EP4200107A1 - Überwachung bei einem robotergestützten prozess - Google Patents
Überwachung bei einem robotergestützten prozessInfo
- Publication number
- EP4200107A1 EP4200107A1 EP21762031.9A EP21762031A EP4200107A1 EP 4200107 A1 EP4200107 A1 EP 4200107A1 EP 21762031 A EP21762031 A EP 21762031A EP 4200107 A1 EP4200107 A1 EP 4200107A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- model
- carried out
- data
- process data
- test
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 245
- 230000008569 process Effects 0.000 title claims abstract description 218
- 238000012544 monitoring process Methods 0.000 title claims abstract description 48
- 238000012360 testing method Methods 0.000 claims abstract description 95
- 238000012790 confirmation Methods 0.000 claims abstract description 12
- 238000011156 evaluation Methods 0.000 claims description 81
- 238000012795 verification Methods 0.000 claims description 27
- 238000012549 training Methods 0.000 claims description 25
- 238000012423 maintenance Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 4
- 238000012550 audit Methods 0.000 claims 2
- 238000013528 artificial neural network Methods 0.000 description 16
- 238000011161 development Methods 0.000 description 15
- 230000018109 developmental process Effects 0.000 description 15
- 125000004122 cyclic group Chemical group 0.000 description 6
- 238000002372 labelling Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000007257 malfunction Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
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
- B25J9/1602—Programme controls characterised by the control system, structure, architecture
- B25J9/161—Hardware, e.g. neural networks, fuzzy logic, interfaces, processor
-
- 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
- B25J9/1628—Programme controls characterised by the control loop
- B25J9/163—Programme controls characterised by the control loop learning, adaptive, model based, rule based expert control
-
- 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
- B25J9/1628—Programme controls characterised by the control loop
- B25J9/1653—Programme controls characterised by the control loop parameters identification, estimation, stiffness, accuracy, error analysis
-
- 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
- G05B23/0243—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 model based detection method, e.g. first-principles knowledge model
- G05B23/0254—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 model based detection method, e.g. first-principles knowledge model based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
-
- 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
- G05B2219/33—Director till display
- G05B2219/33296—ANN for diagnostic, monitoring
Definitions
- the present invention relates to a method for monitoring a robot-supported process and a system and a computer program product for carrying out the method.
- Robot-supported processes i.e. processes that are carried out with the help of one or more robots, are often highly automated. For this reason in particular, but also especially in the case of human-robot collaborations, monitoring is particularly advantageous.
- Particularly advantageous applications are predictive maintenance of the robot(s) and monitoring for errors in the process and/or errors in process products.
- the object of the present invention is to improve robot-supported processes.
- the model-based evaluation (for the respective cycle) is carried out using the machine-learned model based on this recorded process data (for the respective cycle).
- the process is carried out in one embodiment using one or more robots.
- the or one or more of the robots has/have in one embodiment (each) a multi-axis robot arm, in one embodiment at least five, in particular at least six, in one embodiment at least seven-axis robot arm and/or a stationary or mobile base.
- one or more of the robots work through the same program, in one embodiment (each) travel the same path(s) and/ or (respectively) carries out the same activity(s), which may in particular include component transport and/or component processing.
- the present invention is particularly suitable for such processes, in particular industrial processes, due to the high degree of automation.
- the evaluation comprises a two-stage or multi-stage classification, in particular a classification into (the classes) ⁇ "fault-free” and “fault/anomaly” ⁇ or ⁇ "fault-free", "first fault/anomaly”, “second (r) error/anomaly” and, if applicable, one or more other errors or anomalies ⁇ .
- the assessment may also include an anomalies score, anticipated maintenance required, or other.
- step (a.2) carried out model-based evaluation meets a verification criterion, in a Execution (additionally) depending on an external confirmation, in one execution only if both the evaluation meets the verification criterion and the external confirmation is available or positive: (b.1) a test evaluation carried out using a test instance, in a
- step (a.1) Execution based on the process data recorded in step (a.1), in another embodiment independent of this, in a further development based on the process data used in step (a.2), in another further development independent of in step (a.2) process data used, in particular (instead) on the basis of other process data recorded in step (a.1) or another part of the process data recorded in step (a.1) or on the basis not in step (a.1), in one embodiment instead during a separate test drive or the like, recorded data; and
- step (b.2) further trains the machine-learned model on the basis of this test evaluation and in one embodiment also the (associated) process data recorded in step (a.1) and/or used in step (a.2).
- model-based assessment carried out therein or in its step (a.2) meets a verification criterion, in one embodiment depending on an external confirmation, in one embodiment only if both the evaluation meets the verification criterion and the external confirmation is available or positive:
- a test evaluation is carried out using a test instance, in an execution based on the process data recorded in this cycle or its step (a.1) or independently, in a further development based on the in this cycle or its Process data used in step (a.2), in another development independently of the process data used in step (a.2), in particular (instead) on the basis of other process data recorded in step (a.1) or another part of the process data recorded in step ( a.1) Process data recorded or on the basis of data not recorded in step (a.1), in one embodiment instead during a separate test drive or the like; and
- (b.2) further trains the model on the basis of this test evaluation and, in one embodiment, also of the process data recorded in the (corresponding) cycle and used in one embodiment in its step (a.2).
- (c.1) process data are recorded;
- step (b.2) By further training the model on the basis of process data according to step (b.2), the performance of the model and thus the monitoring using the model in step (c.3) can be improved.
- the use of process data from the first process, which is recorded in step (a.1) can improve subsequent monitoring in step (c.3), since the model is further trained in a process-specific manner.
- the use of process data from the second process, which is recorded in step (a.1) can advantageously enable (further) training without completely processing the first process for the purpose of (further) training.
- step (b.2) the model is further trained on the basis of test evaluations, in particular a so-called "label", by a test instance, for example a human who checks the functional status of the robot or manually evaluates process products, this can be machine learning be significantly improved.
- the test assessment in step (b.1) includes labeling with the aid of, in particular by the test instance.
- the present invention therefore proposes only triggering such a test assessment if it is (probably particularly) sensible, particularly necessary and/or particularly efficient or effective, which is decided by the verification criterion.
- labeling can only be carried out (more) specifically, in particular only for part of the cycles.
- the performance of the machine-learned model or the (evaluation for) monitoring using the model can be improved and at the same time the effort for labeling or further training can be reduced and thus productivity can be increased in one embodiment.
- step (b.1) By performing the test evaluation in step (b.1) using the test instance on the basis of the process data used in step (a.2), the effort for data acquisition and/or management, in particular storage, can be reduced will.
- step (b.1) the test evaluation using the test instance is independent of the process data used in step (a.2), in particular on the basis of other process data recorded in step (a.1) or another part of the process data recorded in step (a.1) or instead on the basis of other data recorded in one embodiment during a separate test drive or the like, in one embodiment the quality of the test evaluation can be improved and/or diversity can be used .
- a human being as a test instance can particularly advantageously use image, video and/or audio data for the test evaluation and the machine-learned model for the model-based evaluation of kinematic and/or dynamic data of the robot(s) or in step ( a.2) kinematic and/or dynamic robot data are used for the model-based evaluation and in step (b.1) a test evaluation using a human test instance based on captured images, audio and/or video recordings or the like independently of this robot data be performed.
- an advantageous, in particular (more precise) test evaluation can be carried out, for example on the basis of data recorded during a separate test drive.
- the reviewer includes one or more people.
- particularly complex processes can be labeled advantageously, particularly reliably and/or quickly and/or without using the recorded process data in step (b.2).
- the test instance has at least one further machine-learned model, with an evaluation using this further machine-learned model being more complex and/or more reliable in one embodiment than using the model further trained in step (b.2).
- process data can also be used for labeling which, for example, are difficult(er) to interpret for a human being or only with great effort.
- the test instance determines one or more parameters, in one embodiment one or more parameters of the robot(s) with which the respective process is or has been carried out, and/or or one or more parameters of the respective process and/or respective process product, for example a coefficient of friction of a robot gear or the like, the test evaluation being carried out in one embodiment on the basis of the determined parameter(s).
- the test evaluation includes a test or
- test evaluation is carried out using the test instance on the basis of the data generated during such a test or
- the process data and/or the data used in the test evaluation using the test instance have data, in particular time profiles, from one or more robot(s) with which the process is/are carried out.
- this data includes kinematic data, in particular poses and/or pose changes and/or rates of change, of the robot(s), in one embodiment axis positions, velocities and/or accelerations and/or positions and/or orientations at least one robot-fixed reference such as the TCP or the like, and/or their speeds and/or accelerations, in particular time profiles thereof.
- the process data and/or data used in the test evaluation using the test instance include dynamic data, in particular forces, torques, energies, power or the like, of the robot(s) in one embodiment
- drive forces, torques, energies and/or power, in particular voltages and/or currents, and/or external forces and/or torques which in one embodiment are determined using corresponding sensors on the robot, in particular time profiles thereof.
- Such (process) data is particularly suitable for monitoring and especially for predictive maintenance using the machine-learned model.
- the process data and / or the data used in the test evaluation using the test instance data, in one embodiment image data, one or more process products (s), the / in the respective process, in particular the respective Cycle, handled, in particular transported and/or processed, is/are handled, in particular transported and/or processed, on, in particular with the help of the robot(s).
- Audio and/or video data in particular recordings, of the respective process.
- Such (process) data is particularly suitable for monitoring errors in the process and/or errors in process products using the machine-learned model.
- steps (b.1), (b.2) are not (or no longer) carried out for one or more cycles, although the model-based evaluation carried out therein meets the verification criterion if it is detected that a abort criterion, which can be set in a further development, is met.
- the termination criterion can be, for example, the achievement of a specified number of test evaluations that can be set in a further development and/or a specified quality and/or degree of convergence (extent) of the model that can be set in a further development, in particular falling below a predetermined, in a training adjustable, learning progress include.
- the further training of the model on the basis of specifically initiated test evaluations is ended at an advantageous point in time, thereby (further) increasing productivity.
- monitoring is (already) carried out in the process (run) or cycle whose process data are or have been recorded in step (a.1).
- evaluations carried out in step (a.2) for monitoring purposes can also be used for further training in step (b.2).
- the evaluation in one embodiment in the ongoing process it is possible to react advantageously to a stop of the corresponding monitoring, in one embodiment the evaluation can be validated or checked by the test evaluation or test instance and corrected if necessary. This is particularly expedient in the monitoring, in particular predictive maintenance, of at least one robot with which the first or second process is carried out.
- steps (a.1), (a.2) are carried out multiple times, in particular for multiple cycles of the robot-assisted first or second process, and subsequent to this multiple execution, in particular after these cycles, steps (b1.), (b.2) are carried out on the basis of the model-based assessments and, if applicable, process data collected during or in the cycles, in one execution one or more of the collected, in one execution stored, model-based assessments and optionally process data based on the verification criterion or from the collected model I-based assessments and optionally process data stored in an embodiment, those selected for which the assessment meets the verification criterion, and selected process (runs) for these or cycles, in one embodiment based on the respective stored process data, (in each case) the test evaluation or step (b.1) is carried out, which is then used for further training of the model in step (b.2).
- step (b.2) can advantageously be improved and/or the test evaluation in step (b.1) can be carried out more efficiently. This is particularly useful when monitoring for errors in the first process and/or process products of the first process,
- the present invention is particularly suitable for the monitoring of robots and in particular the predictive robot maintenance as well as the process monitoring for errors in the process and/or process products, but is not limited thereto.
- the monitoring carried out in step (c.3) and, if applicable, also the monitoring that is carried out on the basis of the evaluation(s) carried out in step (a.2) in one embodiment includes monitoring, in a further development predictive maintenance, one or more robot(s) with which the process is/are carried out, and/or monitoring for errors in the process and/or process products.
- the verification criterion is specified in such a way that the model-based evaluation satisfies the verification criterion if it outputs or evaluates a specific error or a predefined number of repetitions of the error, in particular the occurrence of the error in at least a predefined number of cycles .
- a specific error can in particular be a predetermined error type or group of error types, but also any include possible errors.
- the verification criterion is specified in one embodiment in such a way that the model-based evaluation meets the verification criterion if it evaluates or generates or outputs an error of a predetermined error type or group of error types or a predetermined number of repetitions thereof, in another embodiment such that the model-based evaluation satisfies the verification criterion if it evaluates or outputs any error or a predetermined number of repetitions thereof.
- An error within the meaning of the present invention can in particular be a current (already present) error or a predicted or emerging error.
- an alarm is issued if the model-based evaluation meets the verification criterion or evaluates a specific error. As a result, security can be increased in one embodiment.
- a model-based evaluation of one or more robots, with which the process is carried out is carried out using a machine-learned model based on process data that is executed in cycles, and on the basis of this evaluation a monitoring, in particular a predictive maintenance, of the robot(s), in particular a diagnosis of a condition of the robot(s) and/or a prognosis of a malfunction, in particular a failure, of the robot(s) , accomplished.
- a test evaluation is carried out using a test instance, with the the robots carry out a test run that deviates from the process or cycle and/or is/are dismantled.
- an alarm is issued and/or the test assessment is carried out as a function of the external confirmation, in particular only if the external confirmation is also present or positive.
- the machine-learned model is then based on this test evaluation, in particular on the basis of those recorded in step (a.1) and in step (b.1) process data labeled by the test instance, further trained (step (b.2) and then used in further monitoring (steps (c.1)-(c.3)).
- such complex(er)n test evaluations are only carried out or initiated when required or in the event of a (supposed) error. It can equally be provided to carry out the test evaluation for each error evaluated by the model I-based evaluation or only for certain, for example serious(er)n, particularly dangerous, and/or only reliably recognizable errors through test drives or disassembly .
- the verification criterion is specified in such a way that the model-based assessment using the model based on recorded process data meets the verification criterion and the model I-based assessment using the same model based on other recorded process data does not meet the verification criterion, with this process data being considered without loss of generality first or second process data are referred to and wherein an expected information gain with further training of the model on the basis of the first process data is greater than with further training of the model on the basis of the second process data.
- the information gain to be expected is or is determined by means of Uncertainty Sampling, Query-By-Committee, Expected Model Change, Expected Error Reduction, Variance Reduction, Density-Weighted Methods or the like, as described, for example, in Burr Settles: Active Learning Literature Survey, Computer Sciences Technical Report 1648, University of Wisconsin-Madison, 2009 are described with further evidence, with additional reference being made to this article and the further literature mentioned therein and the content of which is fully incorporated into the present disclosure.
- a model I-based evaluation can meet the verification criterion if its reliability falls below a predetermined minimum or if the change in the model or its expected error reduction expected during further training of the model on the basis of the corresponding process data exceeds a predetermined minimum or the like.
- a model-based assessment of the or the (product/s of the) process(es) or cycle/cycles, in an execution of the quality or quality, and in an execution based on this evaluation (already) a monitoring for errors in the Process and/or errors of process products are carried out, for example process products (assessed as) incorrect are sorted out and/or reworked and/or process parameters are adjusted or the like.
- a test evaluation is carried out using a test instance and preferably based on the process data recorded in the cycle, in one embodiment the process and/or the process product, in particular an image, in one embodiment an audio and/or video recording of the process and/or process product, by the test instance , in one embodiment a human and/or another machine learned model.
- the same process data is used for the model-based evaluation and the test evaluation, for example the further machine-learned model can use the same process data that was used in step (a.2).
- different process data are used for the model-based evaluation and the test evaluation, for example, the further machine-learned model or a human can use image, audio and/or video data as a test instance and the further trained in step (b.2).
- machine learned model use kinematic and/or dynamic robot data or the like in step (a.2) instead.
- the further training of a machine-learned model for or during monitoring of a cyclic robot-supported process for errors in the process and/or process products or the performance achieved can be improved.
- the model is additionally trained before step (c.1) on the basis of recorded process data, in a further development of the process data recorded in at least one of steps (a.1), without taking into account a test evaluation using the test instance , carried out in one execution at all, is.
- step (b.1) labeled process data and unlabeled process data (from at least one step (a.1)) are used together.
- the further training of a machine-learned model or its performance can be improved .
- step (a.1) and/or (c.1) is carried out during the respective process (run) or cycle, for example data, in particular time profiles, of at least one robot with which the process is carried out, and/ or audio and/or video data, in particular recordings, of the respective process are recorded and optionally stored, or carried out at the end of the cycle or afterwards, for example data, in particular image data, of at least one process product are recorded and optionally stored.
- step (a.2), (c.2) and/or (c.3) is/are carried out during the respective process (run) or cycle.
- robot monitoring in particular, in particular predictive maintenance, can react early to the onset of malfunctions.
- step (a.2), (c.2) and/or (c.2) is carried out at the end of the respective process (run) or cycle or thereafter, in one embodiment (only) after several process (runs). )s or cycles.
- training and/or monitoring for errors in the process and/or process products can be improved and/or the next cycle can already be started in parallel, thereby improving productivity.
- step (b.1) and/or step (b.2) is carried out after the respective process (run) or cycle, on the basis of which process data the test evaluation is carried out, in a further development immediately after the process (run ) or cycle, carried out in another development after several process (runs) or cycles.
- step (b.1) and/or step (b.2) is carried out after the respective process (run) or cycle, on the basis of which process data the test evaluation is carried out, in a further development immediately after the process (run ) or cycle, carried out in another development after several process (runs) or cycles.
- a system in particular hardware and/or software, in particular programming, is set up to carry out a method described here and/or has:
- step (b.1) a test evaluation using a test instance, in particular on the basis of this recorded process data or independently of it, in one further development on the basis of the process data used in step (a.2), in another further development independently of in step (a. 2) process data used, in particular (instead) on the basis of other process data recorded in step (a.1) or another part of the process data recorded in step (a.1) or on the basis not in step (a.1), in a Execution instead in a separate test drive or the like, to carry out collected data; and
- the system or its means(s) has: means for performing monitoring on the process whose process data is collected in step (a.1) based on the assessment performed in step (a.2); and or
- a means within the meaning of the present invention can be configured as hardware and/or software, in particular a processing unit, in particular a microprocessor unit (CPU), graphics card (GPU ) or the like, and/or one or more programs or program modules.
- the processing unit can be designed to process commands that are implemented as a program stored in a memory system, to acquire input signals from a data bus and/or to output output signals to a data bus.
- a storage system can have one or more, in particular different, storage media, in particular optical, magnetic, solid-state and/or other non-volatile media.
- a computer program product can have, in particular be a, in particular non-volatile, storage medium for storing a program or with a program stored thereon, with the execution of this program causing a system or a controller, in particular a computer, to carry out the method described here or one or more of its steps.
- one or more, in particular all, steps of the method are carried out fully or partially automatically, in particular by the system or its means.
- the system includes the robot.
- the machine learned model includes at least one artificial neural network.
- Such machine-learned models are particularly advantageous for the present invention, in particular because of their learning behavior and/or their precision, reliability and/or speed.
- Fig. 1 a system according to an embodiment of the present invention in a cyclic robotic process
- Fig. 2 shows a method for monitoring in the cyclic robotic process according to an embodiment of the present invention
- Fig. 3 a method for monitoring in the cyclic robotic process according to another embodiment of the present invention.
- FIG. 1 shows an example of a robot 10 with a robot arm 11 , which processes a workpiece 20 in one cycle with a tool 12 , which is transported on and off or further on a conveyor belt 21 and is recorded by a camera 30 after processing .
- a controller of the robot 10 is denoted by 13 .
- Fig. 2 shows a method for monitoring in the cyclic robotic process according to an embodiment of the present invention.
- a cycle of the robot-supported process outlined with reference to FIG. 1 is carried out, and process data, in the exemplary embodiment drive forces of the robot 10 or the like, are recorded in the process.
- a model-based evaluation of the robot for predictive maintenance carried out (Fig. 2: step S20)
- the model or artificial neural network 13.1 in the exemplary embodiment again purely as an example, based on the detected drive forces, the robot 10 as currently error-free, not in need of maintenance or as defective or in need of maintenance classified, ie the model-based assessment for monitoring using the machine-learned model results in, outputs, or evaluates an error.
- step S50 If all the intended cycles have been carried out (S40: "Y"), the process is ended (FIG. 2: step S50).
- a reference run is carried out with the robot, during which particularly significant drive forces or the like occur in the case of faulty robots.
- a test instance On the basis of this reference run or the data recorded in the process, a test instance carries out a test evaluation, for example in the form of another machine-learned model or a signal processing method, or labels the process data recorded in step S10, which led to the error message, accordingly, with these in execution can also distinguish between different errors of the robot.
- the machine-learned model or artificial neural network 13.1 is trained further on the basis of this test evaluation and the next cycle is then optionally run through. Without external confirmation (S70: "N"), an appropriate action is taken in the event of an alarm (S60), for example the robot is repaired (S90).
- the "F" branch is not taken in step S30 until a predetermined error evaluation repetition number has been reached.
- the artificial neural network 13.1 will initially trigger more false alarms on the basis of the process data recorded in the normal work process, in particular if the alarm threshold is initially selected to be low as a precaution.
- step S80 By using the reference runs by the test instance to distinguish the applicable alarms from the false alarms and by continuing to train the artificial neural network 13.1 on the basis of this labeling (step S80), the number of false alarms is reduced as the duration increases.
- FIG 3 shows a method for monitoring in the cyclic robotic process according to another embodiment of the present invention.
- the artificial neural network 13.1 can also use other data, in particular kinematic and/or dynamic robot data, in addition or as an alternative.
- step S31 The image of the machined workpiece from the camera 30 and the associated evaluation by the artificial neural network 13.1 (in the other embodiment the corresponding robot data) are each stored (FIG. 3: step S31).
- step S41 those cycles are selected from this collected process data and model-based assessments in which the reliability of the classification falls below a predetermined minimum, since further training of the artificial neural network 13.1 with these cycles or images can be expected to yield the greatest information.
- information gain, entropy or the like can also be used as a verification criterion or these can depend on it.
- step S51 These selected images are labeled by the human (step S51). If the artificial neural network 13.1 uses kinematic and/or dynamic robot data or the like in the above-mentioned modification, the model-based assessment and the test assessment are based on different process data, while the same process data can alternatively also be used.
- step S61 the artificial neural network 13.1 is trained further with the process data labeled in step S21 and additionally with the process data labeled in step S51.
- step S11 As long as no termination criterion has been met, for example the learning progress falls below a specified minimum or a specified number of repetitions has been reached (S71: “N”), the method jumps back to step S11.
- step S81 If the termination criterion is met (S71: "Y"), further training is ended and the artificial neural network 13.1 is used for quality monitoring in the process (step S81).
- the artificial neural network 13.1 is (further) trained in a particularly effective manner on the basis of the images that are particularly suitable for this purpose, and its performance is thereby significantly increased.
- unlabeled process data can also be used for further training of the artificial neural network 13.1.
- the exemplary implementations are only examples and are not intended to limit the scope, applications, or construction in any way. Rather, the person skilled in the art is given a guideline for the implementation of at least one exemplary embodiment by the preceding description, with various changes, in particular with regard to the function and arrangement of the described components, being able to be made without leaving the scope of protection, as it emerges from the claims and these equivalent combinations of features.
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Fuzzy Systems (AREA)
- Software Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Manipulator (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE102020210530.8A DE102020210530A1 (de) | 2020-08-19 | 2020-08-19 | Überwachung bei einem robotergestützten Prozess |
PCT/EP2021/072496 WO2022038039A1 (de) | 2020-08-19 | 2021-08-12 | Überwachung bei einem robotergestützten prozess |
Publications (1)
Publication Number | Publication Date |
---|---|
EP4200107A1 true EP4200107A1 (de) | 2023-06-28 |
Family
ID=77519122
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP21762031.9A Pending EP4200107A1 (de) | 2020-08-19 | 2021-08-12 | Überwachung bei einem robotergestützten prozess |
Country Status (6)
Country | Link |
---|---|
US (1) | US20230311326A1 (de) |
EP (1) | EP4200107A1 (de) |
KR (1) | KR20230051532A (de) |
CN (1) | CN116324657A (de) |
DE (1) | DE102020210530A1 (de) |
WO (1) | WO2022038039A1 (de) |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2998894B1 (de) * | 2005-07-11 | 2021-09-08 | Brooks Automation, Inc. | Intelligentes zustandsüberwachungs- und fehlerdiagnosesystem |
WO2019216975A1 (en) * | 2018-05-07 | 2019-11-14 | Strong Force Iot Portfolio 2016, Llc | Methods and systems for data collection, learning, and streaming of machine signals for analytics and maintenance using the industrial internet of things |
JP6869755B2 (ja) | 2017-03-07 | 2021-05-12 | オークマ株式会社 | 状態診断装置 |
JP6698603B2 (ja) | 2017-09-29 | 2020-05-27 | ファナック株式会社 | 数値制御システム、及び運転状態異常検知方法 |
US12099344B2 (en) * | 2017-11-03 | 2024-09-24 | R4N63R Capital Llc | Workspace actor selection systems and methods |
CN107817787B (zh) * | 2017-11-29 | 2020-04-28 | 华南理工大学 | 一种基于机器学习的智能产线机械手故障诊断方法 |
JP6810097B2 (ja) | 2018-05-21 | 2021-01-06 | ファナック株式会社 | 異常検出器 |
-
2020
- 2020-08-19 DE DE102020210530.8A patent/DE102020210530A1/de active Pending
-
2021
- 2021-08-12 KR KR1020237008587A patent/KR20230051532A/ko active Search and Examination
- 2021-08-12 WO PCT/EP2021/072496 patent/WO2022038039A1/de unknown
- 2021-08-12 EP EP21762031.9A patent/EP4200107A1/de active Pending
- 2021-08-12 US US18/042,048 patent/US20230311326A1/en active Pending
- 2021-08-12 CN CN202180071154.8A patent/CN116324657A/zh active Pending
Also Published As
Publication number | Publication date |
---|---|
DE102020210530A1 (de) | 2022-02-24 |
US20230311326A1 (en) | 2023-10-05 |
KR20230051532A (ko) | 2023-04-18 |
CN116324657A (zh) | 2023-06-23 |
WO2022038039A1 (de) | 2022-02-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
DE102013010406B4 (de) | Verfahren zur Diagnose von Bremsstörungen und Vorrichtung zur Diagnose von Bremsstörungen | |
DE102008021671A1 (de) | Verfahren und Vorrichtung zur Überwachung eines Manipulators | |
DE102010048369A1 (de) | Verfahren und Vorrichtung zur Sicherheitsüberwachung eines Manipulators | |
DE102016014989B4 (de) | Kollisionsüberwachung eines Roboters | |
DE102008015948A1 (de) | Verfahren und Vorrichtung zur Überwachung eines Manipulators | |
DE102018004437B4 (de) | Vorrichtung und Verfahren für die Unterstützung der Detektion einer Kabelbeschädigung in einem Robotermechanismus | |
WO2021156328A1 (de) | Betreiben einer applikation eines robotersystems | |
DE102017007909A1 (de) | Produktionssteuerung, ausgestattet mit einer Funktion zur Identifizierung der Ursache nach einem Betriebsstop einer Produktionseinrichtung, die Fertigungseinrichtungen umfasst | |
EP2355957A1 (de) | Verfahren und vorrichtung zum steuern eines manipulatorsystems | |
DE102014221798A1 (de) | Verfahren und Robotersystem zur Nutzung eines Industrieroboters für Prüfaufgaben | |
DE102012011108A1 (de) | Verfahren und Vorrichtung zum Steuern einer Manipulatoranordnung | |
WO2021219389A1 (de) | Detektieren von störungen in roboteranordnungen | |
WO2022038039A1 (de) | Überwachung bei einem robotergestützten prozess | |
DE102021126188B4 (de) | Überwachungsverfahren für einen Roboter und zugehörige Roboteranlage | |
EP3710207B1 (de) | Qualitätssicherung eines prozesses | |
EP2479630A1 (de) | Verfahren zur kollisionsfreien Überführung einer Anlage aus einem Scheinausmodus in einen Betriebsmodus | |
EP2864845B1 (de) | Automatisierte rekonfiguration eines ereignisdiskreten regelkreises | |
EP3955077A1 (de) | Steuerungseinrichtung einer betriebsanlage | |
DE102022205357A1 (de) | Verfahren zum Überprüfen einer bewegbaren Komponente einer technischen Vorrichtung | |
DE102021102206A1 (de) | Autonomer mobiler Roboter mit Fehlererkennung und Verfahren zum Betrieb eines autonomen mobilen Roboters mit Fehlererkennung | |
WO2021028209A1 (de) | Emergente risiken eines technischen systems | |
DE102015222168A1 (de) | Verfahren und computerprogramm zur korrektur von fehlern eines manipulatorsystems | |
DE102022205351A1 (de) | Verfahren zum Überprüfen einer technischen Vorrichtung | |
DE102021209867A1 (de) | Bewerten und/oder Steuern eines Roboterarbeitsprozesses | |
DE102022205349A1 (de) | Verfahren zum Erzeugen eines Sicherheitsprogramms für eine Recheneinheit einer technischen Vorrichtung |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
17P | Request for examination filed |
Effective date: 20230213 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
DAV | Request for validation of the european patent (deleted) | ||
DAX | Request for extension of the european patent (deleted) | ||
P01 | Opt-out of the competence of the unified patent court (upc) registered |
Free format text: CASE NUMBER: APP_33453/2024 Effective date: 20240605 |