CN105247430A - Data processing device and data processing method - Google Patents
Data processing device and data processing method Download PDFInfo
- Publication number
- CN105247430A CN105247430A CN201480030655.1A CN201480030655A CN105247430A CN 105247430 A CN105247430 A CN 105247430A CN 201480030655 A CN201480030655 A CN 201480030655A CN 105247430 A CN105247430 A CN 105247430A
- Authority
- CN
- China
- Prior art keywords
- data
- template
- pattern
- similarity
- comparison pattern
- 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
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1674—Programme controls characterised by safety, monitoring, diagnostic
-
- 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
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0221—Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
-
- 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
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/37—Measurements
- G05B2219/37508—Cross correlation
Abstract
The present application provides a data processing device and a data processing method whereby it is possible to easily and highly precisely extract data for comparison from a large quantity of data which occurs in time series. As an example, a data processing device (40) comprises: a storage unit (44) which stores a large quantity of inputted data which occurs in time series; and a pattern extraction unit (45) which respectively extracts a template of an operation pattern which is a reference, and a comparison pattern which is similar to the template, from the data accumulated in the storage unit (44). The pattern extraction unit (45) further comprises: a template creation unit (451) which creates the template from the data stored in the storage unit; a similarity acquisition unit (453) which acquires a similarity between the comparison pattern which is selected from the data stored in the storage unit and the template; and a comparison pattern extraction unit (454) which extracts and stores a comparison pattern which is a similar operation pattern to the template, on the basis of the similarity which is acquired from the similarity acquisition unit.
Description
Technical field
The present invention relates to a kind of data processing equipment and the data processing method that process the mass data that temporally sequence produces.
Background technology
Process the data processing method of the mass data that temporally sequence produces, such as, be used in the method for diagnosing faults (with reference to patent documentation 1) of industrial robot.
In general, the method for diagnosing faults of industrial robot is, collects and extracts the data of servo motor as the industrial robot of drive source, the process that professional etiquette of going forward side by side is fixed, using these data to monitor the aging of industrial robot.
Specifically, in order to monitor mechanism part aging as mechanical part, such as form following system: the action making certain determine the action moved into by the semiconductor crystal wafer be contained in storage box etc. as utilized the hand of industrial robot, taken out of etc. is implemented termly, and the data obtained at that time, compare with reference data when importing at first, if abnormal, then give the alarm.
Prior art document
Patent documentation
Patent documentation 1: Japanese Unexamined Patent Publication 2007-172150 publication
Summary of the invention
Invent technical matters to be solved
But the collection action of reference data in the past separates with common production action and prepares and implement.Therefore, because operation condition and the patterns etc. such as speed operate different from common production, therefore identify that aging analysis is also inaccurate.Namely, exist and be difficult to accurately, from the mass data that temporally sequence produces, easily extract the data wishing to compare, the preconditions such as the operation condition of the data that this hope is compared, environmental baseline and pattern are identical with the data of the pattern as benchmark.
The object of the present invention is to provide a kind of can accurately, from the mass data that temporally sequence produces, easily extract the data processing equipment and data processing method of wishing the data compared.
The technical scheme that technical solution problem adopts
A first aspect of the present invention relates to a kind ofly collects and processes the data processing equipment of the mass data that temporally sequence produces, and described data processing equipment has: storage part, the described data that described storage part storage is transfused to; And schema extraction portion, described schema extraction portion extracts template as the pattern of benchmark and the comparison pattern similar to described template respectively from the described data being stored in described storage part, described schema extraction portion comprises: template generation portion, and described template generation portion generates described template from described data; Similarity obtaining section, described similarity obtaining section obtains the similarity of comparison pattern and the described template selected from described data; And comparison pattern extraction unit, described comparison pattern extraction unit is extracted based on the similarity obtained by described similarity obtaining section and is stored as the comparison pattern of the pattern similar to described template.
By this data processing equipment, can accurately, from the mass data that temporally sequence produces, easily wish the data that compare in advance.
The described similarity identification part in described schema extraction portion of the present invention also can for identifying the structure of described similarity based on the correlation between described template and described comparison pattern.
According to this structure, owing to using correlation to try to achieve similarity, therefore formula is not the complicated formula of high order, therefore, it is possible to shorten the processing time, can be the computing machine without high throughput yet, have good versatility.
The present invention also can for following structure: the servo data that the described data that temporally sequence produces comprise the director data that inputs to carry out action control to industrial robot and export from the described industrial robot of the action based on this director data, described template and described comparison pattern generate from described director data, and the more described template in described schema extraction portion and the described data corresponding to described comparison pattern.Preferred described servo data comprises torque data, and the more described template in described schema extraction portion and the described torque data corresponding to described comparison pattern.
According to this structure, although exist in the prior art and be difficult to the problem setting identical initial conditions, but be identical be used as prerequisite with identical initial conditions (template and pattern) due to this structure, therefore have advantage comparing on this servo data.
A second aspect of the present invention relates to a kind ofly collects and processes the data processing method of the mass data that temporally sequence produces, and described data processing method has: storing step, in described storing step, the described data of input is stored in storage part; And template extraction step, in described template extraction step, the template as the pattern of benchmark and the comparison pattern similar to described template is extracted respectively from the described data being stored in described storage part, described schema extraction step comprises: template generation step, in described template generation step, generate described template from described data; Similarity obtains step, obtains in step the similarity of comparison pattern and the described template obtaining and select from described data in described similarity; And comparison pattern extraction step, extract based on being obtained the similarity that step obtains by described similarity and be stored as the comparison pattern of the pattern similar to described template in described comparison pattern extraction step.
Invention effect
According to the present invention, can accurately, from the mass data that temporally sequence produces, easily extract the data wishing to compare, the preconditions such as the operation condition of the data that this hope is compared, environmental baseline and pattern are identical with reference data.
Accompanying drawing explanation
Fig. 1 is the figure of the overview of the robot information handling system illustrated involved by embodiments of the present invention.
Fig. 2 is the block diagram of the configuration example of the data processing equipment illustrated involved by present embodiment.
Fig. 3 is the figure that the example data obtained from the controller in the data processing equipment involved by present embodiment stored to storage part is shown.
Fig. 4 is the figure of the example that generation action template is shown.
Fig. 5 is the figure of the action example of the hand of the robot body illustrated when to generate template in the data processing equipment involved by present embodiment.
Fig. 6 is the figure illustrated explicitly by the storage box of the hand in robot body and moving object object.
Fig. 7 is the figure using hand and storage box to specifically illustrate the action of the hand corresponding to Fig. 5.
The figure of the search method of comparison pattern when Fig. 8 is the similarity for illustration of using correlation as the data processing equipment involved by present embodiment.
Fig. 9 is the figure illustrated by the action pictorialization of the encircled portion in Fig. 8.
Figure 10 is the process flow diagram of the action of entirety for illustration of the robot information handling system involved by present embodiment.
Embodiment
Below, by reference to the accompanying drawings embodiments of the present invention are described.
In the following embodiments, be described for the robot information handling system with data processing equipment, this data processing equipment collects director data and servo data, described director data inputs to carry out action control to the industrial robot as industrial machinery, and described servo data exports from the industrial robot of the action based on director data.
Fig. 1 is for illustrating the block diagram of the overview of the robot information handling system involved by embodiments of the present invention.
As shown in Figure 1, robot information handling system 10 is configured to have robot body 20, the controller 30 of action of control machine human agent 20 and data processing equipment 40.
(structure of robot body and controller and function)
In the present embodiment, by robot body 20 and the controller 30 configuration example industrial robot 50 as carrying semiconductor crystal wafer or LCDs glass substrate (calling semiconductor crystal wafer etc. in the following text).The action of controller 30 control machine human agent 20, and using as the input data in the common production of robot body 20 director data (position, speed) and as export data servo data (torque, deviation etc.) temporally sequence export data processing equipment 40 to.That is, director data and servo data are the mass data of temporally sequence generation.Say further, the director data (position, speed) as input data is the data generated by program etc., and is not by the data of aging impact.In addition, be easily by the aging data of the mechanical part of formation industrial robot and mechanism part etc. as exporting the servo data (torque, deviation etc.) of data.
Drive source as the robotic arm 20ARM of industrial robot 50 uses servo motor, the controller 30 of present embodiment described later realizes following control: export director data to servo motor from controller 30, and the detected value as servo data exported from the scrambler being installed on servo motor is fed back to controller 30, by making the changes of following desired value (director data) consistent with desired value (director data) such as the position of rotation of servo motor.
Robot body 20 has robot base 21, arm standard shaft portion 22, first arm 23, second arm 24 and such as hand 25.
In robot body 20, the arm standard shaft portion 22 that the Z-direction (above-below direction in Fig. 1) along the rectangular coordinate system set in FIG extends is lifting and is rotatably configured at robot base 21 place.
The base end part of the first arm 23 is fixed on the upper end in arm standard shaft portion 22.The base end part of the second arm 24 is rotatably installed on the terminal part of the first arm 23.To keep and the hand 25 carried as the semiconductor crystal wafer etc. of moving object object is rotatably configured at the terminal part of the second arm 24.
The mode that this hand 25 such as can load thereon with conveying object bodies such as semiconductor crystal wafers is formed.Hand 25 is formed as to keep loading the moving object objects such as semiconductor crystal wafer thereon in the mode that can decontrol.
Second arm 24 is fixedly installed in the second arm turning axle 26, second arm turning axle 26 by its base end part and is rotatably arranged at the terminal part of the first arm 23.Hand 25 is fixedly installed in hand (Hand) axle 27 by its base end part, and hand axle 27 is rotatably configured at the terminal part of the second arm 24.
Like this, the robotic arm 20ARM of robot body 20 comprises arm standard shaft portion 22, first arm 23, second arm 24, hand 25, second arm turning axle 26 and hand axle 27.This robotic arm 20ARM is called as SCARA (SelectiveComplianceAssemblyRobotArm: select compliance to put together machines arm) the horizontal articulated arm of type, by the action of the controller 30 control machine arm 20ARM by the control part as robot, hand 25 can move to the desired position of X-axis, Y-axis and Z-direction.
Robot body 20 has: the first arm drive division 201 driving the first arm 23 to rotate, the second arm drive division 202 driving the second arm 24 to rotate, the hand axle drive division 203 driving hand axle 27 to rotate and the lifting drive division 204 that arm standard shaft portion 22 is driven along Z-direction lifting.
First arm drive division 201 is configured at the inner space of robot base 21, and comprises the servo motor and Poewr transmission mechanism thereof that are built-in with scrambler.Second arm drive division 202 is configured at the inner space of the first arm 23, and comprises the servo motor and Poewr transmission mechanism thereof that are built-in with scrambler.Hand axle drive division 203 is configured at the inner space of the second arm 24, and comprises the servo motor and Poewr transmission mechanism thereof that are built-in with scrambler.
Poewr transmission mechanism such as has reductor, and the power transmission of servo motor is to the input side of reductor, and its torque is exaggerated by the amplification ratio of setting in advance, and its rotational speed is decelerated by the reduction gear ratio of setting in advance, and exports from the outgoing side of reductor.Thus, by the power that exports from the outgoing side of reductor, actuating arm standard shaft portion 22, second arm turning axle 26 and hand axle 27 rotate respectively.Thus, the first arm 23, second arm 24 and hand 25 is driven to rotate respectively.
Lifting drive division 204 is configured at the inside of robot base 21, and is realized by the ball screw framework employing the motor that can adjust angular displacement.Such as, the motor that lifting drive division 204 has ball-screw, the screw thread fixed body be threaded with this ball-screw and drives ball-screw to rotate, and screw thread fixed body is fixed in arm standard shaft portion 22.The motor of lifting drive division 204 such as uses the servo motor being built-in with scrambler.
In the robot body 20 with this structure, arm standard shaft portion 22 passes through the first arm drive division 201 relative to robot base 21 around rotation RX1 rotary actuation.Thus, the first arm 23 relative to robot base 21 around rotation RX1 rotary actuation.
Second arm turning axle 26 passes through the second arm drive division 202 relative to the first arm 23 around rotation RX2 rotary actuation.Thus, the second arm 24 relative to the first arm 23 around rotation RX2 rotary actuation.
Hand axle 27 passes through hand axle drive division 203 relative to the second arm 24 around rotation RX3 rotary actuation.Thus, hand 25 relative to the second arm 24 around rotation RX3 rotary actuation.
About robot body 20, general in common production action, first arm drive division 201, second arm drive division 202, hand axle drive division 203 and lifting drive division 204 each servo motor drive according to the position inputted by controller 30 or speed command, and by the servo data such as torque, deviation temporally sequence export controller 30 to.
Controller 30 drives 202 to the first arm drive division 201, second arm as control object, hand axle drive division 203 and upgrading drive division 204 outgoing position instruction or speed command, and carrys out each drive division of FEEDBACK CONTROL by the angle position that the scrambler of the servo motor from each drive division obtains each servo motor.Thus, hand 25 can align with target location accurately.
Controller 30 exports instruction (position or speed) and these a large amount of data S30 that temporally sequence produces of servo data (torque, deviation etc.) to data processing equipment 40, and described instruction is the information relevant to the drived control of the servo system drive source (servo motor) of the robot body 20 in common production action.Here, the position data (or speed data) as director data is the input data inputted to carry out action control to industrial robot 50, and for inputing to the data of the drive division corresponding to robot body 20 from controller 30.Servo data is the output data exported from the industrial robot 50 carrying out action control based on director data, and for exporting the data of controller 30 to from the drive division of robot body 20.Controller 30 is formed in such a way: using the position data (or speed data) as director data and as exporting the servo data (torque, deviation etc.) of data in conjunction with the common production action of robot body 20, can export the CPC40 as data processing equipment to.In addition, in the present embodiment, from controller 30 by above-mentioned position data (or speed data) and servo data (torque, deviation etc.) temporally sequence export.
Controller 30 is such as configured to comprise CPU (CentralProcessingUnit: central processing unit) and storer 32.The operation program that CPU31 implements to be stored in storer 32 carrys out control machine human agent 20.Storer 32 can store the data (position data) relevant to the taught point of the action for control machine human agent 20, and hand 25 moves to assigned position based on the data relevant to the taught point being stored in storer 32.Further, also store in memory 32 the data relevant to the shape of hand 25, size and with shape, the data that size is relevant of semiconductor crystal wafer etc. being held in hand 25.
Above, the structure of robot body 20 and controller 30 and function are illustrated.Next, the structure of the data processing equipment involved by present embodiment and function are specifically described.
(structure of data processing equipment and function)
Data processing equipment 40 involved by present embodiment has following function: collect export from controller 30 in conjunction with the common production action of robot body 20, as the position data (or speed data) of the information relevant to the drived control of servo system drive source and servo data (torque, deviation etc.), and implement the process of regulation, from the mass data that temporally sequence produces, such as easily extract the data that the hope roughly the same with reference data of the preconditions such as operation condition, environmental baseline and pattern is compared accurately.
The mass data that in general data processing equipment 40 involved by present embodiment produces the temporally sequence inputted from controller 30 implements the process of regulation, and is configured to comprise: the storage part of storaging position data (or speed data) and servo data; And schema extraction portion, this schema extraction portion from the extracting data being stored in storage part as the template (reference data) of the pattern of benchmark and the comparison pattern similar with this template (data that hope is compared).
The schema extraction portion of data processing equipment 40 has: based on the director data as input data in the mass data of temporally sequence generation, namely position-based data (or speed data) generate the function as the template of the pattern of reference data; The similarity obtaining the similarity as the comparison pattern of the data of wishing to compare and the template of generation selected the mass data produced from temporally sequence obtains function; And to extract based on the similarity obtained and to be stored as the comparison pattern abstraction function of the comparison pattern of the pattern similar to template.
The similarity recognition function in schema extraction portion identifies similarity according to the correlation such as between template and comparison pattern.Like this, owing to using similar value to try to achieve similarity, therefore formula is not the complicated formula of high order, therefore, it is possible to shorten the processing time, can be the computing machine without high throughput yet, have good versatility.
In the present embodiment, the mass data that temporally sequence produces comprises: the director data that inputs to carry out action control to industrial machine human agent 20 and the output data from the industrial machine human agent 20 according to instruction action, template and comparison pattern generate from director data.In addition, schema extraction portion compares the data corresponding to template and comparison pattern.As described later, template and the comparison pattern servo data respectively with corresponding is interrelated, and schema extraction portion compares the servo data corresponding to template and comparison pattern.That is, in the present embodiment, by comparing the servo data corresponding to template and comparison pattern, aging impact can be quantized.Servo data such as comprises torque data, and schema extraction portion compares the torque data corresponding to template and comparison pattern.
Fig. 2 is for illustrating the block diagram of the configuration example of the data processing equipment involved by present embodiment.The data processing equipment 40 of Fig. 2 is made up of individual computing machine (PC) etc., and has control part 41, input part 42, efferent 43, storage part 44 and schema extraction portion 45.
Control part 41 is configured to comprise the various storeies such as CPU and ROM (Read-OnlyMemory: ROM (read-only memory)), RAM (Random-AccessMemory: random access memory) and external memory storage, and carries out the control of each several part, the storage control of the temporary transient of data, the transfer control etc. of data.
Input part 42 receives the information inputted by keyboard, mouse etc. by user, and the information of reception is supplied to control part 41.Further, the data exported from robot body 20 are also received by control part 41 and are sent to storage part 44.
Efferent 43 is shown by display screens such as CRT (CathodeRayTube: cathode-ray tube (CRT)) or LCD (LiquidCrystalDisplaydevice: liquid crystal display), or is printed by printer.
The input data (position, speed command) of the mass data produced as temporally sequence that storage part 44 stores various setting data (program etc.), provide to industrial machine human agent 20, as the servo data (torque data, deviation data etc.) of the output data exported from industrial machine human agent 20 and the data of comparative result.
Schema extraction portion 45 generates template by the input data (position command, speed command) being supplied to industrial machine human agent 20, and extracts the preconditions such as operation condition, environmental baseline, the pattern pattern (comparison pattern) roughly the same with template (reference data).This schema extraction portion 45 has template generation portion 451, compares (action) schema creation portion 452, similarity obtaining section 453 and comparison pattern extraction unit (comparing section) 454.
Template generation portion 451 generates the template of the pattern as benchmark from the mass data that the temporally sequence being stored in storage part 44 occurs.Comparison pattern generating unit 452 generates the comparison pattern (pattern) as the object compared with the template be generated from the mass data that temporally sequence occurs.
Similarity obtaining section 453 obtains the similarity of comparison pattern and the template selected from the mass data that temporally sequence produces.Here, as mentioned above, the instruction (position, speed) as input data of template and comparison pattern (pattern) is the data generated by program etc., is not by the data of burn-in effects.Therefore, think that template is roughly consistent with the precondition such as operation condition, environmental baseline, pattern of comparison pattern (pattern).Comparison pattern extraction unit 454 extracts the comparison pattern as the pattern similar to template based on the similarity obtained by similarity obtaining section 453, and is stored in storage part 44 under the control of such as control part 41.
Next, the process of extraction comparison pattern is described in conjunction with data example, and this comparison pattern is the pattern roughly the same with the templating of the pattern in this schema extraction portion 45.Here, to identify that the process of similarity is described based on the correlation between template and comparison pattern.
Fig. 3 is for illustrating the figure of the example data obtained from the controller in the data processing equipment involved by present embodiment stored to storage part.Fig. 4 is the figure of the example that generation action template is shown.
(storing the data obtained from controller)
The position command obtained from controller 30 and torque data are stored in storage part 44 in mode as shown in Figure 3 by data processing equipment 40.In figure 3, clear record collect data moment, as the position command pulse of the input data in each moment and the Driving Torque as the output data corresponding with position command pulse.Position command pulse represents the output number (umber of pulse) of the scrambler being installed on each motor.Further, torque data is electric current respective value, is namely the internal control value proportional with the current value being supplied to motor.The example of Fig. 3 is the figure of the part extracting position command and the torque data obtained in units of about 20msec.In addition, Hand, Z, TH etc. correspond to the title of the action axle of robot, and the hand axle 27 of such as action axle Hand and Fig. 1 is corresponding, and action axle Z is corresponding with the arm standard shaft in the arm standard shaft portion 22 being elevated driving in the Z-axis direction.Further, action axle TH is such as corresponding with the second arm turning axle 26 of Fig. 1.
(login of action template)
Hope is confirmed that in template generation portion 451 group of the continuous data of the position command (or speed command) of the axle of the change of servo data in the past that (check) storage part 44 stores selects as such as template TMP as shown in Figure 4 by the schema extraction portion 45 of data processing equipment 40.Specifically, template TMP when industrial robot 50 for the first time for the production of complete from test working time data generate, generate from data when can run well when implementing repairing or part exchanging to the Poewr transmission mechanism of industrial robot.In other words, reference data is obtained when Poewr transmission mechanism etc. does not occur aging.Like this, if do not occur aging, then the time in office obtains reference data, therefore, does not need strict regulations reference data to obtain opportunity.In the diagram, horizontal axis representing time, the longitudinal axis represents the displacement of hand 25 with relative value.Fig. 4 is an example after all, is to figure the figure by logging in using the position command pulse continuous print data group of the Hand axle of Fig. 3 as action template TMP to easily understand.
Fig. 5 is for illustrating the figure of the action example of the hand of robot body when to generate template in the data processing equipment involved by present embodiment.Fig. 6 is the figure illustrated explicitly by the storage box of the hand in robot body and storing semiconductor wafer etc. (moving object object).Fig. 7 is the figure using hand and storage box to specifically illustrate the action of the hand corresponding to Fig. 5.In Fig. 6 and Fig. 7, the storage box of 60 expression storing semiconductor wafers etc. (moving object object), 70 represent the semiconductor crystal wafer etc. as an example of moving object object.Further, in the figure 7,61 expressions are formed at some framves of the multiple mounting framves in storage box 60.In addition, Fig. 7 represents the action that semiconductor crystal wafer etc. takes out of from storage box by robot.The Z-direction of hand comprises the action of the movement from bottom to top in figure.
In Figure 5, step ST1 to ST4 is corresponding with the symbol ST1 to ST4 used in the chart of the template of Fig. 4.In step ST1, make the hand 25 being positioned at primary importance PT1 start to stretch to the direction of storage box 60, and as shown in Fig. 7 (A), be moved horizontally to the second place PT2 of semiconductor crystal wafer 70 grade that can keep carrying.This state is step ST2.Next, as as shown in Fig. 7 (B), kept with the method for the lower surface portion of clamping of semiconductor wafers 70 grade by hand 25 at second place PT2, and as shown in Fig. 7 (C), make arm standard shaft portion 22 start to rise (vertically moving) at second place PT2 along Z-direction.Then, in step ST3, as shown in Fig. 7 (C), complete rising at the 3rd PT3 place, position of Z-direction.Next, hand 25 is moved horizontally to the direction left from storage box 60, and be back to the 4th position PT4, thus complete a series of action.
(content that retrieval is identical with above-mentioned action template in the data area of the time that hope is compared)
Next, the pattern (comparison pattern) that retrieval is identical with above-mentioned action template in the data area of the time that hope is compared.
Fig. 8 is the figure of the search method in order to comparison pattern when using the similarity of correlation as the data processing equipment involved by present embodiment is described.In this example embodiment, obtain action template position command pulse and as the correlation between the position command pulse in the data area of object.As an example of intercept method, Fig. 8 is by correlation pictorialization.Therefrom extract the part (with the encircled portion shown in symbol A in Fig. 8) that correlation is more than 0.999 as much as possible.In addition, why extract correlation be more than 0.999 in part because: as shown in Figure 6, industrial robot 50 and hand 25 different because of frame with the relation of the mounting frame 61 being formed at storage box 60.Further, although do not illustrate, storage box 60 is configured with multiple, and therefore, industrial robot 50 is also different from the position relationship of each storage box.This difference causes the difference of correlation as shown in Figure 8.
(intercepting action)
Fig. 9 is the figure illustrated by the action pictorialization of the encircled portion in Fig. 8.When the comparison pattern of Fig. 9 compares with the action template of Fig. 4, both can be considered as identical in fact action (roughly the same action) or similar action.
(comparing the torque data of (intercepting) pattern of template and extraction)
As mentioned above, servo data, the such as torque data of comparison template and the identical pattern (comparison pattern) of extraction.Thereby, it is possible to confirm the so-called aging of the servo system drive division of (check) such as robot body 20.In other words, owing to achieving the extraction of the same action pattern launched in period, therefore confirm that (check) is aging by torque data more therebetween.In addition, comparative approach can adopt the various methods such as the method described in patent documentation 1.
That is, above-mentioned template and the pattern data (being the torque data as servo data in the present embodiment) respectively with corresponding are interrelated.In the past, although there is the problem being difficult to set identical initial conditions, according to the present embodiment, be identical be used as condition with identical initial conditions (template and pattern), therefore have advantage comparing on this servo data.
Like this, in the present embodiment, as mentioned above, because the template in position command and pattern are identical, therefore initial conditions is identical.By the output of robot more at that time, namely by comparing servo data, the aging of robot can be confirmed, and can fault diagnosis be carried out.Specifically, although the aging of robot such as causes the arrival term of life, the deterioration that produce motor, bearing etc., but can by comparing the output (servo data) of robot body relative to identical initial conditions (position command), and investigate output difference, thus judge the term of life, replacing period etc. of each part.
(action of the entirety of robot information handling system)
Figure 10 is the process flow diagram of the action of entirety in order to the robot information handling system involved by present embodiment is described.Next, centered by the action of data processing equipment 40, and in conjunction with the process flow diagram of Figure 10, the action of entirety of the robot information handling system with said structure is described.
In the robot information handling system 10 of present embodiment, instruction (data of machine entered people) and servo data (data from robot exports) can be extracted from the industrial robot common production action.In the common production action of operation robot (step ST11), controller 30 exports as the instruction of the information relevant to the drived control of servo system drive source and servo data (step ST12) to data processing equipment 40 as data S30.Thus, data processing equipment 40 collects the instruction and servo data that are transfused in units of such as about 20msec, and the data of collection is stored in storage part 44 (step ST13, ST14).
Such as, as the industrial robot of object as such as Fig. 6 and as shown in Figure 7, for storage box 60 position relationship in opposite directions with holding semiconductor wafer, the hand 25 of robot carries out semiconductor crystal wafer to move in storage box 60 or the action taken out of from storage box.
Although the robot carrying out action shown in Fig. 7 processes each layer in storage box as shown in Figure 6, due to be configured with the position of one or more storage box, each storage box and robot position relationship etc. and make the position of the hand of robot slightly different.Specifically the value of the longitudinal axis of the pattern shown in Fig. 4 etc. is slightly different.Therefore, the pattern (embodiment, correlation is 0.999) that the extracting data of a large amount of (huge) that the data processing equipment 40 involved by present embodiment can accurately, easily produce from temporally sequence is in short time roughly the same from slightly different patterns.
Specifically, data processing equipment 40 generates template from the mass data that temporally sequence produces, and this template is the pattern (step ST15) becoming reference or benchmark.Next, data processing equipment 40 extracts similarity (such as correlation) (the step ST16) as obtaining with the comparison pattern of the roughly the same pattern of template generated from above-mentioned mass data.Namely, recognition template whether consistent with the comparison pattern of selection (step ST17), time consistent ("Yes" of step ST17), extract the comparison pattern (step ST18) consistent with this, and be stored in storage part 44 (step ST19).In step ST17, be identified as the comparison pattern of template and selection inconsistent time ("No" of step ST17), return the process of step ST16.In addition, the generation of template, pattern (comparison pattern) is implemented repeatedly.
Then, such as compare extraction comparison pattern and as template pattern between servo data (step ST20), carry out aging confirmation (step ST21).
As described above, the robot information handling system 10 comprising the data processing equipment of present embodiment has following structure.Robot information handling system 10 has can the controller 30 of outgoing position data (or speed data) and servo data (torque, deviation etc.) all the time, and data processing equipment 40 is collected, preserve and store the mass data that temporally sequence produces.Data processing equipment 40 by desired pattern, such as, will confirm aging pattern as template (reference data) for position command (or speed command).Data processing equipment 40 obtains the similarity with the position command stored (or speed command) in the past by template, and extracts consistent (similarity is high) pattern.Then, although be random, confirm that (check) is aging by the servo data compared between the pattern of extraction and the template of pattern.
Consequently, by the data processing equipment 40 of present embodiment, subsidiary torque data, deviation data etc. can be analyzed in the past and between current arbitrary pattern.When industrial robot, production run substantially all carries out perseveration, thus necessarily there is identical action with the passing of time.Such as, to the action that each framves of multiple mounting framves 61 of storage box 60 or multiple storage box are implemented as shown in figure 5 and figure 7.Further, in Fig. 5 and Fig. 7, be the action that the semiconductor crystal wafer 70 etc. being stored in storage box 60 is taken out of, but also have the action with hand 25, semiconductor crystal wafer 70 etc. being moved into the frame of the defined of the mounting frame 61 of storage box 60 conversely.
The template of velocity mode can carry out the extraction of individual part, the template of mode position can carry out the extraction of combinative movement, and, if use multiple templates of each axle, then can also carry out the comparison between longer pattern, and the comparison of arbitrary pattern can be carried out.Specifically, in the present embodiment, the action of industrial robot 50 is the actions shown in Fig. 5 and Fig. 7, and therefore, template and the pattern as pattern adopt the position data as position command.That is, be Fig. 4 and simple trapezoidal shape as shown in Figure 9.During the action more complicated of industrial robot 50, exist when representing with mode position and be not simple trapezoidal shape but the situation of the shape of complexity.In this case, also only can use an action, namely also only can use the action of the step ST1 of Fig. 5 (Fig. 7), be used as the speed data of speed command.That is, when considering the action of step ST1 with speed command, be make speed increase in an accelerated manner from the state that stopped to the action becoming regulation, then, being constant speed, next reducing speed to make action stop, make speed vanishing, and stop in assigned position.If this action Negotiation speed, position, relation between the time are carried out pictorialization, then can represent with the simple trapezoidal shape identical with Fig. 4 with Fig. 9.
Further, if utilize this function, then pattern when producing mistake also can be become template, and trace back to over and extract identical pattern, comparing the change of servo data, thus can misjudgment reason whether relevant to servo system.
Further, according to notebook data treating apparatus, pattern when mistake can be produced becomes template, and confirms whether servo data changes from the same action in past.Consequently, can misjudgment reason whether be caused by the servo system with mechanism part.Such as, when the mistake of servo system produces, by pattern templating when producing wrong from the position be stored at that time (or speed) data.Review over from this template and extract identical pattern.With the servo data of the pattern producing mistake, the servo data of the identical pattern of relatively passing by checks whether error reason is that servo causes.Although need the situation in assurance cycle more about industrial robot, if apply this function, automatically the cycle can be derived.The pattern in cycle is logged in as template.Retrieve pattern in the past and extract identical pattern, thus calculating the cycle.
As described above, data processing equipment 40 according to the present embodiment, collects and the data processing equipment 40 processing the mass data that temporally sequence produces has the storage part 44 of mass data and the schema extraction portion 45 of the template extracted respectively from the data being stored in storage part 44 as the pattern of benchmark and the comparison pattern similar to template that store input.And, schema extraction portion 44 has the template generation portion 451 generating template from the data stored, the similarity obtaining section 453 of the similarity obtaining the comparison pattern selected and template from the data stored and comparison pattern extraction unit 454, and this comparison pattern extraction unit 454 is extracted based on the similarity obtained by similarity obtaining section 453 and is stored as the comparison pattern of the pattern similar to template.Consequently, if data processing equipment according to the present embodiment, then easily can extract the data wishing to compare from the mass data that temporally sequence produces accurately.
Further, based on the correlation between template and comparison pattern, the similarity identification part 453 of the data processing equipment 40 of present embodiment extraction unit 45 in mode identifies that the mode of similarity is formed.By having this structure, if data processing equipment according to the present embodiment 40, then owing to using correlation to try to achieve similarity, therefore formula is not the complicated formula of high order, therefore, it is possible to shorten the processing time, also can be the computing machine without high throughput, there is the advantage of good versatility.
And, in the data processing equipment 40 of present embodiment, data that temporally sequence produces comprise the output data of the instruction that inputs to carry out action control to industrial robot 50 and the industrial robot 50 from the action based on this instruction, and template and comparison pattern generate from director data.In addition, schema extraction portion 45 is formed to compare the mode of the data corresponding to template and comparison pattern.Servo data comprises torque data, and schema extraction portion 45 compares the torque data corresponding to template and comparison pattern.
By having this structure, data processing equipment 40 according to the present embodiment, although exist in the prior art and be difficult to the problem setting identical initial conditions, but according to this structure, by identical initial conditions (template and pattern) be identical premised on, therefore have advantage when comparing this servo data.In other words, according to the data processing equipment 40 with this structure, following effect can be obtained.
(1) in order to observe the aging of servo system as prior art, that prepare as basic data obtains action, different from common production action actions is not needed.
(2) be the torque data of production action itself due to what check, therefore, it is possible to Direct Analysis real data.
(3) following content can be realized by the template function of this arbitrary specific action.
When producing suddenly mistake aborning, using pattern at that time as template also retrieval data in the past, and can compare with torque data time normal, can judge whether servo system is main cause.
When can confirm there is the symptom of sounding such in specific action, pattern at that time is also extracted pattern in the past as template, compares with the torque data in past, can decision mechanism system whether change
(4) by logging in the template representing circulation, the aging of cycle can be held.
In addition, the above method specifically illustrated can be formed as the program corresponding with said sequence, and computer-implemented by CPU etc.In the present embodiment, although have selected action when not producing aging as template, be not limited thereto.Such as, as shown in the embodiment, also hope can be confirmed that (check) is stored in the group of the continuous data of the position command (or speed command) of the axle of the change of servo data in the past of storage part 44 as template in template generation portion 451.Further, such a program can be configured to by storage mediums such as semiconductor memory, disk, CD, floppy disks (registered trademark) and is provided with the computer access of this storage medium and implements said procedure.
Symbol description
10 robot information handling systems
20 robot bodies
21
30 controllers
40 data processing equipments
41 control parts
42 input parts
43 efferents
44 storage parts
45 schema extraction portions
451 template generation portions
452 compare (action) schema creation portion
453 similarity obtaining sections (similarity operational part)
454 comparison pattern extraction units (comparing section).
Claims (8)
1. a data processing equipment, collect and process the mass data that temporally sequence produces, it is characterized in that having:
Storage part, the described data that described storage part storage is transfused to; And
Schema extraction portion, described schema extraction portion extracts template as the pattern of benchmark and the comparison pattern similar to described template respectively from the described data being stored in described storage part,
Described schema extraction portion comprises:
Template generation portion, described template generation portion generates described template from described data;
Similarity obtaining section, described similarity obtaining section obtains the similarity of comparison pattern and the described template selected from described data; And
Comparison pattern extraction unit, described comparison pattern extraction unit is extracted based on the similarity obtained by described similarity obtaining section and is stored as the comparison pattern of the pattern similar to described template.
2. data processing equipment according to claim 1, is characterized in that,
The described similarity obtaining section in described schema extraction portion identifies described similarity based on the correlation between described template and described comparison pattern.
3. data processing equipment according to claim 1 and 2, is characterized in that,
The described data that temporally sequence produces comprise:
Director data, described director data is transfused to carry out action control to industrial robot; And
Servo data, described servo number is exported by the described industrial robot from the action based on described director data,
Described template and described comparison pattern generate from described director data, the more described template in described schema extraction portion and the described data corresponding to described comparison pattern.
4. data processing equipment according to claim 3, is characterized in that,
Described servo data comprises torque data, the more described template in described schema extraction portion and the described torque data corresponding to described comparison pattern.
5. a data processing method, collect and process the mass data that temporally sequence produces, it is characterized in that having:
The described data of input, in described storing step, are stored in storage part by storing step; And
Template extraction step, in described template extraction step, extracts the template as the pattern of benchmark and the comparison pattern similar to described template respectively from the described data being stored in described storage part,
Described schema extraction step comprises:
Template generation step, in described template generation step, generates described template from described data;
Similarity obtains step, obtains in step in described similarity, obtains the similarity of comparison pattern and the described template selected from described data; And
Comparison pattern extraction step, in described comparison pattern extraction step, is extracted according to being obtained the similarity that step obtains by described similarity and is stored as the comparison pattern of the pattern similar to described template.
6. data processing method according to claim 5, is characterized in that,
Obtain in step in described similarity, identify described similarity based on the correlation between described template and described comparison pattern.
7. the data processing method according to claim 5 or 6, is characterized in that,
The described data that temporally sequence produces comprise:
Director data, described director data is transfused to carry out action control to industrial robot; And
Servo data, described servo number is exported by the described industrial robot from the action based on described director data,
Described template and described comparison pattern generate from described director data, in described schema extraction step, and more described template and the described data corresponding to described comparison pattern.
8. data processing method according to claim 7, is characterized in that,
Described servo data comprises torque data, in described schema extraction step, and more described template and the described torque data corresponding to described comparison pattern.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2013113363A JP6211802B2 (en) | 2013-05-29 | 2013-05-29 | Data processing apparatus and data processing method |
| JP2013-113363 | 2013-05-29 | ||
| PCT/JP2014/062647 WO2014192525A1 (en) | 2013-05-29 | 2014-05-13 | Data processing device and data processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105247430A true CN105247430A (en) | 2016-01-13 |
| CN105247430B CN105247430B (en) | 2017-09-05 |
Family
ID=51988563
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201480030655.1A Active CN105247430B (en) | 2013-05-29 | 2014-05-13 | Data processing equipment and data processing method |
Country Status (3)
| Country | Link |
|---|---|
| JP (1) | JP6211802B2 (en) |
| CN (1) | CN105247430B (en) |
| WO (1) | WO2014192525A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110308669A (en) * | 2019-07-27 | 2019-10-08 | 南京市晨枭软件技术有限公司 | A kind of modular robot selfreparing analogue system and method |
| CN113211426A (en) * | 2020-12-02 | 2021-08-06 | 格创东智(深圳)科技有限公司 | Robot fault diagnosis method and device, computer equipment and storage medium |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6563210B2 (en) * | 2015-02-17 | 2019-08-21 | 日本電産サンキョー株式会社 | Operation history management system, management apparatus, and operation history management method |
| JP2016179527A (en) | 2015-03-24 | 2016-10-13 | ファナック株式会社 | Robot controller having function of diagnosing abnormality of robot |
| JP7192251B2 (en) * | 2018-05-29 | 2022-12-20 | 富士通株式会社 | Information processing device, robot operation program generation assistance method, and robot operation program generation assistance program |
| JP2021135780A (en) * | 2020-02-27 | 2021-09-13 | 株式会社東芝 | State monitor, method, and program |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4503507A (en) * | 1981-06-17 | 1985-03-05 | Hitachi, Ltd. | Robot control data processing apparatus |
| JP2003058248A (en) * | 2001-08-10 | 2003-02-28 | Mitsubishi Electric Corp | Operation supporting device |
| JP2006123157A (en) * | 2004-09-28 | 2006-05-18 | Nidec Sankyo Corp | Robot teaching program for teaching robot, cassette used therefor, position measuring apparatus and robot operation method using these |
| US20080308533A1 (en) * | 2007-06-14 | 2008-12-18 | Fanuc Ltd | Spot welding system and method for adjusting welding-gun closing speed |
| US20120136823A1 (en) * | 2010-05-17 | 2012-05-31 | Tomohiro Konuma | Audio classification device, method, program and integrated circuit |
| CN102528809A (en) * | 2010-11-05 | 2012-07-04 | 佳能株式会社 | Robot controlling device |
| CN102756374A (en) * | 2011-04-28 | 2012-10-31 | 精工爱普生株式会社 | Robot controller, simple installation-type robot, and method of controlling simple installation-type robot |
| CN103003803A (en) * | 2010-08-11 | 2013-03-27 | 株式会社日立制作所 | Time-series data processing apparatus and method thereof |
| JP2013061853A (en) * | 2011-09-14 | 2013-04-04 | Toshiba Corp | Process monitoring/diagnosis support device |
| CN103049459A (en) * | 2011-10-17 | 2013-04-17 | 天津市亚安科技股份有限公司 | Feature recognition based quick video retrieval method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3221194B2 (en) * | 1993-12-02 | 2001-10-22 | 日産自動車株式会社 | Operational waveform diagnostic device for industrial robots |
| JP2002268728A (en) * | 2001-03-08 | 2002-09-20 | Yamatake Sangyo Systems Co Ltd | Synchronization diagnosing and monitoring system and its device and its program |
| JP4739556B2 (en) * | 2001-03-27 | 2011-08-03 | 株式会社安川電機 | Remote adjustment and abnormality judgment device for control target |
| JP2007172150A (en) * | 2005-12-20 | 2007-07-05 | Honda Motor Co Ltd | Method for diagnosing failure of industrial machinery |
| JP2009169556A (en) * | 2008-01-12 | 2009-07-30 | Kanto Auto Works Ltd | Tracing system |
-
2013
- 2013-05-29 JP JP2013113363A patent/JP6211802B2/en active Active
-
2014
- 2014-05-13 CN CN201480030655.1A patent/CN105247430B/en active Active
- 2014-05-13 WO PCT/JP2014/062647 patent/WO2014192525A1/en active Application Filing
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4503507A (en) * | 1981-06-17 | 1985-03-05 | Hitachi, Ltd. | Robot control data processing apparatus |
| JP2003058248A (en) * | 2001-08-10 | 2003-02-28 | Mitsubishi Electric Corp | Operation supporting device |
| JP2006123157A (en) * | 2004-09-28 | 2006-05-18 | Nidec Sankyo Corp | Robot teaching program for teaching robot, cassette used therefor, position measuring apparatus and robot operation method using these |
| US20080308533A1 (en) * | 2007-06-14 | 2008-12-18 | Fanuc Ltd | Spot welding system and method for adjusting welding-gun closing speed |
| US20120136823A1 (en) * | 2010-05-17 | 2012-05-31 | Tomohiro Konuma | Audio classification device, method, program and integrated circuit |
| CN103003803A (en) * | 2010-08-11 | 2013-03-27 | 株式会社日立制作所 | Time-series data processing apparatus and method thereof |
| CN102528809A (en) * | 2010-11-05 | 2012-07-04 | 佳能株式会社 | Robot controlling device |
| CN102756374A (en) * | 2011-04-28 | 2012-10-31 | 精工爱普生株式会社 | Robot controller, simple installation-type robot, and method of controlling simple installation-type robot |
| JP2013061853A (en) * | 2011-09-14 | 2013-04-04 | Toshiba Corp | Process monitoring/diagnosis support device |
| CN103049459A (en) * | 2011-10-17 | 2013-04-17 | 天津市亚安科技股份有限公司 | Feature recognition based quick video retrieval method |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110308669A (en) * | 2019-07-27 | 2019-10-08 | 南京市晨枭软件技术有限公司 | A kind of modular robot selfreparing analogue system and method |
| CN110308669B (en) * | 2019-07-27 | 2021-07-30 | 南京市晨枭软件技术有限公司 | Modular robot self-repairing simulation system and method |
| CN113211426A (en) * | 2020-12-02 | 2021-08-06 | 格创东智(深圳)科技有限公司 | Robot fault diagnosis method and device, computer equipment and storage medium |
| CN113211426B (en) * | 2020-12-02 | 2023-02-28 | 格创东智(深圳)科技有限公司 | Robot fault diagnosis method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105247430B (en) | 2017-09-05 |
| JP6211802B2 (en) | 2017-10-11 |
| WO2014192525A1 (en) | 2014-12-04 |
| JP2014232450A (en) | 2014-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105247430A (en) | Data processing device and data processing method | |
| Azamfar et al. | Intelligent ball screw fault diagnosis using a deep domain adaptation methodology | |
| US20200198128A1 (en) | Learning data confirmation support device, machine learning device, and failure predicting device | |
| US20060085091A9 (en) | Electronic fingerprints for machine control and production machines | |
| KR102494819B1 (en) | Dynamic modification of production plans responsive to manufacturing deviations | |
| CN109693146B (en) | Service life estimation device and machine learning device | |
| US20200101603A1 (en) | Controller and control system | |
| US9823648B2 (en) | Method and apparatus for an automated configuration of a monitoring function of an industrial robot | |
| US8185238B2 (en) | System for carrying out automatic drilling / riveting process in aeronautical assembly pieces | |
| CN104423317A (en) | Data acquisition device for acquiring cause of stoppage of drive axis and information relating thereto | |
| US20230161337A1 (en) | Diagnostic device, server, and diagnostic method | |
| Wöstmann et al. | A retrofit approach for predictive maintenance | |
| Vukelic et al. | A rule-based system for fixture design | |
| US20200254671A1 (en) | State determination device and state determination method | |
| DE102018106812A1 (en) | Interactive gripping system | |
| WO2020144939A1 (en) | Diagnosis system, diagnosis method, program, and recording medium | |
| JP7407784B2 (en) | Production system information gathering device, information gathering method and program | |
| CN111788042A (en) | Predictive analysis of robots | |
| JP2010513039A (en) | Control system | |
| Ciani | Future trends in I&M: Diagnostics, maintenance and condition monitoring for cyber-physical systems | |
| JP2019076993A (en) | Abnormality diagnosis device and abnormality diagnosis method for direct-acting shaft in machine tool, and machine tool | |
| EP4088886A1 (en) | Rapid change mechanism for complex end effectors | |
| WO2022234678A1 (en) | Machine learning device, classification device, and control device | |
| US20210311454A1 (en) | Maintenance support system, numerical controller, and control method of maintenance support system | |
| Espíndola et al. | Improving maintenance operations through application of mixed reality systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |