A method, device and system for rapid fault location of robot components

Abstract

This invention relates to the field of robot fault technology, specifically to a method, device, and system for rapid fault localization of robot components. The method includes: monitoring abnormal data and timestamps of each connector of the robot to obtain the alignment frequency between the target connector and the other connectors; determining a set of associated connectors using the alignment frequency; determining the correlation between the target connector and the other connectors based on the associated connector set and the alignment frequency; obtaining the similarity between the target connector and the downstream connectors of the downstream connector set; calculating the fault probability of the target connector; and determining whether the target connector is truly faulty based on the fault probability. Through the technical solution of this invention, accurate fault localization of robot components is achieved, improving the accuracy of fault diagnosis.

Description

Technical Field

[0001] This invention relates to the field of robot fault technology, and specifically to a method, device and system for rapid fault location of robot components. Background Technology

[0002] Modern robots are evolving towards higher precision, higher speed, and higher integration. To meet the demands of complex motion control, robots typically employ a distributed control architecture. This involves connecting numerous sensors, servo drivers, controllers, and end effectors via a bus. While the current robot component connection architecture enhances system modularity, it also dramatically increases the number of physical connection points within the system. During long-term operation, factors such as high-frequency vibration, reciprocating motion traction, and temperature variations can cause connector failures. Unlike "hard failures" where components are completely destroyed, connector failures often manifest as intermittent poor contact or abnormal data transmission. Maintenance personnel must inspect every component of the robot, consuming significant time and effort, thus increasing maintenance costs. Therefore, when a robot component fails, rapid localization of the faulty component is crucial.

[0003] In existing technologies, traditional methods for rapid localization of robot components primarily rely on maintaining a stable power supply to these components under normal circumstances, using the stability of the power supply to each component to determine if a fault exists. Since connectors are critical components of the robot, analyzing the consistency between their output and input can help diagnose connector faults. However, relying on a single connector lacks information sharing between connectors. Furthermore, robots require varying torques and withstand different loads during command execution or actions. Therefore, relying solely on a single connector cannot effectively diagnose robot component faults, leading to incorrect localization and hindering subsequent robot maintenance and testing. Summary of the Invention

[0004] To address the technical problem that existing technologies cannot effectively diagnose robot component faults using a single connector, and that fault location is not precise enough, the present invention aims to provide a method, device, and system for rapid fault location of robot components. The specific technical solution adopted is as follows:

[0005] This invention provides a method for rapid fault location of robot components, the method comprising:

[0006] Monitor abnormal data and timestamps of each connector of the robot to obtain the alignment frequency between the target connector and the other connectors;

[0007] Each connector is sequentially selected as the target connector and its alignment frequency with the remaining connectors is obtained. The alignment frequency is used to determine the associated connector set. The association between the target connector and the remaining connectors is determined based on the associated connector set and the alignment frequency.

[0008] Based on the robot's circuit topology, the downstream connector set of the target connector is determined. Combining the correlation between the target connector and each connector in the downstream connector set, as well as the correlation set of each connector in the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0009] Based on the similarity of the downstream connectors and the correlation between the target connector and the other connectors, the failure probability of the target connector is calculated, and the failure probability is used to determine whether the target connector is a real failure.

[0010] Further, obtaining the alignment frequency between the target connector and the other connectors includes:

[0011] Determine the time difference in the transmission of abnormal data between the target connector and the other connectors in the robot;

[0012] If the transmission time difference is less than or equal to the preset time difference threshold, it is determined that the abnormal data between the connectors is a case of time alignment being completed.

[0013] If the transmission time difference is greater than the preset time difference threshold, it is determined that the abnormal data between the connectors is due to incomplete time alignment.

[0014] The ratio of the number of times time alignment is completed to the total number of times time alignment is used as the alignment frequency between the target connector and the other connectors.

[0015] Further, the step of sequentially identifying each connector as a target connector and obtaining its alignment frequency with the remaining connectors, and using the alignment frequency to determine the associated connector set, includes:

[0016] When the target connector is used as the connector to be analyzed, the alignment frequencies are determined, and multiple connectors with similar alignment frequencies are selected using an unsupervised learning algorithm.

[0017] By considering multiple connectors with similar alignment frequencies as associated connectors of the target connector, a set of associated connectors for the target connector is obtained.

[0018] Furthermore, determining the association between the target connector and the remaining connectors based on the associated connector set and alignment frequency includes:

[0019] The initial correlation between the target connector and the other connectors is obtained by using their respective alignment frequencies. The correction coefficient of the initial correlation is obtained by using the alignment frequencies of the target connector and the other connectors and the set of associated connectors. The initial correlation is then corrected using the correction coefficient to obtain the correlation between the target connector and the other connectors.

[0020] Furthermore, the step of obtaining the correction coefficient for the initial correlation using the alignment frequencies of the target connector and the other connectors, and the associated connector set, includes:

[0021] The set of abnormally consistent connectors is obtained by using the intersection of the target connector and the respective associated connector sets of the other connectors.

[0022] Determine the percentage of abnormally consistent connectors in the union of the abnormally consistent connector set with respect to the target connector and the respective associated connector sets of the other connectors.

[0023] The correction coefficient for the initial correlation is obtained by using the proportion of abnormally consistent connectors and the difference in alignment frequency between the target connector and the other connectors.

[0024] Furthermore, the step of combining the correlation between the target connector and each connector in the downstream connector set, and the correlation set of each connector in the downstream connector set itself, to obtain the similarity between the target connector and the downstream connectors in the downstream connector set includes:

[0025] By utilizing the correlations between the target connector and the set of downstream connectors in its line topology, the corresponding correlation standard deviation is obtained.

[0026] The initial similarity between the target connector and the downstream connector set is obtained by using the correlation standard deviation between the target connector and the downstream connector set.

[0027] By utilizing the initial similarity and the intersection between each associated connector set corresponding to the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0028] Furthermore, the step of obtaining the downstream connector similarity between the target connector and the downstream connector set by utilizing the intersection between the initial similarity and the corresponding associated connector sets of the downstream connector set includes:

[0029] Determine the percentage of the number of connectors in the intersection of each associated connector set corresponding to the downstream connector set, relative to the total number of connectors in the downstream connector set.

[0030] The corrected similarity between the target connector and the downstream connector set is obtained by using the initial similarity and the proportion of connectors in the number of connectors.

[0031] By using the corrected similarity and the mean correlation between the target connector and the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0032] Further, the step of calculating the failure probability of the target connector based on the similarity of the downstream connectors and the correlation between the target connector and the other connectors, and determining whether the target connector is truly faulty based on the failure probability, includes:

[0033] The similarity between upstream connectors and downstream connector sets is determined by using the mean and standard deviation of the correlation between the upstream and downstream connector sets of the target connector.

[0034] By utilizing the similarity between downstream and upstream connectors, the failure probability of the target connector is obtained to determine whether the target connector is a real fault.

[0035] The present invention also provides a rapid fault location device for robot components, the device comprising: a processor, a memory, a communication interface, and a computer program stored in the memory; the communication interface is used to receive abnormal data and timestamps uploaded from edge nodes of each connector of the robot; the processor is used to execute the computer program to implement the steps of the method described above.

[0036] The present invention also provides a rapid fault location system for robot components, the system being used to implement the rapid fault location method for robot components as described in any of the preceding claims; the system includes:

[0037] The timing analysis module is used to monitor abnormal data and timestamps of each connector of the robot, and to obtain the alignment frequency between the target connector and the other connectors.

[0038] The correlation analysis module is used to sequentially identify each connector as a target connector and obtain its alignment frequency with the remaining connectors, and use the alignment frequency to determine the associated connector set; determine the correlation between the target connector and the remaining connectors based on the associated connector set and the alignment frequency; determine the downstream connector set of the target connector according to the robot circuit topology, and obtain the downstream connector similarity between the target connector and the downstream connector set by combining the correlation between the target connector and each connector in the downstream connector set, as well as the associated connector set of each connector in the downstream connector set itself.

[0039] The fault screening module is used to calculate the fault probability of the target connector based on the similarity of the downstream connectors and the correlation between the target connector and the other connectors, and to determine whether the target connector is a real fault based on the fault probability.

[0040] The present invention has the following beneficial effects:

[0041] This invention utilizes local anomaly monitoring of robot components and processes abnormal data via edge nodes, reducing processing load by uploading only abnormal data. It obtains abnormal data uploaded by all robot components, enabling collaborative fault localization and avoiding the inaccuracy of traditional methods that only locate single components. The correlation between connectors is determined by the consistency of the anomaly times of these components and the change in the alignment of anomaly times after replacing the reference connector. By combining the specific connections of the robot and the data flow direction relationship of connectors experiencing simultaneous anomalies, downstream and upstream similarities are determined. Finally, by analyzing the impact of connector failures on their upstream and downstream components, precise fault localization of robot components is achieved, improving the accuracy of fault diagnosis. Attached Figure Description

[0042] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart illustrating the steps of a method for rapid fault localization of robot components according to an embodiment of the present invention.

[0044] Figure 2 This is a detailed flowchart of step S1 in a method for rapid fault localization of robot components provided in an embodiment of the present invention;

[0045] Figure 3 This is a detailed flowchart of step S2 in a method for rapid fault localization of robot components provided in an embodiment of the present invention;

[0046] Figure 4 This is a detailed flowchart of step S23 in a method for rapid fault localization of robot components provided in an embodiment of the present invention;

[0047] Figure 5 This is a detailed flowchart of step S3 in a method for rapid fault localization of robot components provided in an embodiment of the present invention;

[0048] Figure 6 This is a detailed flowchart of step S4 in a method for rapid fault localization of robot components provided in an embodiment of the present invention;

[0049] Figure 7This is a schematic diagram of the hardware operating environment of the robot component fault rapid localization device involved in the embodiments of the present invention;

[0050] Figure 8 This is a schematic diagram of the framework structure of the robot component fault rapid localization system involved in the embodiment of the present invention. Detailed Implementation

[0051] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for rapid fault location of robot components proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0053] The specific solution of the rapid fault location method for robot components provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0054] Example 1:

[0055] For a method for rapid fault location of robot components provided by this invention, please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of a method for rapid fault localization of robot components according to an embodiment of the present invention.

[0056] The method includes:

[0057] Step S1: Monitor the abnormal data and timestamps of each connector of the robot to obtain the alignment frequency between the target connector and the other connectors;

[0058] Each connector on the robot is treated as an edge node. A local monitoring model (anomaly detection) is deployed on each edge node to identify abnormal data from the connector. The method for identifying abnormal data is the same as traditional methods, such as judging by the stability of the connector's time-series data or the consistency of the connector's input and output. The edge node uploads the detected abnormal data and its corresponding timestamp; normal data is not uploaded. All abnormal data detected by the robot's connector edge nodes is uploaded to the robot's data processing module, which acts as an edge server. The data is then analyzed on the edge server.

[0059] A robot consists of multiple components, each performing a different function, and data and power are transmitted between them via connectors. Therefore, when a robot performs a certain action, it actually affects the entire data and power transmission process. Consequently, there are interrelationships between the components. This makes traditional fault analysis based solely on a single connector less accurate.

[0060] Specifically, please refer to Figure 2 Step S1 includes:

[0061] Step S11: Determine the transmission time difference of abnormal data between the target connector and the other connectors in the robot;

[0062] Step S12: If the transmission time difference is less than or equal to the preset time difference threshold, it is determined that the abnormal data between the connectors is a case of time alignment completion.

[0063] Step S13: If the transmission time difference is greater than the preset time difference threshold, it is determined that the abnormal data between the connectors is due to incomplete time alignment.

[0064] Step S14: The ratio of the number of time alignment cases to the total number of time alignment cases is used as the alignment frequency between the target connector and the other connectors.

[0065] On the edge server, select an abnormal data uploaded by a connector edge node, obtain the corresponding timestamp, and determine the alignment of the abnormal data among all abnormal data from other connectors. Since the robot's components work together to transmit data and power, under normal circumstances, a failure in one robot component will cause a certain degree of failure in the other components, thus ensuring temporal consistency. However, there is a transmission time difference (value) between these abnormal times; therefore, the transmission time difference of abnormal data from all edge nodes is limited to a certain range. Within 100ns, This is recorded as the preset time difference threshold. The determination of this value requires comprehensive consideration of the robot control system's communication cycle, signal transmission delay, and sensor sampling frequency. Typically, It can be set to 1 / 2 to 1 / 5 of the system's minimum control cycle. For example, for a system with a control cycle of 1 ms, It can be initially set to 200μs, and then fine-tuned based on the actual temporal aggregation of abnormal data. The specific value can be determined according to the actual transmission line length and data flow direction; this is just an example.

[0066] In other words, the transmission time difference between the abnormal data uploaded by the other connectors and the target connector (any connector) is within If the transmission time difference is within the preset time difference threshold, the two are considered to have completed time alignment. The timestamp for the completion of time alignment is marked as 1. If there are no abnormal events uploaded within a certain period, it is marked as 0. That is, if the transmission time difference is greater than the preset time difference threshold, it is determined that the two have not completed time alignment.

[0067] The above analysis is performed on the timestamps of the abnormal data uploaded by the edge nodes of the target connector to be analyzed, obtaining the corresponding marking information for the remaining connectors. If two connectors exhibit high frequency of time alignment, it indicates a strong temporal correlation between the anomalies of the target connector and the anomalies of the other connectors. The following alignment information is used. and alignment frequency Characterization:

[0068]

[0069] Indicates the connector to be analyzed After the upload of abnormal data (as the target connector), the tag is aligned with the time of the other connectors b; This represents the alignment of each outlier data point of connector a with connector b, expressed as a vector with a length equal to the number of outliers in connector a. The entire vector is represented by 0s and 1s.

[0070]

[0071] Indicates the selected connector After uploading abnormal data, the alignment frequency with the other connectors b in time; or They represent connectors respectively. The number of outlier data points marked as 1 or 0. The number of 1s or 0s marked in the vector actually represents the sum of the number of elements that have completed time alignment and the number that have not. This is the total number of time-aligned cases; The larger the value, the more it indicates that connectors a and b share the same anomalous common characteristics (when analyzing connector a).

[0072] Step S2: Sequentially select each connector as the target connector and obtain the alignment frequency with the remaining connectors. Use the alignment frequency to determine the associated connector set. Determine the association between the target connector and the remaining connectors based on the associated connector set and the alignment frequency.

[0073] Specifically, please refer to Figure 3 Step S2 includes:

[0074] Step S21: Determine the alignment frequencies when the target connector is used as the connector to be analyzed, and use an unsupervised learning algorithm to select multiple connectors with similar alignment frequencies.

[0075] As another alternative implementation, associated connectors can also be filtered by setting a fixed threshold: the alignment frequencies are sorted in descending order, and the top K connectors are selected, or the connectors with alignment frequencies higher than a preset threshold θ (e.g., θ=0.7) are selected as associated connectors of the target connector.

[0076] Step S22: Select multiple connectors with similar alignment frequencies as associated connectors of the target connector to obtain the associated connector set of the target connector.

[0077] Based on the above embodiments, the connector to be analyzed is obtained. For each alignment frequency of all other connectors b, the ocsvm (One-Class Support Vector Machine, an unsupervised learning algorithm) algorithm (an existing algorithm) is used to perform anomaly detection and thresholding of the alignment frequencies, thereby selecting connectors whose alignment frequencies are close to those of the connectors (i.e., alignment frequencies within the ocsvm decision boundary) as connectors. The associated connectors are then used to obtain the set of associated connectors, denoted as . The elements inside are the connector number and the corresponding alignment frequency.

[0078] The unsupervised learning algorithm uses the alignment frequencies of all connector pairs collected during historical operation as the training dataset, which is considered to mainly contain "normal" association patterns. A radial basis function (RBF) is used as the kernel function, and the decision boundary is controlled by adjusting the outlier scaling parameter ν (typically set to 0.05~0.1). Connectors whose alignment frequencies fall within the decision boundary are classified as having "similar alignment frequencies".

[0079] The above-mentioned anomalies are obtained through the ocsvm algorithm mechanism instead of a fixed threshold segmentation method, because the number of anomalies uploaded by different connectors and at different time periods is different, so the segmentation results should be dynamically adjusted.

[0080] Step S23: Based on the alignment frequencies of the target connector and the other connectors and the associated connector set, obtain the correlation between the target connector and the other connectors.

[0081] More specifically, please refer to Figure 4 Step S23 includes:

[0082] Step S231: Using the alignment frequencies of the target connector and the other connectors, the initial correlation between the target connector and the other connectors is obtained;

[0083] Based on the above embodiments, the analyzed connector is changed to another connector b, and the alignment between the selected connector b and each connector a is analyzed. Connectors a and b themselves are not different; they are simply referred to as a or b for ease of description. It is important to note the difference from the analysis of fixed connector a and the other connectors b (i.e., when connector a of the robot component malfunctions, b also malfunctions; however, this does not mean that if b malfunctions, a will malfunction simultaneously). Therefore, a and b are not equivalent in the anomaly correlation analysis, requiring the aforementioned replacement analysis.

[0084] If, after replacement, both are still in the set of associated connectors, then connectors a and b have a high degree of abnormal association. Therefore:

[0085]

[0086] This indicates the initial association between connectors a and b; , These represent the alignment frequencies of the connector and the other connectors relative to each other.

[0087] If a higher alignment frequency is maintained after changing the connector being analyzed, then the correlation between the two is stronger.

[0088] Step S232: Using the alignment frequencies of the target connector and the other connectors, and the set of associated connectors, obtain the correction coefficient for the initial correlation.

[0089] Specifically, step S232 includes:

[0090] The set of abnormally consistent connectors is obtained by using the intersection of the target connector and the respective associated connector sets of the other connectors.

[0091] Determine the percentage of abnormally consistent connectors in the union of the abnormally consistent connector set with respect to the target connector and the respective associated connector sets of the other connectors.

[0092] The correction coefficient for the initial correlation is obtained by using the proportion of abnormally consistent connectors and the difference in alignment frequency between the target connector and the other connectors.

[0093] Based on the above embodiments, if the connectors are not time-aligned, a certain auxiliary judgment is required. If the robot has a connector component failure, it will cause serious loss of data flow or machine power transmission, leading to abnormal operation of downstream machine components or stopping operation due to receiving unparsed instructions. Therefore, when replacing the analyzed robot connector, there will be a large difference in alignment frequency, making it impossible to classify them into associated connectors. Furthermore, the sets of associated connectors differ significantly, influenced by the selected connector, resulting in significant differences in analysis results. Conversely, if it is a normal robot control signal, it will require adjustments to the robot's power supply to adapt to the load. Therefore, data anomalies during data or power transmission are synchronous until they reach the robot's end effector, and do not depend on the selected connector. Therefore, if they do not exist simultaneously in associated connectors, the following method is used to further obtain the association between connectors:

[0094]

[0095]

[0096] This represents the set of remaining associated connectors for connector 'a' after excluding connector 'b' when the connector to be analyzed is 'a'; similarly... This represents the set of remaining associated connectors of connector b after excluding connector a when the connector to be analyzed is b.

[0097] Let be the intersection of the (remaining) associated connector sets of connectors a and b respectively. Let be the set of connectors that can maintain a high degree of consistency with the anomalies uploaded by the corresponding connectors after selecting different connectors a and b. Let be the anomaly-consistent connector set.

[0098] This represents the percentage of abnormally consistent connectors in the union of the set of abnormally consistent connectors that can still maintain abnormal consistency and the respective associated connector sets of connectors a and b. The larger this percentage, the less likely it is that changing the analyzed connectors will affect the final set judgment, thus increasing the probability that it is a normal phenomenon. The aim is to reduce the interference on the alignment frequency caused by differences in the number of data points uploaded by different connectors.

[0099] This reflects that the smaller the difference in alignment frequencies between the two at this point, the stronger the correlation between connectors a and b; to avoid the denominator being zero here, we can... When the denominator is 0, a minimum value x is set in the denominator, x = 0.01.

[0100] Jab represents the correction coefficient for the initial correlation between connectors a and b;

[0101] Step S233: Correct the initial correlation using correction coefficients to obtain the corrected correlation between the target connector and the other connectors.

[0102] It can be represented as:

[0103]

[0104] This is a preset positive correlation normalization function used to map input values ​​to the interval [0,1]. Indicates the correlation after correction; That is, initial correlation.

[0105] Step S3: Determine the downstream connector set of the target connector based on the robot circuit topology. Combine the correlation between the target connector and each connector in the downstream connector set, as well as the correlation set of each connector in the downstream connector set, to obtain the downstream connector similarity between the target connector and the downstream connector set.

[0106] In this embodiment, the topology of the associated set is analyzed. It should be noted that the route topology can be obtained from the robot's structural topology diagram. The topology diagram reveals the connection status between the robot's components.

[0107] The location of the connector set can be obtained from the topology diagram, and the abnormal connection set can be obtained from the topology diagram. The specific analysis process is as follows:

[0108] As described in the above embodiments, if the connector is in an abnormal state, it will affect subsequent data transmission and stable power supply. Therefore, although the state of the connector and the downstream components may both appear abnormal, the abnormality of the downstream robot components is actually caused by the failure of the connector or other robot components connected to the connector. Alternatively, the downstream robot components may maintain a normal and stable state, indicating that the downstream equipment itself is in a normal working state. Therefore, if the connector experiences a genuine failure, the abnormal behavior of its downstream connectors may be similar, or the downstream components may appear normal, but the abnormality will differ significantly from that of the connector itself with a genuine failure. Conversely, if the abnormality is normal but the power supply changes due to load variations, both data and power will change, but the connector and the downstream connector will exhibit similar abnormal characteristics.

[0109] Specifically, please refer to Figure 5 Step S3 includes:

[0110] Step S31: Utilize the correlations between the target connector and the set of downstream connectors in its line topology to obtain the corresponding correlation standard deviation.

[0111] Step S32: Use the correlation standard deviation between the target connector and the downstream connector set to obtain the initial similarity between the target connector and the downstream connector set;

[0112] Select connector 'a' to be analyzed, and obtain the connectors connected to 'a' from the line topology (obtaining its upstream and downstream sets). Upstream and downstream connections are determined by the data flow direction. If it is the first or last connector, it belongs to the first position of signal or data transmission or the position of the end effector. In this case, connector anomalies are normal and do not require analysis.

[0113] Obtain the initial similarity between connector a and the downstream connector set. :

[0114]

[0115] This represents the standard deviation of the correlation between connector a and each connector in the downstream connector set; As mentioned above, this indicates the revised relevance; This represents the positive correlation normalization function.

[0116] Step S33: Using the initial similarity and the intersection between each associated connector set corresponding to the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0117] More specifically, step S33 includes:

[0118] Determine the percentage of the number of connectors in the intersection of each associated connector set corresponding to the downstream connector set, relative to the total number of connectors in the downstream connector set.

[0119] The corrected similarity between the target connector and the downstream connector set is obtained by using the initial similarity and the proportion of connectors in the number of connectors.

[0120] By using the corrected similarity and the mean correlation between the target connector and the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0121] Based on the above embodiments, taking connectors a and b as examples, the correlation between each pair of connectors is obtained. and the corresponding set of exception-consistent connectors If it is necessary to determine whether connector a is faulty, connector a is compared with the other connectors b. This involves iterating through connectors b, checking each... The intersection of these sets yields a set of connectors that share the same anomalous changes as a. This can be denoted as the set of connectors with common abnormal changes. The corresponding expression is as follows:

[0122]

[0123] For a set of connectors whose downstream connector sets share the same common changes, the greater the similarity between connector a and the downstream connector set, the better.

[0124]

[0125] This represents the number of connectors in the intersection of the common abnormal changes among downstream (excluding a) connectors. In fact, it is also the intersection of the various associated connector sets corresponding to the downstream connector set. Here, each connector in the downstream connector set is also represented by b. This indicates the total number of downstream connectors; The proportion of connectors represents the consistency of the set of common changes downstream; therefore:

[0126]

[0127] To correct for similarity; This represents the initial similarity.

[0128] Subsequently, if the relationship between connector a and connector b in the downstream connector set... A higher value indicates that the downstream connector has a similar alignment frequency to the connector being analyzed (a), exhibiting the same abnormal behavior, and thus falls under the normal state, more likely due to normal load dynamic changes. Therefore:

[0129]

[0130] This indicates the calculation of the mean correlation between connector a and the set of downstream connectors;

[0131] This indicates the similarity between connector a and the downstream connectors in the downstream connector set.

[0132] Step S4: Calculate the failure probability of the target connector based on the similarity of the downstream connectors and the correlation between the target connector and the other connectors, and determine whether the target connector is a real failure based on the failure probability.

[0133] Specifically, please refer to Figure 6 Step S4 includes:

[0134] Step S41: Use the mean and standard deviation of the correlation between the upstream connectors and downstream connector sets of the target connector to determine the similarity between the upstream connectors and the upstream connectors of the downstream connector sets.

[0135] Step S42: Using the similarity between downstream connectors and upstream connectors, the failure probability of the target connector is obtained to determine whether the target connector is a real fault.

[0136] If the selected connector's malfunction is caused by interference from faults in other locations, it may appear abnormal even though it is functioning normally. Therefore, to achieve accurate fault location, the status of upstream robot components needs to be considered. If the connector is affected by interference, the nearest upstream connector will exhibit the same abnormal behavior, mirroring the downstream connector's abnormal behavior. If it is a genuine fault, data and power supply will be normal before the fault, but an abnormality will occur afterward, resulting in a significant difference from the downstream situation.

[0137] Therefore, select the upstream nearest connector of connector a. Obtain connector The correlation between a and the downstream connector set of a yields the corresponding similarity. :

[0138]

[0139] Indicates connector The mean correlation between a and the set of downstream connectors of a;

[0140] Indicates connector The standard deviation of the correlation between a and the downstream connector set of a; This represents a positive correlation normalization function, which can be specifically, for example, a maximum and minimum value normalization function;

[0141] Indicates that the connector is considered after upstream. Similarity to upstream connectors in downstream connector sets;

[0142] This leads to the failure probability of connector a. :

[0143]

[0144] This represents the positive correlation normalization function. Indicates similarity between downstream connectors; This indicates the similarity of the upstream connectors. The denominator is increased by 0.1 to prevent it from being zero, which would invalidate the formula.

[0145] The higher the similarity between connector a and the downstream, the lower the similarity after considering the upstream, indicating that connector a has a higher probability of failure.

[0146] During the offline training phase, historical operational data of the robot is collected, and connectors that actually malfunction in each time period are identified through manual labeling or known fault injection experiments, thus forming a fault probability sample set labeled "faulty" and "normal". Based on this sample set, a classifier is trained using Fisher's discriminant method, and the resulting decision boundary value is the fault probability threshold.

[0147] The failure probability is greater than the failure probability threshold. The corresponding connector is marked as a real fault.

[0148] The edge server transmits the marked real faults to the corresponding connector edge nodes to complete the fault location of the robot parts.

[0149] This invention utilizes local anomaly monitoring of robot components and processes abnormal data via edge nodes, reducing processing load by uploading only abnormal data. It obtains abnormal data uploaded by all robot components, enabling collaborative fault localization and avoiding the inaccuracy of traditional methods that only locate single components. The correlation between connectors is determined by the consistency of the anomaly times of these components and the change in the alignment of anomaly times after replacing the reference connector. By combining the specific connections of the robot and the data flow direction relationship of connectors experiencing simultaneous anomalies, downstream and upstream similarities are determined. Finally, by analyzing the impact of connector failures on their upstream and downstream components, precise fault localization of robot components is achieved, improving the accuracy of fault diagnosis.

[0150] Example 2:

[0151] This invention also proposes a rapid fault location device for robot components. This rapid fault location device can be used with a robot. Figure 7 As shown, Figure 7 This is a schematic diagram of the hardware operating environment of the robot component fault rapid localization device involved in the embodiments of the present invention.

[0152] like Figure 7As shown, the robot component fault rapid location device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display or an input unit such as a control panel; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed RAM or a stable, non-volatile memory, such as a disk storage device. Optionally, the memory 1005 may also be a storage system independent of the aforementioned processor 1001. The memory 1005, as a computer storage medium, may include a robot component fault rapid location program.

[0153] Those skilled in the art will understand that Figure 7 The hardware structure shown does not constitute a limitation on the device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0154] Continue to refer to Figure 7 , Figure 7 The memory 1005, which is a computer-readable storage medium, may include an operating system, a user interface module, a network communication module, and a rapid fault location program for robot components.

[0155] exist Figure 7 In this embodiment, the network communication module is mainly used to connect to the server and can communicate with the server for data; while the processor 1001 can call the robot component fault rapid location program stored in the memory 1005 and execute the steps in the above embodiments.

[0156] Based on the hardware structure of the above-mentioned rapid fault location device for robot components, various embodiments of the rapid fault location method for robot components of the present invention are implemented.

[0157] In addition, the present invention also provides a rapid fault location system for robot components, please refer to... Figure 8 The rapid fault location system for robot components includes:

[0158] The timing analysis module A10 is used to monitor abnormal data and timestamps of each connector of the robot, and to obtain the alignment frequency between the target connector and the other connectors.

[0159] The correlation analysis module A20 is used to sequentially identify each connector as a target connector and obtain its alignment frequency with the remaining connectors, and use the alignment frequency to determine the associated connector set; determine the correlation between the target connector and the remaining connectors based on the associated connector set and the alignment frequency; determine the downstream connector set of the target connector according to the robot circuit topology, and obtain the downstream connector similarity between the target connector and the downstream connector set by combining the correlation between the target connector and each connector in the downstream connector set, as well as the associated connector set of each connector in the downstream connector set itself.

[0160] The fault screening module A30 is used to calculate the fault probability of the target connector based on the similarity of the downstream connectors and the correlation between the target connector and the other connectors, and to determine whether the target connector is a real fault based on the fault probability.

[0161] Furthermore, the timing analysis module A10 is also used for:

[0162] Determine the time difference in the transmission of abnormal data between the target connector and the other connectors in the robot;

[0163] If the transmission time difference is less than or equal to the preset time difference threshold, it is determined that the abnormal data between the connectors is a case of time alignment being completed.

[0164] If the transmission time difference is greater than the preset time difference threshold, it is determined that the abnormal data between the connectors is due to incomplete time alignment.

[0165] The ratio of the number of times time alignment is completed to the total number of times time alignment is used as the alignment frequency between the target connector and the other connectors.

[0166] Furthermore, the correlation analysis module A20 is also used for:

[0167] When the target connector is used as the connector to be analyzed, the alignment frequencies are determined, and multiple connectors with similar alignment frequencies are selected using an unsupervised learning algorithm.

[0168] By taking multiple connectors with similar alignment frequencies as associated connectors of the target connector, a set of associated connectors of the target connector is obtained.

[0169] The correlation between the target connector and the other connectors is obtained based on their respective alignment frequencies and associated connector sets.

[0170] Furthermore, the correlation analysis module A20 is also used for:

[0171] The initial correlation between the target connector and the other connectors is obtained by using their respective alignment frequencies.

[0172] The correction coefficient for the initial correlation is obtained by using the alignment frequencies of the target connector and the other connectors, as well as the set of associated connectors.

[0173] The initial correlation is corrected using a correction factor to obtain the corrected correlation between the target connector and the other connectors.

[0174] Furthermore, the correlation analysis module A20 is also used for:

[0175] The set of abnormally consistent connectors is obtained by using the intersection of the target connector and the respective associated connector sets of the other connectors.

[0176] Determine the percentage of abnormally consistent connectors in the union of the abnormally consistent connector set with respect to the target connector and the respective associated connector sets of the other connectors.

[0177] The correction coefficient for the initial correlation is obtained by using the proportion of abnormally consistent connectors and the difference in alignment frequency between the target connector and the other connectors.

[0178] Furthermore, the correlation analysis module A20 is also used for:

[0179] By utilizing the correlations between the target connector and the set of downstream connectors in its line topology, the corresponding correlation standard deviation is obtained.

[0180] The initial similarity between the target connector and the downstream connector set is obtained by using the correlation standard deviation between the target connector and the downstream connector set.

[0181] By utilizing the initial similarity and the intersection between each associated connector set corresponding to the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0182] Furthermore, the correlation analysis module A20 is also used for:

[0183] Determine the percentage of the number of connectors in the intersection of each associated connector set corresponding to the downstream connector set, relative to the total number of connectors in the downstream connector set.

[0184] The corrected similarity between the target connector and the downstream connector set is obtained by using the initial similarity and the proportion of connectors in the number of connectors.

[0185] By using the corrected similarity and the mean correlation between the target connector and the downstream connector set, the downstream connector similarity between the target connector and the downstream connector set is obtained.

[0186] Furthermore, the fault screening module A30 is also used for:

[0187] The similarity between upstream connectors and downstream connector sets is determined by using the mean and standard deviation of the correlation between the upstream and downstream connector sets of the target connector.

[0188] By utilizing the similarity between downstream and upstream connectors, the failure probability of the target connector is obtained to determine whether the target connector is a real fault.

[0189] The specific implementation of the robot component fault rapid location system of the present invention is basically the same as the embodiments of the robot component fault rapid location method described above, and will not be repeated here.

[0190] Furthermore, the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a robot component fault rapid location program, wherein when executed by a processor, the robot component fault rapid location program implements the steps of the robot component fault rapid location method as described above.

[0191] The method implemented when the robot component fault rapid location program is executed can be referred to in various embodiments of the robot component fault rapid location method of the present invention, and will not be repeated here.

[0192] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0193] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0194] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied 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.

[0195] The above description is only a preferred embodiment of the present invention and does not limit the scope of protection of the present invention. All equivalent structural / method transformations made under the inventive concept of the present invention using the contents of the present invention specification and drawings, or direct / indirect applications in other related technical fields, are included within the scope of protection of the present invention.

Claims

1. A method for rapid localization of a robot component failure, the method comprising: The method comprises: Monitoring abnormal data of each connector of the robot and its timestamp, obtaining alignment frequency between the target connector and the remaining connectors; determining the transmission time difference of abnormal data between the target connector and the remaining connectors in the robot; if the transmission time difference is less than or equal to the preset time difference threshold, it is determined that the abnormal data between the connectors is in the case of completed time alignment; if the transmission time difference is greater than the preset time difference threshold, it is determined that the abnormal data between the connectors is in the case of incomplete time alignment; the ratio of the number of completed time alignment cases to the total number of time alignment cases is taken as the alignment frequency between the target connector and the remaining connectors; Each connector is sequentially taken as the target connector, and the alignment frequency with the remaining connectors is obtained, the alignment frequency is used to determine the associated connector set, each alignment frequency is determined when the target connector is taken as the to-be-analyzed connector, and the unsupervised learning algorithm is used to screen out multiple connectors with similar alignment frequencies, i.e., within the ocsvm decision boundary; the multiple connectors with similar alignment frequencies, i.e., within the ocsvm decision boundary, are taken as the associated connectors of the target connector, and the associated connector set of the target connector is obtained; the association between the target connector and the remaining connectors is determined based on the associated connector set and the alignment frequency; the initial association between the target connector and the remaining connectors is obtained by using the alignment frequency of each of the target connector and the remaining connectors; the correction coefficient of the initial association is obtained by using the alignment frequency of each of the target connector and the remaining connectors and the associated connector set; the initial association is corrected by using the correction coefficient, and the association between the target connector and the remaining connectors is obtained. The downstream connector set of the target connector is determined according to the robot line topology structure, the association between the target connector and each connector in the downstream connector set is combined, and the association between each connector in the downstream connector set and the associated connector set of each connector in the downstream connector set is obtained; the similarity between the target connector and the downstream connectors of the downstream connector set is obtained by using the similarity between the target connector and each connector in the downstream connector set in the line topology structure; the corresponding association standard deviation is obtained by using the association between the target connector and the downstream connector set; the initial similarity between the target connector and the downstream connector set is obtained by using the association standard deviation between the target connector and the downstream connector set; the intersection between the initial similarity and each associated connector set corresponding to the downstream connector set is obtained, and the similarity between the target connector and the downstream connectors of the downstream connector set is obtained. According to the downstream connector similarity and the association between the target connector and the remaining connectors, the failure probability of the target connector is calculated, and whether the target connector is a real failure is judged according to the failure probability; the average value and the standard deviation of the association between the upstream connectors and the downstream connector set are used to determine the upstream connector similarity between the upstream connectors and the downstream connector set; the failure probability of the target connector is obtained by using the downstream connector similarity and the upstream connector similarity, so as to judge whether the target connector is a real failure.

2. The method of claim 1, wherein, The correction coefficient of the initial relevance is obtained by using the intersection between the respective sets of associated connectors of the target connector and the remaining connectors, including: An abnormal consistent connector set is obtained by using the intersection between the respective sets of associated connectors of the target connector and the remaining connectors; The abnormal consistent connector quantity proportion is determined by comparing the number of abnormal consistent connectors in the intersection between the respective sets of associated connectors of the target connector and the remaining connectors with the number of abnormal consistent connectors in the union between the respective sets of associated connectors of the target connector and the remaining connectors; The correction coefficient of the initial relevance is obtained by using the abnormal consistent connector quantity proportion and the difference between the alignment frequencies of the target connector and the remaining connectors.

3. The method of claim 1, wherein, The downstream connector similarity between the target connector and the downstream connector set is obtained by using the intersection between the respective sets of associated connectors corresponding to the initial similarity and the downstream connector set, including: The connector quantity proportion in the intersection between the respective sets of associated connectors corresponding to the downstream connector set is determined by comparing the total number of connectors in the downstream connector set with the number of connectors in the intersection between the respective sets of associated connectors corresponding to the downstream connector set; The corrected similarity between the target connector and the downstream connector set is obtained by using the initial similarity and the connector quantity proportion; The downstream connector similarity between the target connector and the downstream connector set is obtained by using the corrected similarity and the average relevance between the target connector and the downstream connector set.

4. A robotic component failure rapid localization apparatus, comprising: The device includes a processor, a memory, a communication interface, and a computer program stored in the memory; The communication interface is configured to receive abnormal data and timestamps uploaded from each connector edge node of the robot; The processor is configured to execute the computer program to implement the method according to any one of claims 1 to 3.

5. A robotic component failure rapid localization system, comprising: The system is configured to implement the method for quickly locating faults of robot components according to any one of claims 1 to 3; the system includes: A time series analysis module is configured to monitor abnormal data and timestamps of each connector of the robot, obtain the alignment frequency between the target connector and the remaining connectors, determine the transmission time difference of abnormal data between the target connector and the remaining connectors in the robot, determine that the abnormal data between the connectors is in a completed time alignment state if the transmission time difference is less than or equal to a preset time difference threshold, determine that the abnormal data between the connectors is in an incomplete time alignment state if the transmission time difference is greater than the preset time difference threshold, and take the ratio of the number of completed time alignment states to the total number of time alignment states as the alignment frequency between the target connector and the remaining connectors. The association analysis module is used for sequentially taking each connector as a target connector and acquiring alignment frequencies with the remaining connectors, determining an associated connector set by using the alignment frequencies, determining each alignment frequency when the target connector is taken as a to-be-analyzed connector, screening out multiple connectors with similar alignment frequencies, i.e., within the ocsvm decision boundary, by using an unsupervised learning algorithm; taking the multiple connectors with similar alignment frequencies, i.e., within the ocsvm decision boundary, as the associated connectors of the target connector to obtain the associated connector set of the target connector; determining the association between the target connector and the remaining connectors based on the associated connector set and the alignment frequencies; obtaining the initial association between the target connector and the remaining connectors by using the alignment frequencies of the target connector and the remaining connectors; obtaining a correction coefficient of the initial association by using the alignment frequencies of the target connector and the remaining connectors and the associated connector set; correcting the initial association by using the correction coefficient to obtain the association between the target connector and the remaining connectors; determining a downstream connector set of the target connector according to the robot line topology structure, combining the association of the target connector with each connector in the downstream connector set and the associated connector set of each connector in the downstream connector set to obtain a downstream connector similarity of the target connector and the downstream connector set; obtaining a corresponding association standard deviation by using each association of the target connector and the downstream connector set in the line topology structure; obtaining an initial similarity of the target connector and the downstream connector set by using the association standard deviation of the target connector and the downstream connector set; obtaining the downstream connector similarity of the target connector and the downstream connector set by using the initial similarity and the intersection between each associated connector set corresponding to the downstream connector set; The fault screening module is used for calculating a fault probability of the target connector according to the downstream connector similarity and the association between the target connector and the remaining connectors, judging whether the target connector is a real fault according to the fault probability; determining an upstream connector similarity of the target connector and the downstream connector set by using the mean value and the standard deviation of the association between the upstream connector of the target connector and the downstream connector set; obtaining the fault probability of the target connector by using the downstream connector similarity and the upstream connector similarity to judge whether the target connector is a real fault.

Images