Fault early warning method and device, nonvolatile storage medium and electronic equipment

Abstract

The invention discloses a fault early warning method and device, a nonvolatile storage medium and electronic equipment. The method comprises the steps that target waveform data generated when the robot works in a target time period and target log information corresponding to the target waveform data are acquired; according to the target waveform data, predicting future waveform data generated when the robot works in a future time period; performing feature extraction on the future waveform data and the target log information to obtain to-be-detected fault features; and according to the to-be-detected fault features, early warning is carried out on possible faults of the robot. The technical problem that the robot fault detection efficiency is low in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of robot fault diagnosis, and in particular to a fault early warning method, device, non-volatile storage medium and electronic equipment. Background Art

[0002] With the rapid development of industrial automation, robots are increasingly used in manufacturing, logistics, medical and other fields. The efficient and stable operation of robot systems is crucial to ensuring production efficiency and product quality. However, robots will inevitably fail during long-term operation. If these failures are not detected and handled in a timely manner, they may lead to production interruptions, increased costs and even safety accidents. There are two existing methods for robot fault detection: (1) Manual inspection method. Technical service personnel usually conduct a series of inspections based on the robot's fault performance, combined with system log analysis, control signal testing, software diagnostic tools, spare parts replacement methods, etc., and observe the robot's operating status. The technician can preliminarily determine the possible location of the fault. Manual inspection may lead to misjudgment or missed judgment, and often takes a long time, especially when the fault is not obvious or complex. (2) Traditional machine learning models, that is, they rely on a large amount of labeled data to train the model. When faced with new or unknown fault modes, this traditional machine learning model has limited generalization ability, and model updates and maintenance require additional labeled data and time, and the detection efficiency is also low.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a fault warning method, device, non-volatile storage medium and electronic device to at least solve the technical problem of low efficiency in robot fault detection in the prior art.

[0005] According to one aspect of an embodiment of the present invention, a fault warning method is provided, comprising: obtaining target waveform data generated when a robot works in a target time period, and target log information corresponding to the target waveform data; predicting future waveform data generated when the robot works in a future time period based on the target waveform data; performing feature extraction on the future waveform data and the target log information to obtain fault features to be detected; and issuing a warning for possible faults of the robot based on the fault features to be detected.

[0006] Optionally, based on the target waveform data, future waveform data generated when the robot is working in a future time period is predicted, including: inputting the target waveform data into a pre-trained long short-term memory network, and the long short-term memory network outputs the future waveform data, wherein the long short-term memory network is obtained by training the original long short-term memory network with a first training sample, and the first training sample is generated according to the historical waveform data of the sample robot in a historical time period.

[0007] Optionally, the first training sample is obtained in the following manner: obtaining historical waveform data of the sample robot within a historical period; determining a window length for dividing the training sample; sliding the window with the window length as a step size, and taking out waveform data corresponding to the window length in turn to obtain multiple groups of waveform data; and dividing the sample waveform data and the corresponding label waveform data in the multiple groups of waveform data to obtain the first training sample.

[0008] Optionally, based on the fault features to be detected, an early warning is given for possible faults in the robot, including: inputting the fault features to be detected into a pre-trained small sample learning model, and the small sample learning model outputs the probability of fault occurrence and the corresponding fault type, wherein the small sample learning model is obtained by training the original small sample learning model with a second training sample, and the second training sample includes sample fault features corresponding to each sample included in the first training sample, and sample fault types corresponding to each sample included in the first training sample, wherein the sample fault features corresponding to each sample included in the first training sample are generated by sample waveform data and corresponding log data.

[0009] Optionally, feature extraction is performed on future waveform data and target log information to obtain fault features to be detected, including: feature extraction is performed on future waveform data and target log information respectively to obtain waveform features and text features; and waveform features and text features are concatenated to obtain fault features to be detected.

[0010] Optionally, feature extraction is performed on the future waveform data and the target log information respectively to obtain waveform features and text features, including: preprocessing the future waveform data, extracting features from the preprocessed waveform data to obtain waveform features; using natural language processing technology to identify keywords in the target log information to obtain text features.

[0011] According to another aspect of an embodiment of the present invention, a fault warning device is provided, including: an acquisition module, which acquires target waveform data generated when the robot works in a target time period, and target log information corresponding to the target waveform data; a prediction module, which predicts future waveform data generated when the robot works in a future time period based on the target waveform data; a feature extraction module, which extracts features from the future waveform data and the target log information to obtain fault features to be detected; and a warning module, which warns of possible faults of the robot based on the fault features to be detected.

[0012] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided. The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing any one of the fault warning methods.

[0013] According to another aspect of an embodiment of the present invention, there is provided an electronic device, comprising: one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement any one of the fault warning methods.

[0014] According to yet another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, any one of the above-mentioned fault warning methods is implemented.

[0015] In an embodiment of the present invention, by acquiring target waveform data generated when the robot works in a target time period, and target log information corresponding to the target waveform data; predicting future waveform data generated when the robot works in a future time period based on the target waveform data; performing feature extraction on the future waveform data and the target log information to obtain fault features to be detected; and issuing early warnings for possible faults of the robot based on the fault features to be detected, the technical problem of low efficiency of robot fault detection in the prior art is solved, the purpose of efficiently detecting faults in the robot is achieved, and the technical effect of improving the efficiency of robot fault detection is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0017] Figure 1 is a flow chart of a fault warning method provided according to an embodiment of the present invention;

[0018] Figure 2 is a schematic diagram of a fault warning device provided according to an embodiment of the present invention;

[0019] Figure 3 is a schematic diagram of an electronic device provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0021] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] According to an embodiment of the present invention, a method embodiment of a fault warning is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0023] Figure 1 is a flow chart of a fault warning method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0024] Step S102, acquiring target waveform data generated when the robot works within a target period, and target log information corresponding to the target waveform data;

[0025] In this step, during the target period, the system uses sensors and log recording systems to capture the waveform data and log information generated by the robot during operation in real time. Waveform data usually refers to the continuously changing signals generated during the operation of the robot, such as the current waveform of the motor, the waveform of the joint position change, or the waveform output by the force sensor. These waveform data can reflect the dynamic behavior of each component of the robot when performing a specific task, and play a key role in fault detection and diagnosis. The target log information includes various records generated by the robot control system, such as error codes, warning messages, execution of operating instructions, and changes in system status. The log information provides a detailed operating status of the robot at the software level, which can help the system understand the context when the fault occurs. It is an important basis for diagnosing robot software faults or hardware faults and software interaction problems.

[0026] Step S104, predicting the future waveform data generated when the robot works in the future period according to the target waveform data;

[0027] In this step, the prediction of future waveform data is done by using deep learning technology and small sample learning models. By learning the target waveform data, the problems that the robot may encounter in the subsequent working period are predicted, so that measures can be taken in advance to reduce the losses caused by failures.

[0028] In an optional embodiment, based on the target waveform data, future waveform data generated when the robot is working in a future time period is predicted, including: inputting the target waveform data into a pre-trained long short-term memory network, and the long short-term memory network outputs the future waveform data, wherein the long short-term memory network is obtained by training the original long short-term memory network with a first training sample, and the first training sample is generated according to the historical waveform data of the sample robot in a historical time period.

[0029] Optionally, the original long short-term memory network can recognize and memorize the characteristics and change patterns of waveform data under normal operation and fault conditions by learning the first training samples generated by a large amount of historical waveform data. It can predict possible abnormal changes in future waveform data based on the currently input target waveform data.

[0030] In an optional embodiment, the first training sample is obtained in the following manner: obtaining historical waveform data of the sample robot within a historical period; determining a window length for dividing the training sample; sliding the window with the window length as a step size, and sequentially taking out waveform data corresponding to the window length to obtain multiple groups of waveform data; respectively dividing the sample waveform data and the corresponding label waveform data in the multiple groups of waveform data to obtain the first training sample.

[0031] Optionally, in order to train the model to recognize patterns in time series, a "window length" needs to be determined. For example, 30 days can be used as a window length. The system uses the above-determined window length to slide in the historical waveform data. Each moving window captures waveform data equal to the window length. In this way, a series of "multiple groups of waveform data" can be extracted from the historical data, and each group of data is a time series fragment. For each group of extracted waveform data, the system needs to divide it into "sample waveform data" and "label waveform data". The sample waveform data is the first part of the data in the window; the label waveform data is the part of the data immediately after the sample waveform data, which is used as a comparison standard for the network output during the training process. If an abnormal waveform (i.e., a fault feature) appears in the label waveform data, it is marked as a fault state, and the waveform data that operates normally is marked as a normal state. For example, the data collected from January 1 to January 31 can be used as sample waveform data, and the data collected from February 1 to March 2 can be used as label waveform data. In turn, the data collected from January 2 to February 1 can be used as new sample waveform data, and the data collected from February 2 to March 3 can be used as new label waveform data. Through the above steps, the system can generate multiple groups of "first training samples", each group of samples includes a sample waveform data and corresponding label waveform data, as well as a corresponding state label (normal / fault).

[0032] In an optional embodiment, based on the fault features to be detected, a warning is given for possible faults in the robot, including: inputting the fault features to be detected into a pre-trained small sample learning model, and the small sample learning model outputs the probability of fault occurrence and the corresponding fault type, wherein the small sample learning model is obtained by training the original small sample learning model through a second training sample, and the second training sample includes sample fault features corresponding to each sample included in the first training sample, and sample fault types corresponding to each sample included in the first training sample, wherein the sample fault features corresponding to each sample included in the first training sample are generated by sample waveform data and corresponding log data.

[0033] Optionally, the "fault features to be detected" obtained through the data acquisition module and the multi-dimensional data processing unit are input into a pre-trained small sample learning model. The model will output the corresponding "fault probability" and "fault type" based on the fault mode and classification rules it has learned.

[0034] Step S106, extracting features from the future waveform data and target log information to obtain features of the fault to be detected;

[0035] In this step, future waveform data refers to the waveform data that the system predicts the robot may generate in the next period of time. The fault feature extraction module uses a specific algorithm to analyze these future waveform data. By identifying abnormal patterns, trend changes, abnormal values, etc., the system can extract a series of features related to potential faults. For example, a sudden increase in motor current may indicate an overload, and irregular changes in joint position may mean a joint position failure. The target log information refers to the log records generated by the robot system in the same period as the target waveform data. The fault feature extraction module processes the target log information through text analysis technology, such as natural language processing based on a small sample learning model. It identifies keywords or phrases related to the fault, as well as the execution of operating instructions, so as to extract fault features at the software level. For example, the appearance of error codes, frequent failures of specific operating instructions, and degradation of system performance may all become objects of feature extraction.

[0036] In an optional embodiment, feature extraction is performed on future waveform data and target log information to obtain fault features to be detected, including: feature extraction is performed on future waveform data and target log information respectively to obtain waveform features and text features; waveform features and text features are concatenated to obtain fault features to be detected.

[0037] Optionally, future waveform data is the waveform data of the robot in the future period of time predicted by the long short-term memory network, which contains the predicted information of the robot's operating status. The system will use time series analysis technology to identify patterns in these data, such as abnormal waveforms (current overload, torque abnormality, etc.) and trend changes. These abnormal patterns and changing trends will be extracted as "waveform features". The target log information is the information extracted from the system log of the robot's current operation, which contains text data such as system status, operation instructions, error codes, etc. After extracting the waveform features and text features separately, the system needs to combine these features to form a comprehensive "fault feature to be detected".

[0038] In an optional embodiment, feature extraction is performed on the future waveform data and the target log information respectively to obtain waveform features and text features, including: preprocessing the future waveform data, extracting features from the preprocessed waveform data to obtain waveform features; using natural language processing technology to identify keywords in the target log information to obtain text features.

[0039] Optionally, for future waveform data, preprocessing may involve removing noise, smoothing signals, performing data normalization, and possible signal segmentation or time-frequency conversion, such as converting time domain signals into frequency domain features to better capture the periodicity or frequency characteristics of the waveform. After preprocessing, the system will perform feature extraction on the waveform data. This process may include calculating statistical features (such as mean, variance, peak), frequency features (features based on spectrum analysis, such as main frequency, harmonics, etc.), morphological features (such as waveform shape, duration, start and end points, etc.), and other possible timing features. Using timing analysis technology, the fault-related patterns and trends in the waveform data are automatically identified, and these patterns and trends are converted into waveform features, that is, a series of numerical values ​​or vectors, for subsequent fault diagnosis model input. In addition, using natural language processing technology, the system will identify keywords in the target log information. By identifying keywords, the system can extract text features, that is, specific words or phrases describing the fault information in the log. These features will also be converted into numerical representations for model processing.

[0040] Step S108, based on the fault characteristics to be detected, an early warning is issued for possible faults of the robot.

[0041] In this step, early warning is achieved by analyzing and matching these extracted fault features with known fault patterns in the historical database. The system uses the fault features previously extracted from the future waveform data and target log information. These features contain information such as possible abnormal patterns, trend changes, and abnormal values. By matching them with the fault code fields stored in the historical database, the system can identify known fault patterns similar to these features.

[0042] By the above-mentioned method, the target waveform data generated by the robot when working in the target time period and the target log information corresponding to the target waveform data are obtained; based on the target waveform data, the future waveform data generated by the robot when working in the future time period is predicted; the features of the future waveform data and the target log information are extracted to obtain the fault features to be detected; based on the fault features to be detected, the possible faults of the robot are warned, which solves the technical problem of low efficiency of robot fault detection in the prior art, achieves the purpose of efficiently detecting faults of the robot, and further realizes the technical effect of improving the efficiency of robot fault detection.

[0043] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation manner.

[0044] Step S1, acquiring target waveform data generated when the robot works within a target period, and target log information corresponding to the target waveform data;

[0045] In this step, during the target period, the system uses sensors and log recording systems to capture the waveform data and log information generated by the robot during operation in real time. Waveform data usually refers to the continuously changing signals generated during the operation of the robot, such as the current waveform of the motor, the waveform of the joint position change, or the waveform output by the force sensor. These waveform data can reflect the dynamic behavior of each component of the robot when performing a specific task, and play a key role in fault detection and diagnosis. The target log information includes various records generated by the robot control system, such as error codes, warning messages, execution of operating instructions, and changes in system status. The log information provides a detailed operating status of the robot at the software level, which can help the system understand the context when the fault occurs. It is an important basis for diagnosing robot software faults or hardware faults and software interaction problems.

[0046] Step S2, based on the target waveform data, predicting the future waveform data generated when the robot works in the future time period, including: inputting the target waveform data into a pre-trained long short-term memory network, and outputting the future waveform data from the long short-term memory network, wherein the long short-term memory network is obtained by training the original long short-term memory network through a first training sample, and the first training sample is generated according to the historical waveform data of the sample robot in the historical time period.

[0047] Optionally, the original long short-term memory network can recognize and memorize the characteristics and change patterns of waveform data under normal operation and fault conditions by learning the first training samples generated by a large amount of historical waveform data. It can predict possible abnormal changes in future waveform data based on the currently input target waveform data.

[0048] Step S21, the first training sample is obtained in the following manner: obtaining historical waveform data of the sample robot in a historical period; determining the window length used to divide the training sample; sliding the window with the window length as the step length, and taking out the waveform data corresponding to the window length in turn to obtain multiple groups of waveform data; respectively dividing the sample waveform data and the corresponding label waveform data in the multiple groups of waveform data to obtain the first training sample.

[0049] Step S22, based on the fault features to be detected, early warning of possible faults in the robot, including: inputting the fault features to be detected into a pre-trained small sample learning model, and the small sample learning model outputs the probability of fault occurrence and the corresponding fault type, wherein the small sample learning model is obtained by training the original small sample learning model through the second training sample, the second training sample includes the sample fault features corresponding to each sample included in the first training sample, and the sample fault types corresponding to each sample included in the first training sample, wherein the sample fault features corresponding to each sample included in the first training sample are generated by sample waveform data and corresponding log data.

[0050] Optionally, the "fault features to be detected" obtained through the data acquisition module and the multi-dimensional data processing unit are input into a pre-trained small sample learning model. The model will output the corresponding "fault probability" and "fault type" based on the fault mode and classification rules it has learned.

[0051] Optionally, in order to train the model to recognize patterns in time series, a "window length" needs to be determined. For example, 30 days can be used as a window length. The system uses the above-determined window length to slide in the historical waveform data. Each moving window captures waveform data equal to the window length. In this way, a series of "multiple groups of waveform data" can be extracted from the historical data, and each group of data is a time series fragment. For each group of extracted waveform data, the system needs to divide it into "sample waveform data" and "label waveform data". The sample waveform data is the first part of the data in the window; the label waveform data is the part of the data immediately after the sample waveform data, which is used as a comparison standard for the network output during the training process. If an abnormal waveform (i.e., a fault feature) appears in the label waveform data, it is marked as a fault state, and the waveform data that operates normally is marked as a normal state. For example, the data collected from January 1 to January 31 can be used as sample waveform data, and the data collected from February 1 to March 2 can be used as label waveform data. In turn, the data collected from January 2 to February 1 can be used as new sample waveform data, and the data collected from February 2 to March 3 can be used as new label waveform data. Through the above steps, the system can generate multiple groups of "first training samples", each group of samples includes a sample waveform data and corresponding label waveform data, as well as a corresponding state label (normal / fault).

[0052] Step S3, extracting features from the future waveform data and the target log information to obtain features of the fault to be detected, including: extracting features from the future waveform data and the target log information to obtain waveform features and text features; and concatenating the waveform features and the text features to obtain features of the fault to be detected.

[0053] Optionally, future waveform data is the waveform data of the robot in the future period of time predicted by the long short-term memory network, which contains the predicted information of the robot's operating status. The system will use time series analysis technology to identify patterns in these data, such as abnormal waveforms (current overload, torque abnormality, etc.) and trend changes. These abnormal patterns and changing trends will be extracted as "waveform features". The target log information is the information extracted from the system log of the robot's current operation, which contains text data such as system status, operation instructions, error codes, etc. After extracting the waveform features and text features separately, the system needs to combine these features to form a comprehensive "fault feature to be detected".

[0054] Step S31, extracting features from the future waveform data and the target log information respectively to obtain waveform features and text features, including: preprocessing the future waveform data, extracting features from the preprocessed waveform data to obtain waveform features; using natural language processing technology to identify keywords in the target log information to obtain text features.

[0055] Optionally, for future waveform data, preprocessing may involve removing noise, smoothing signals, performing data normalization, and possible signal segmentation or time-frequency conversion, such as converting time domain signals into frequency domain features to better capture the periodicity or frequency characteristics of the waveform. After preprocessing, the system will perform feature extraction on the waveform data. This process may include calculating statistical features (such as mean, variance, peak), frequency features (features based on spectrum analysis, such as main frequency, harmonics, etc.), morphological features (such as waveform shape, duration, start and end points, etc.), and other possible timing features. Using timing analysis technology, the fault-related patterns and trends in the waveform data are automatically identified, and these patterns and trends are converted into waveform features, that is, a series of numerical values ​​or vectors, for subsequent fault diagnosis model input. In addition, using natural language processing technology, the system will identify keywords in the target log information. By identifying keywords, the system can extract text features, that is, specific words or phrases describing the fault information in the log. These features will also be converted into numerical representations for model processing.

[0056] The above optional implementation achieves at least the following effects: solves the technical problem of low efficiency of robot fault detection in the prior art, achieves the purpose of efficient robot fault detection, and further achieves the technical effect of improving the efficiency of robot fault detection.

[0057] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0058] In this embodiment, a fault warning device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the terms "module" and "device" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0059] According to an embodiment of the present invention, a device embodiment for implementing a fault warning method is also provided. Figure 2is a schematic diagram of a fault warning device according to an embodiment of the present invention. Figure 2 As shown, the above-mentioned fault warning device includes an acquisition module 21, a prediction module 22, a feature extraction module 23, and a warning module 24. The device is described below.

[0060] An acquisition module 21 is used to acquire target waveform data generated when the robot works in a target period, and target log information corresponding to the target waveform data;

[0061] The prediction module 22 is connected to the acquisition module 21 and is used to predict the future waveform data generated when the robot works in the future period according to the target waveform data;

[0062] The feature extraction module 23 is connected to the prediction module 22 and is used to extract features from future waveform data and target log information to obtain features of the fault to be detected;

[0063] The early warning module 24 is connected to the feature extraction module 23 and is used to issue an early warning for possible faults of the robot according to the fault features to be detected.

[0064] In a fault warning device provided by an embodiment of the present invention, the target waveform data generated when the robot works in a target period of time and the target log information corresponding to the target waveform data are obtained by setting; based on the target waveform data, the future waveform data generated when the robot works in a future period of time is predicted; the future waveform data and the target log information are feature extracted to obtain the fault features to be detected; based on the fault features to be detected, the possible faults of the robot are warned. The technical problem of low efficiency of robot fault detection in the prior art is solved, the purpose of efficient fault detection of the robot is achieved, and the technical effect of improving the efficiency of robot fault detection is achieved.

[0065] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0066] It should be noted that the acquisition module 21, prediction module 22, feature extraction module 23, and early warning module 24 correspond to steps S102 to S108 in the embodiment, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments. It should be noted that the modules as part of the device can be run in a computer terminal.

[0067] It should be noted that the optional or preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0068] The above-mentioned fault warning device may also include a processor and a memory. The acquisition module 21, the prediction module 22, the feature extraction module 23, the warning module 24, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0069] The processor includes a kernel, which retrieves the corresponding program unit from the memory. There can be one or more kernels. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0070] An embodiment of the present invention provides a non-volatile storage medium on which a program is stored. When the program is executed by a processor, a fault early warning method is implemented.

[0071] like Figure 3 As shown, an embodiment of the present invention provides an electronic device, the electronic device 10 includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: the memory is used to store a computer program, wherein when the computer program is executed by the processor, the processor implements the above-mentioned fault warning method. The device in this article may be a server, a PC, etc.

[0072] The present invention also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that is initialized with the following method steps: computer instructions are executed by a processor to perform the above-mentioned fault warning method.

[0073] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0074] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0075] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0076] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0077] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0078] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0079] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0080] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0081] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0082] The above are only embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A fault warning method, characterized in that: include: Acquire target waveform data generated when the robot works within a target period of time, and target log information corresponding to the target waveform data; According to the target waveform data, predicting and obtaining future waveform data generated when the robot works in a future time period; Performing feature extraction on the future waveform data and the target log information to obtain features of a fault to be detected; According to the fault characteristics to be detected, an early warning is given for possible faults of the robot.

2. The method according to claim 1, characterized in that The step of predicting, based on the target waveform data, future waveform data generated when the robot works in a future period of time comprises: The target waveform data is input into a pre-trained long short-term memory network, and the long short-term memory network outputs the future waveform data, wherein the long short-term memory network is obtained by training the original long short-term memory network through a first training sample, and the first training sample is generated according to the historical waveform data of the sample robot in a historical period.

3. The method according to claim 2, characterized in that The first training sample is obtained by: Acquire historical waveform data of the sample robot in the historical period; Determine the window length used to divide the training samples; Sliding the window with the window length as the step length, sequentially taking out the waveform data corresponding to the window length, and obtaining multiple groups of waveform data; Sample waveform data and corresponding label waveform data are respectively divided from the multiple groups of waveform data to obtain the first training samples.

4. The method according to claim 2, characterized in that: The step of providing an early warning of a possible fault of the robot according to the fault feature to be detected includes: The fault feature to be detected is input into a pre-trained small sample learning model, and the small sample learning model outputs the probability of fault occurrence and the corresponding fault type, wherein the small sample learning model is obtained by training the original small sample learning model through a second training sample, and the second training sample includes sample fault features corresponding to each sample included in the first training sample, and sample fault types corresponding to each sample included in the first training sample, wherein the sample fault features corresponding to each sample included in the first training sample are generated by sample waveform data and corresponding log data.

5. The method according to any one of claims 1 to 4, characterized in that: The extracting features of the future waveform data and the target log information to obtain the fault features to be detected includes: Extracting features from the future waveform data and the target log information respectively to obtain waveform features and text features; The waveform feature and the text feature are concatenated to obtain the fault feature to be detected.

6. The method according to claim 5, characterized in that Feature extraction is performed on the future waveform data and the target log information respectively to obtain the waveform features and text features, including: Preprocessing the future waveform data, extracting features from the preprocessed waveform data, and obtaining the waveform features; Natural language processing technology is used to identify keywords in the target log information to obtain the text features.

7. A fault warning device, characterized in that: include: An acquisition module, used to acquire target waveform data generated when the robot works within a target period, and target log information corresponding to the target waveform data; A prediction module, used for predicting, based on the target waveform data, future waveform data generated when the robot is working in a future period; A feature extraction module, used for extracting features from the future waveform data and the target log information to obtain features of a fault to be detected; The early warning module is used to warn of possible faults of the robot according to the fault characteristics to be detected.

8. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the fault warning method described in any one of claims 1 to 6.

9. An electronic device, characterized in that: include: One or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the fault warning method described in any one of claims 1 to 6.

10. A computer program product comprising computer instructions, characterized in that The computer instructions are executed by the processor to implement the fault warning method described in any one of claims 1 to 6.

Images