- 238000004458 analytical method Methods 0.000 claims description 27
- 238000013523 data management Methods 0.000 claims description 18
- 238000001514 detection method Methods 0.000 claims description 10
- 238000007906 compression Methods 0.000 claims description 9
- 230000006835 compression Effects 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 abstract description 2
- 238000007781 pre-processing Methods 0.000 description 15
- 238000012423 maintenance Methods 0.000 description 13
- 238000012545 processing Methods 0.000 description 9
- 239000010705 motor oil Substances 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 239000003921 oil Substances 0.000 description 5
- 230000005856 abnormality Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 230000005484 gravity Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000002860 competitive effect Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000003449 preventive effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000002826 coolant Substances 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 239000010720 hydraulic oil Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
-
- 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
- G06N3/088—Non-supervised learning, e.g. competitive learning
Definitions
- Data compression apparatus and method data analysis apparatus and method, and data management system
- the present invention relates to a data compression apparatus and method, a data analysis apparatus and method, and a data management system suitable for use in compressing enormous data.
- Predictive maintenance refers to predicting deterioration and remaining life by diagnosing soundness by inference based on understanding of the load during operation'environmental information, past maintenance history database, failure physics, etc.
- the operating state of the machine is usually detected by a sensor attached to the target machine, and the raw data of the above operating state is collected by a data collecting device provided in the machine.
- This raw data is sent to the management center (for example, the service department of the company that performs machine maintenance) in real time or at a predetermined period, and the management center analyzes the raw data and diagnoses the soundness of the machine. It is like this.
- Patent Document 1 discloses a method for compressing a time series operation signal from a sensor into frequency distribution data or frequency distribution data.
- Patent Document 2 discloses a method of changing the transmission time interval of operation signals in accordance with the failure probability (bathtub curve).
- Patent Document 3 discloses a technique for integrating the detection frequencies for each unit time and determining the state of the machine from the frequency distribution in order to reduce the storage capacity of the memory.
- Patent Document 1 JP 2003-083848
- Patent Document 2 JP 2002-180502
- Patent Document 3 Japanese Patent Laid-Open No. 10-273920
- the present invention was devised in view of such a problem, and is capable of compressing a large amount of data, and more accurately reproducing the characteristics of the data of the compressed data source. It is an object to provide a compression apparatus and method.
- the data compression device detects a plurality of data sets having the value of n (n: natural number) parameters that vary according to the operation of the object.
- n natural number
- the neuron is trained by the unsupervised learning method of the Ural network. It is characterized by comprising a compression means for compressing by converting into a Euron model parameter.
- the data compression device of the present invention according to claim 2 is the same as the device according to claim 1, wherein the compression means is closest to each data set, and the neuron is the winner.
- the neuron model parameters represent the coordinate information of the neuron in the n-dimensional space, the average distance information from the neuron to each data set, and how many data sets the neuron represents. It is characterized by the fact that it includes weight information that expresses RU.
- the data compression device of the present invention according to claim 3 is the device according to claim 2, wherein the data compression means includes the predetermined number of -euroons after completion of the learning. It is characterized by erasing the Euron, who has never been a winner-a force that has not been set to Euron.
- a data compression apparatus is the apparatus according to any one of the first to third aspects, further comprising transmission means for transmitting the -Euron model parameter to the outside. Characterized by that! /
- the data analysis device of the present invention is a reception unit that receives the neuron model parameter transmitted by the transmission unit of the data compression device according to claim 4, and is received by the reception unit. And an analysis means for analyzing the data set based on the -Euron model parameter.
- the data analysis apparatus of the present invention is the apparatus according to claim 5, wherein the analysis means is a moving average of the neurons based on coordinate information and weight information included in the -Euron model parameter. It is characterized by analyzing by obtaining.
- the data analysis device of the present invention according to claim 7 is the device according to claim 5, wherein the analysis means is based on coordinate information, weight information, and average distance information included in the -Euron model parameter. It is characterized by analyzing the distribution density of the data set.
- the data management system according to the present invention described in claim 8 is characterized by comprising the data compression device according to claim 4 and the data analysis device according to any one of claims 5 to 7.
- the data management system of the present invention is the system according to claim 8, wherein the object is a construction machine, and the n parameters fluctuate according to the operation of the construction machine. It is characterized by being a parameter.
- the data compression method of the present invention described in claim 10 includes a detection step of detecting a plurality of data sets having n (n: natural number) parameter values that vary according to the motion of the object, and the detection Each data set detected in the step is input into an n-dimensional space, and a predetermined number of -euroons less than the number of the data sets are placed in the n-dimensional space--by the unsupervised learning method of the Euler network. It is characterized by the fact that neuron learning is performed, and the above-mentioned plurality of data sets are provided with a data compression step for compressing by converting the data set obtained by the learning into -characterizing the Euron model-converting into the Euron model parameters.
- the data compression method of the present invention according to claim 11 is the method according to claim 10, wherein in the data compression step, for each data set, the nearest distance is set to -Euron as a winner neuron, and the neuron
- the model parameters include coordinate information of the neuron in the n-dimensional space, average distance information to each data set of the neuron force, and weight information representing the force that the neuron represents. It's a feature! /
- the data compression method of the present invention according to claim 12 is the method according to claim 11, wherein, in the data compression step, the neuron that has never been set to the winner-Euron among the predetermined number of -Eurons. It is characterized by erasing.
- the data analysis method of the present invention described in claim 13 is based on the acquisition step of acquiring the neuron model parameter obtained by the data compression method of claim 12, and the -Euron model parameter acquired in the acquisition step. And an analysis step for analyzing the data set.
- the data analysis method of the present invention according to claim 14 is the method according to claim 13, The analysis step is characterized in that the analysis is performed by obtaining a moving average of the neuron based on the coordinate information and the weight information included in the -Euron model parameter.
- the data analysis method of the present invention according to claim 15 is the method according to claim 13, wherein, in the analysis step, the data is based on the coordinate information, the weight information, and the average distance information included in the -Euron model parameter. It is characterized by analyzing the distribution density of the set.
- a plurality of data sets (raw data) having the value power of n parameters can be obtained from a predetermined number of eurones, which is smaller than the number of data sets.
- Characterize the Euron model-it can be compressed into Euron model parameters.
- the compressed-Euron model parameters (compressed data) characterize the neuron model, i.e., characterize multiple datasets, so that the compressed data force also more accurately reproduces the characteristics of the original dataset. Is possible.
- the data analysis apparatus and method of the present invention it is possible to analyze an original data set (raw data) using neuron model parameters.
- the object can be diagnosed by analyzing the moving average of neurons and the distribution density of the data set.
- the effects of both the data compression device and the data analysis device can be obtained.
- FIG. 1 is a block diagram schematically showing a data management system according to an embodiment of the present invention.
- FIG. 2 is a plot of input data set points for engine speed and boost pressure for a certain operation dial.
- FIG. 3 This is a diagram when the input data set of engine speed and boost pressure shown in Fig. 2 is converted to the Euron model (idling-Euron and weak-Euron have been deleted).
- FIG. 4 is a diagram for explaining a moving average of Euron considering weight.
- FIG. 5 A graph showing the relationship between engine speed and boost pressure obtained by moving average of Euron considering weight.
- FIG. 6 is a graph when the moving average of the input data set is directly obtained.
- FIG. 7 is a diagram showing the distribution density of the input data set.
- FIG. 8 is a flowchart showing a procedure for data compression and data analysis according to an embodiment of the present invention.
- FIG. 9 is a flowchart showing a specific procedure for data preprocessing.
- FIG. 10 is a flowchart showing a specific procedure for data compression.
- FIG. 11 Flow chart showing the procedure of analysis with moving average of Euron considering weight.
- FIG. 12 is a flowchart showing the analysis procedure based on the distribution density of the input data set. Explanation of symbols
- the data management system according to the present embodiment is used for diagnosing whether there is an abnormality in machinery such as construction machinery.
- the following describes the data management system when applied to a hydraulic excavator as a construction machine. It should be noted that the application target of this data management system is not limited to this, and can be applied to all various objects that can vary depending on the operation or environment.
- FIG. 1 is a block diagram schematically showing a data management system according to an embodiment of the present invention.
- this data management system 1 is a management center that manages the hydraulic excavator 2 away from the operation site (for example, maintenance of the hydraulic excavator 2) from the operation site such as a construction site.
- the data of the hydraulic excavator 2 is transmitted to the service department of the company 10), and the soundness level of the hydraulic excavator 2 can be diagnosed by inference based on the received data at the management center 10 side.
- the data management system 1 mainly includes an on-board data compression device 3 provided in the hydraulic excavator 2 and a data analysis device 10 provided in the management center.
- the data compression device 3 includes a sensor 4, a data preprocessing unit (preprocessing unit) 5, a data compression unit (compression unit) 6, a transmission unit (transmission unit) 7, and an input device (keyboard, mouse, etc.) 8. It is mainly prepared.
- the sensor 4 and the data preprocessing unit 5 constitute detection means.
- an ECU electronic control unit
- the ECU includes an input / output device, a storage device (memory such as RAM and ROM), a central processing unit (CPU), and the like.
- the sensor 4 is provided corresponding to n parameters (variation factors) related to the hydraulic excavator 2, and when the hydraulic excavator 2 operates, the value of each parameter that varies according to the operation of the hydraulic excavator 2 is provided. X, X,..., X is detected (measured).
- the sensor 4 includes a sensor that directly detects the value of the corresponding parameter, and a sensor that processes certain detection data by calculation or the like to obtain the value of the corresponding parameter as a calculated value or an estimated value.
- parameters related to hydraulic excavator 2 include, for example, engine speed, engine oil temperature, engine oil pressure, engine coolant temperature, boost Hydraulic excavators such as pressure (supply pressure after supercharging), fuel consumption, exhaust gas temperature, power shift pressure, hide lick oil pressure, hide lick oil temperature, hide lick oil particulate counter, engine operating time 2
- boost Hydraulic excavators such as pressure (supply pressure after supercharging), fuel consumption, exhaust gas temperature, power shift pressure, hide lick oil pressure, hide lick oil temperature, hide lick oil particulate counter, engine operating time 2
- the data preprocessing unit 5 is capable of realizing its functions by a required processing program (data preprocessing program).
- This data preprocessing program can realize the processing procedure shown in FIG. . That is, first, it is determined whether or not the engine is rotating (step T10). If the engine is rotating, it is next determined whether or not the engine oil temperature is equal to or higher than a set value (step T20). . When the engine oil temperature becomes equal to or higher than the set value, the raw data detected by the sensor 4 is acquired (step ⁇ 30), and then it is determined whether or not the engine has stopped (step ⁇ 40). In other words, it continues to acquire raw data until the engine stops.
- the raw data is organized for each operation dial (step ⁇ 50), and a file is created for each operation dial and the raw data is saved (step ⁇ 60).
- the above operation dial is a dial in which the operator sets the engine speed according to the work content (work load).
- the data preprocessing unit 5 detects each parameter detected by the sensor 4 when the engine of the hydraulic excavator 2 rotates and the engine oil temperature becomes equal to or higher than the set value (that is, when the hydraulic excavator 2 starts normal operation). For example, a data set with a value of 1 is acquired (collected) every 1 second and stored in a storage device. Whether the engine is rotating or not can be determined by information from the sensor that detects the engine speed. Whether the engine oil temperature has exceeded the set value is determined by the engine oil temperature. It can be judged by the information of the sensor force to be detected. In the present embodiment, the period for acquiring data from the sensor 4 is set to one second, but this period can be arbitrarily set by input from the input device 8.
- the data preprocessing unit 5 continues to save the data set of each parameter force in the storage device until the operation start force of the excavator 2 ends (that is, when the engine stops).
- the data pre-processing unit 5 organizes a plurality of data sets stored in the storage device for each operation dial after the engine of the excavator 2 is stopped. A file is created for each file, and the data set is saved in each corresponding file.
- Fig. 2 is a plot of the data set (ie, raw data) of engine speed (Boost Pressure) for a certain operation dial.
- Boost Pressure engine speed
- the number of data sets is enormous.
- the engine speed axis (horizontal axis) in Fig. 2 indicates the speed range during normal operation.
- FIGS. 3, 5, 6 and 7 used below the axis of engine speed is indicated by the speed range during normal operation.
- the data compression unit 6 is capable of realizing its function by a required processing program (data compression program).
- This data compression program can realize the processing procedure shown in FIG. That is, after reading the operation dial file (step U10), a predetermined number of -Eurons are arranged in a random number in n dimensions (step U20), and neuron learning is performed (step U30). At this time, learning is repeated for a set number of times, and when learning is completed, idling-euron, weakness, and neurons (which will be described later) are deleted (step U40). Then, the compressed file data (hereinafter simply referred to as the compressed file) that includes the neuron's n-dimensional coordinates, average distance, and weight information is created (step U50). Repeat the above steps U10 to U50 for each operation dial! /, And then (step U60).
- the data compression unit 6 functions to compress the enormous number of data sets preprocessed by the data preprocessing unit 5 into a number of data that is much smaller than the number of these data sets.
- a self-organizing map [Self
- Unsupervised learning refers to learning using only the input data set when there is no clear target value (ie, "answer") for the input data set (hereinafter referred to as the input data set). This learning is based on several principles that are described below.
- the self-organization map is composed of an input layer (input data, that is, an input data set group) and a competition layer (neuron group), and features hidden in the input data set group are automatically displayed. This algorithm performs extraction and learning, and automatically finds the similarity of input data sets and arranges similar input data sets close to each other on the network.
- the number of neurons k, the initial learning rate ⁇ , the initial neighborhood region ,, and the set total number of learning steps ⁇ are set that are necessary for the calculation that represents the distribution of the input data set to the neurons.
- This learning condition can be arbitrarily set in advance by the input device 8.
- the input data set has the structure shown in the following formula (1).
- the number of neurons k is set to be much smaller than the number of input data sets 1 (L) (for example, several tens. K ⁇ 1).
- the initial connection weight m is expressed by the following equation (2) It has the structure shown in
- Neuron number k initial neuron number
- Vector X has the structure shown in the following equation (3).
- the Eugrid distance d can be obtained by the following equation (4).
- the Eugrid distance d is the smallest (ie, the most similar).
- the winner-Euron corresponding to the preset neighborhood N around the Euron is also determined.
- the neighborhood region N can be obtained by the following equation (5).
- the weight of the winner neuron m is updated by the learning rate ⁇ and approaches the input data set.
- the updated winner neuron can be obtained by the following equation (6).
- the winner-a force that has become like Euron-Euron (this "Idling-Euron") and very small numbers (below a predetermined number, eg 1 or 2) could only be representative of the input data set-Eurone (this is also called "Weak-Euron”) ) To delete.
- each input data set is input into the n-dimensional space, and a predetermined number k of Euron, which is much smaller than the number of input data sets 1 (el), is n-dimensional. It is placed randomly in the space, and the Euron is learned by the unsupervised learning method of the neural network. Then, after this learning, idling-Euron and weak-Euron are deleted, and a predetermined number of representative input data sets (ie, processed and detected raw data detected by sensor 4) are represented.
- Create a -Eron model neural arrangement
- Fig. 3 shows the power when the input data set of engine speed and boost pressure shown in Fig.
- the neuron model parameters are the remaining information after the idling-Euron and weak-Euron deletions--Eron's coordinate information in the n-dimensional space, and the neuron force It includes average distance information to the data set and weight information that represents how many data sets the neuron represents! Therefore, a huge number of input data sets are compressed into the number of -Euron model parameters shown in the following equation (7).
- the data compression unit 6 compresses the raw data as described above for each file (that is, for each operation dial), and then creates a compressed file corresponding to each file in the storage device. -Eron model parameters are now stored in the file! /
- the transmitting unit 7 transmits the compressed file created by the data compressing unit 6 to the outside. In this embodiment, the transmitting unit 7 transmits wirelessly using an antenna. Good.
- the data analysis device 10 of the management center is mainly configured with a reception unit (reception unit) 11, a data analysis unit (analysis unit) 12, and a determination unit 13.
- a reception unit reception unit
- a data analysis unit analysis unit
- a determination unit 13 Each function of the data analysis unit 12 and the determination unit 13 is realized by a processing program incorporated in an ECU (Electronic Control Unit) such as a computer (not shown).
- the ECU includes an input / output device, a storage device (memory such as RAM and ROM), a central processing unit (CPU), and the like.
- the receiving unit 11 receives the compressed file transmitted from the transmitting unit 6.
- the received compressed file is stored in a storage device (not shown).
- the data analysis unit 12 analyzes the input data set based on the ⁇ Euron model parameter in the compressed file received by the reception unit 11.
- the function of the data analysis unit 12 is realized by several processing programs (data analysis programs), which will be described later, but the analysis is performed by the two methods described below. I'm getting ready.
- the data analysis unit 12 has a processing program for realizing the analysis based on the moving average of Euron in consideration of the weight, and this processing program can realize the processing procedure shown in FIG. That is, first, a compressed file of a certain operation dial is read (step V10), and any two parameters (for example, engine speed and boost pressure) are selected (step V20). Each -Euron has information on coordinates, weights, and average distance for each parameter (measurement item), so it is easy to know the relationship between any parameters. Next, using the Euron coordinate information and weight information included in the neuron model parameters, as shown in Fig. 4, after determining the maximum value RB and minimum value LB of the neuron, it is assigned to a predetermined number of areas (called windows).
- Step V30 Divide (Step V30), and for each window-find the center of gravity (Cx, Cy) considering the Euron weight (Step V40).
- a moving average graph is created by connecting the center of gravity of each window (step V50).
- step V60 After performing the above steps V20 to V50 for all the n-dimensional parameters (step V60), the above steps V20 to V60 are performed for the compressed files of all operation dials (step V70).
- the center of gravity is calculated using the following equation (8).
- xi, yi Neuron point coordinates in the th window
- Fig. 5 is a graph showing the relationship between the engine speed and boost pressure determined by the moving average of neurons taking weights into account.
- the points of the input data set ie, raw data
- the -Euron points after learning are also displayed.
- the graph of the barycentric point obtained by the moving average of neurons has a shape that closely approximates the input data set group (distribution of the input data set).
- Fig. 6 is a graph when the moving average of the input data set is directly obtained.
- the data analysis unit 12 has a processing program for realizing an analysis based on the distribution density of the input data set, and this processing program can realize the processing procedure shown in FIG. That is, first, a compressed file of a certain operation dial is read (step W10), and any two parameters (for example, engine speed and boost pressure) are selected from the n-dimensional parameters (step W20). As described above, each -Euron has information on coordinates, weights, and average distances for each parameter (measurement item), so it is easy to know the relationship between any parameters. Next, using the neuron coordinate information, average distance information, and weight information included in the neuron model parameters, create a three-dimensional graph (in this case, engine speed, boost pressure, input data density) as shown in Fig. 7. (Step W30).
- the neuron's coordinate information power was also obtained-the position of the euron represents the peak (mountain) position
- the neuron's weight information power was also obtained-the size of the euron weight represents the peak height.
- the average distance from the input data set obtained from the average distance information corresponds to the slope of the slope of the peak, for example, the average distance is large (that is, the input data set is relatively close to the neuron position).
- the slope of the slope of the peak is smaller (slower) and the average distance is smaller (ie, the input data set is present at a relatively high density near the position of -Euron). Ho However, the slope of the slope of the peak becomes large (steep).
- the data analysis unit 12 can display the graph created in this way on the display 20.
- the data analysis unit 12 can analyze the Euron model parameters by two methods.
- the force analysis in which the relationship between the engine speed and the boost pressure is analyzed is performed for all parameters for each operation dial.
- the diagnosis of the hydraulic excavator 2 is performed based on each graph obtained by the data analysis unit 12 (actually, each information of the Euron model parameter that is the basis of the graph). Yes. For example, if the similarity with the graph during normal operation is calculated and this similarity is smaller than the preset value (that is, if the deviation from the graph during normal operation becomes large), the hydraulic excavator 2 is judged to be abnormal or deteriorated. In addition, when an abnormality has occurred in the excavator 2, the determination unit 13 displays that fact on the display 20 and notifies the operator of that fact!
- step S10 the raw data (input data set) of each parameter of the hydraulic excavator 2 is detected by the sensor 4 (step S10), and the raw data is preprocessed (step S20).
- step S30 the raw data is compressed using neurons (step S30), and the compressed data (that is, the -Euron model parameter) is transmitted to the management center as a compressed file (step S40).
- the management center analyzes the received compressed file (step S50) and determines whether or not there is any abnormality in the hydraulic excavator 2 (step S60).
- step S50 the process of step S20 shown in FIG. 8 is executed by the data preprocessing unit 5 along the flow shown in FIG. 9, and the process of step S30 shown in FIG. 10 is executed along the flow shown in FIG. 10, and the processing of step S50 shown in FIG. 8 is executed by the data analysis unit 12 along the flow shown in FIG. 11 and FIG.
- a plurality of input data sets (that is, raw data: see FIG. 2) having the value power of n parameters are used as the input data. It can be compressed into parameters that characterize the Euron model (obtained from Euron), which is smaller than the number of sets-Neuron model parameters (see Fig. 3), which makes it possible to reduce communication costs.
- the compressed single model parameters characterize the neuron model, i.e., characterize a plurality of input data sets
- the original input data set i.e., the raw data
- the original input data set i.e., the raw data
- the excavator 2 can be diagnosed by analyzing the moving average of neurons (see Fig. 5) and the distribution density of the input data set (see Fig. 7).
- the effects of both the data compression device and the data analysis device can be obtained.
- the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention.
- the operator may make a diagnosis by outputting the analysis result to a printing device such as a printer (not shown). Oh ,.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Operation Control Of Excavators (AREA)
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP05736519A EP1764728A4 (en) | 2004-06-25 | 2005-04-28 | DATA COMPROMISING DEVICE AND METHOD, DATA ANALYSIS DEVICE AND METHOD AND DATA MANAGEMENT SYSTEM |
US11/628,323 US7664715B2 (en) | 2004-06-25 | 2005-04-28 | Apparatus and method for compressing data, apparatus and method for analyzing data, and data management system |
CN2005800212019A CN1981297B (zh) | 2004-06-25 | 2005-04-28 | 数据压缩设备和方法,数据分析设备和方法,以及数据管理系统 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2004188419A JP4175296B2 (ja) | 2004-06-25 | 2004-06-25 | 建設機械のデータ処理装置及び建設機械のデータ処理方法 |
JP2004-188419 | 2004-06-25 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2006001121A1 true WO2006001121A1 (ja) | 2006-01-05 |
Family
ID=35779045
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2005/008170 WO2006001121A1 (ja) | 2004-06-25 | 2005-04-28 | データ圧縮装置及び方法,データ解析装置及び方法並びにデータ管理システム |
Country Status (5)
Country | Link |
---|---|
US (1) | US7664715B2 (ja) |
EP (1) | EP1764728A4 (ja) |
JP (1) | JP4175296B2 (ja) |
CN (1) | CN1981297B (ja) |
WO (1) | WO2006001121A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111178520A (zh) * | 2017-06-15 | 2020-05-19 | 北京图森智途科技有限公司 | 一种低计算能力处理设备的数据处理方法及装置 |
Families Citing this family (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007323401A (ja) * | 2006-06-01 | 2007-12-13 | Kagawa Univ | データ処理装置,データ復元装置,データ処理方法及びデータ復元方法 |
JP2008009934A (ja) * | 2006-06-30 | 2008-01-17 | Kagawa Univ | データ処理装置,データ処理方法,作業機械の遠隔診断システム及び作業機械の遠隔診断方法 |
JP2008009935A (ja) * | 2006-06-30 | 2008-01-17 | Kagawa Univ | データ処理装置,データ処理方法,作業機械の遠隔診断システム及び作業機械の遠隔診断方法 |
JP4853921B2 (ja) * | 2007-02-14 | 2012-01-11 | キャタピラー エス エー アール エル | 機体診断システム |
JP2008203941A (ja) * | 2007-02-16 | 2008-09-04 | Shin Caterpillar Mitsubishi Ltd | 機体稼働状態検知方法 |
JP2008202221A (ja) * | 2007-02-16 | 2008-09-04 | Shin Caterpillar Mitsubishi Ltd | 作業負荷判定方法および作業負荷判定システム |
JP2008202223A (ja) * | 2007-02-16 | 2008-09-04 | Shin Caterpillar Mitsubishi Ltd | 機体稼働状態検知方法 |
JP2008202222A (ja) * | 2007-02-16 | 2008-09-04 | Shin Caterpillar Mitsubishi Ltd | 機体稼働状態検知方法 |
JP5000539B2 (ja) * | 2008-01-29 | 2012-08-15 | 本田技研工業株式会社 | 内燃機関の排気浄化装置 |
JP4759034B2 (ja) * | 2008-09-29 | 2011-08-31 | 株式会社エー・アンド・デイ | モデル駆動型計測制御データ管理システム |
JP5134605B2 (ja) * | 2009-09-18 | 2013-01-30 | パナソニック デバイスSunx株式会社 | 信号識別装置 |
AU2010344438B2 (en) * | 2010-01-28 | 2014-11-06 | Hitachi Construction Machinery Co., Ltd. | Operation machine monitoring diagnosis device |
WO2011129819A1 (en) * | 2010-04-13 | 2011-10-20 | Empire Technology Development Llc | Combined-model data compression |
WO2011129816A1 (en) | 2010-04-13 | 2011-10-20 | Empire Technology Development Llc | Semantic compression |
WO2011129817A2 (en) | 2010-04-13 | 2011-10-20 | Empire Technology Development Llc | Semantic medical devices |
JP5663393B2 (ja) * | 2011-04-28 | 2015-02-04 | 日立建機株式会社 | 稼働機械及び保守点検情報生成装置 |
JP5963579B2 (ja) * | 2012-06-13 | 2016-08-03 | 住友重機械工業株式会社 | ショベルの管理装置及びショベルの異常検出方法 |
US9111182B1 (en) * | 2012-09-06 | 2015-08-18 | Hrl Laboratories, Llc | System, method, and computer program product for multispectral image processing with spiking dynamics |
CN104809069A (zh) * | 2015-05-11 | 2015-07-29 | 中国电力科学研究院 | 一种基于集成神经网络的源代码漏洞检测方法 |
US9513813B1 (en) | 2015-12-18 | 2016-12-06 | International Business Machines Corporation | Determining prefix codes for pseudo-dynamic data compression utilizing clusters formed based on compression ratio |
JP6960209B2 (ja) * | 2016-03-24 | 2021-11-05 | 三菱重工機械システム株式会社 | デッキクレーンシステム |
CN106779069A (zh) * | 2016-12-08 | 2017-05-31 | 国家电网公司 | 一种基于神经网络的异常用电检测方法 |
EP3340129B1 (en) * | 2016-12-21 | 2019-01-30 | Axis AB | Artificial neural network class-based pruning |
CN106991139B (zh) * | 2017-03-17 | 2018-08-07 | 广东蜂助手网络技术股份有限公司 | 一种基于人工神经网络的应用分发平台作弊检测方法 |
US10693495B2 (en) * | 2017-10-10 | 2020-06-23 | The Boeing Company | Data collection device with efficient data compression |
WO2021111511A1 (ja) * | 2019-12-03 | 2021-06-10 | 株式会社日立ハイテク | 探索装置、探索プログラム及びプラズマ処理装置 |
CN112783056B (zh) * | 2021-01-04 | 2022-09-23 | 潍柴动力股份有限公司 | Ecu的数据烧写方法、装置、设备及存储介质 |
JP7225445B2 (ja) * | 2021-01-14 | 2023-02-20 | 株式会社日立製作所 | ストレージシステム |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0721029A (ja) * | 1993-07-05 | 1995-01-24 | Komatsu Ltd | 推論装置 |
JPH0970048A (ja) * | 1995-06-21 | 1997-03-11 | Studio Gen:Kk | ベクトル量子化方法、画像情報圧縮方法、音声情報圧縮方法、圧縮画像情報記録媒体、圧縮画像情報再生装置及び圧縮画像情報通信装置 |
JP2001076172A (ja) * | 1999-09-07 | 2001-03-23 | Moriaki Sakakura | 曲面発生方法及びその装置並びに記憶媒体 |
Family Cites Families (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0810453B2 (ja) * | 1989-12-27 | 1996-01-31 | 日本電気株式会社 | 入れ子型トークン列の検索装置 |
JPH0652241A (ja) * | 1992-07-28 | 1994-02-25 | Nec Corp | 入れ子構造メッセージ検索装置 |
US5774629A (en) | 1993-07-05 | 1998-06-30 | Komatsu Ltd. | Inference apparatus using occurence rate data and degree of relevance data |
JP3334957B2 (ja) * | 1993-09-03 | 2002-10-15 | 松下電送システム株式会社 | ファクシミリ通信システム |
US7251637B1 (en) * | 1993-09-20 | 2007-07-31 | Fair Isaac Corporation | Context vector generation and retrieval |
JPH0836557A (ja) | 1994-07-25 | 1996-02-06 | Olympus Optical Co Ltd | クラスタ分類装置 |
JP3562126B2 (ja) * | 1996-04-09 | 2004-09-08 | 株式会社デンソー | Dma制御装置 |
JP3765551B2 (ja) | 1997-03-31 | 2006-04-12 | 株式会社小松製作所 | 建設機械の異常監視方法 |
US6847737B1 (en) * | 1998-03-13 | 2005-01-25 | University Of Houston System | Methods for performing DAF data filtering and padding |
EP1107693A4 (en) * | 1998-08-24 | 2003-03-19 | Univ Emory | METHOD AND APPARATUS FOR PREDICTING THE APPEARANCE OF CRISES BASED ON CHARACTERISTICS DERIVED FROM SIGNALS INDICATING BRAIN ACTIVITY |
SE9804127D0 (sv) * | 1998-11-27 | 1998-11-27 | Astra Ab | New method |
GB9902115D0 (en) * | 1999-02-01 | 1999-03-24 | Axeon Limited | Neural networks |
JP2000266570A (ja) | 1999-03-15 | 2000-09-29 | Omron Corp | 定常状態と非定常状態を弁別する信号処理装置 |
AU2610701A (en) * | 2000-01-03 | 2001-07-16 | Efeckta Technologies Corporation | Efficient and lossless conversion for transmission or storage of data |
US6876779B2 (en) * | 2000-01-24 | 2005-04-05 | Sony Côrporation | Method and apparatus of reconstructing audio/video/image data from higher moment data |
JP4024993B2 (ja) * | 2000-06-01 | 2007-12-19 | 株式会社コナミデジタルエンタテインメント | 候補キャラクターデータ登録方法、記録媒体及びビデオゲーム装置 |
US7389208B1 (en) * | 2000-06-30 | 2008-06-17 | Accord Solutions, Inc. | System and method for dynamic knowledge construction |
US6728728B2 (en) * | 2000-07-24 | 2004-04-27 | Israel Spiegler | Unified binary model and methodology for knowledge representation and for data and information mining |
JP5109197B2 (ja) | 2000-09-20 | 2012-12-26 | 株式会社小松製作所 | 作業機械の管理装置および作業機械の通信装置 |
US7170997B2 (en) * | 2000-12-07 | 2007-01-30 | Cryptico A/S | Method of generating pseudo-random numbers in an electronic device, and a method of encrypting and decrypting electronic data |
JP2002221570A (ja) * | 2001-01-26 | 2002-08-09 | Nec Corp | 航跡追尾装置及び航跡追尾方法 |
JP2003083848A (ja) | 2001-09-11 | 2003-03-19 | Hitachi Constr Mach Co Ltd | 建設機械の稼働診断装置 |
NZ515527A (en) * | 2001-11-15 | 2003-08-29 | Auckland Uniservices Ltd | Method, apparatus and software for lossy data compression and function estimation |
JP2003296696A (ja) | 2002-04-02 | 2003-10-17 | Sysmex Corp | クラスタリング装置およびクラスタリング方法 |
US7471724B2 (en) * | 2003-06-23 | 2008-12-30 | Vichip Corp. Limited | Method and apparatus for adaptive multiple-dimensional signal sequences encoding/decoding |
-
2004
- 2004-06-25 JP JP2004188419A patent/JP4175296B2/ja not_active Expired - Fee Related
-
2005
- 2005-04-28 WO PCT/JP2005/008170 patent/WO2006001121A1/ja not_active Application Discontinuation
- 2005-04-28 US US11/628,323 patent/US7664715B2/en not_active Expired - Fee Related
- 2005-04-28 CN CN2005800212019A patent/CN1981297B/zh not_active Expired - Fee Related
- 2005-04-28 EP EP05736519A patent/EP1764728A4/en not_active Withdrawn
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0721029A (ja) * | 1993-07-05 | 1995-01-24 | Komatsu Ltd | 推論装置 |
JPH0970048A (ja) * | 1995-06-21 | 1997-03-11 | Studio Gen:Kk | ベクトル量子化方法、画像情報圧縮方法、音声情報圧縮方法、圧縮画像情報記録媒体、圧縮画像情報再生装置及び圧縮画像情報通信装置 |
JP2001076172A (ja) * | 1999-09-07 | 2001-03-23 | Moriaki Sakakura | 曲面発生方法及びその装置並びに記憶媒体 |
Non-Patent Citations (1)
Title |
---|
See also references of EP1764728A4 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111178520A (zh) * | 2017-06-15 | 2020-05-19 | 北京图森智途科技有限公司 | 一种低计算能力处理设备的数据处理方法及装置 |
CN111178520B (zh) * | 2017-06-15 | 2024-06-07 | 北京图森智途科技有限公司 | 一种构建神经网络的方法及装置 |
Also Published As
Publication number | Publication date |
---|---|
US7664715B2 (en) | 2010-02-16 |
JP4175296B2 (ja) | 2008-11-05 |
CN1981297A (zh) | 2007-06-13 |
CN1981297B (zh) | 2011-12-07 |
JP2006011849A (ja) | 2006-01-12 |
EP1764728A4 (en) | 2010-04-21 |
EP1764728A1 (en) | 2007-03-21 |
US20070233623A1 (en) | 2007-10-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2006001121A1 (ja) | データ圧縮装置及び方法,データ解析装置及び方法並びにデータ管理システム | |
US7464063B2 (en) | Information processor, state judging unit and diagnostic unit, information processing method, state judging method and diagnosing method | |
Lu et al. | Dominant feature selection for the fault diagnosis of rotary machines using modified genetic algorithm and empirical mode decomposition | |
JP5026433B2 (ja) | プロセスモデルベースの仮想センサシステムおよび方法 | |
CN106779200A (zh) | 基于在历史数据中进行相似搜索的风电机组状态预测方法 | |
Mosallam et al. | Component based data-driven prognostics for complex systems: Methodology and applications | |
JP2007257366A (ja) | 診断装置及び診断方法 | |
CN116861331A (zh) | 一种融合专家模型决策的数据识别方法及系统 | |
CN118095813B (zh) | 基于bim技术的地基沉降的可视化监测方法及系统 | |
Kumar et al. | Latest innovations in the field of condition-based maintenance of rotatory machinery: a review | |
CN117969092B (zh) | 一种盾构机主轴承的故障检测方法、设备及介质 | |
Rizvi et al. | From data to insight, enhancing structural health monitoring using physics-informed machine learning and advanced data collection methods | |
Movsessian et al. | Feature selection techniques for modelling tower fatigue loads of a wind turbine with neural networks | |
CN117521912A (zh) | 一种碳排放测算模型及对比评价方法及其应用 | |
JP2008009935A (ja) | データ処理装置,データ処理方法,作業機械の遠隔診断システム及び作業機械の遠隔診断方法 | |
CN117994955A (zh) | 水电机组温度报警模型建立和报警方法及装置 | |
CN117218602A (zh) | 一种结构健康监测数据异常诊断方法及系统 | |
CN114720129B (zh) | 一种基于双向gru的滚动轴承剩余寿命预测方法及系统 | |
CN113744081B (zh) | 窃电行为分析方法 | |
JP2008009934A (ja) | データ処理装置,データ処理方法,作業機械の遠隔診断システム及び作業機械の遠隔診断方法 | |
CN118095023B (zh) | 一种刀盘结构的有限元分析方法和系统 | |
CN118092192B (zh) | 一种流体罐装参数优化控制方法及系统 | |
CN118504847B (zh) | 基于数字孪生技术的智慧梁场管理方法及系统 | |
CN118656617A (zh) | 一种生成瓦斯浓度预测模型的方法和终端设备 | |
CN118747322A (zh) | 一种基于深度学习的医疗器械故障类别检测方法及系统 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS KE KG KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2005736519 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 11628323 Country of ref document: US Ref document number: 2007233623 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 200580021201.9 Country of ref document: CN |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: DE |
|
WWP | Wipo information: published in national office |
Ref document number: 2005736519 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 11628323 Country of ref document: US |