Underwater robot course data anomaly monitoring method and system

Abstract

The invention relates to the technical field of data processing, in particular to an underwater robot course data anomaly monitoring method and system, and the method comprises the steps: obtaining course data of an underwater robot in a preset time period; clustering the course data to obtain a plurality of first clusters; determining a first interaction index value among the indexes in the heading process; clustering the course data according to a first interaction index value between indexes of the course data in the first cluster in different course processes to obtain a second cluster; calculating an abnormal score value of abnormal data of the index data of each index in the second cluster; and determining an abnormal degree value of each index of the underwater robot according to the abnormal score value of the index, the number of the second cluster where the index data of the index is located and the number of the second clusters, and determining the index of which the abnormal degree value is greater than a first threshold value as an abnormal index. According to the invention, the detection accuracy of the anomaly detection of the underwater robot is improved, and the control precision of the underwater robot is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for monitoring abnormal heading data of an underwater robot. Background Art

[0002] Underwater robots, also known as unmanned remote-controlled submersibles, are extreme operation robots that work underwater. As the world pays more and more attention to the marine economy and marine resources gradually become an important competitive resource, the underwater robot industry has ushered in opportunities for rapid development. Underwater robots can not only help humans explore and utilize marine resources more effectively, but also perform various tasks in complex underwater environments, such as seabed mapping, pipeline inspection, and biological research. Therefore, the heading data of underwater robots is crucial to studying the degree of task completion of underwater robots. Therefore, it is necessary to monitor the abnormality of underwater robot industry data.

[0003] In some scenarios, when detecting anomalies in the heading data of an underwater robot, the Local Outlier Factor (LOF) algorithm is usually used to monitor the acquired heading data of the underwater robot for anomalies. However, in actual operation, the underwater environment is complex and changeable. For example, changes in the direction of the water flow, rapid changes in temperature and salinity will cause short-term fluctuations in the heading data, but the fluctuations in the heading data caused by changes in the underwater environment are normal. The LOF algorithm usually misjudges the changes in the heading data caused by fluctuations as anomalies, resulting in errors in the heading control and judgment of the underwater robot. Therefore, when the existing method is used to detect anomalies in the heading data of the underwater robot, the detection accuracy is low, which in turn leads to low control accuracy of the underwater robot. Summary of the invention

[0004] In order to solve the technical problem of low detection accuracy when detecting abnormality of heading data of an underwater robot, the object of the present invention is to provide a method and system for monitoring abnormality of heading data of an underwater robot. The technical scheme adopted is as follows: In a first aspect, an embodiment of the present invention provides a method for monitoring abnormal heading data of an underwater robot, including: obtaining the heading data of the underwater robot within a predetermined time period, where the heading data includes the index data of multiple indicators; clustering the heading data to obtain multiple first clusters, and each first cluster includes the heading data of the same section during multiple heading processes of the underwater robot; determining the first interaction index value between each indicator during the heading process according to the index data of multiple indicators in the heading data within each first cluster; clustering the heading data according to the first interaction index value between the indicators in the first cluster during different heading processes to obtain a second cluster; calculating the abnormal score value of the abnormal data of the index data of each indicator in the second cluster; determining the abnormal degree value of each indicator of the underwater robot according to the abnormal score value of each indicator, the number of the second cluster where the index data of the indicator is located, and the number of the second cluster, and determining the indicator with the abnormal degree value greater than the first threshold as the abnormal indicator.

[0005] Optionally, determining the first interaction index value between each indicator during the heading process according to the index data of multiple indicators in the heading data within each first cluster includes: obtaining the fluctuation curve of the index data of each indicator in each first cluster during multiple heading processes; clustering the index data of each indicator according to the first DTW distance between the fluctuation curves to obtain multiple third clusters; determining the target cluster in which there is a positive correlation or negative correlation between the indicators according to the average value of the second DTW distances of the index data of all indicators in the third cluster; determining the second interaction index value between the indicators during the heading process in the target cluster according to the third DTW distance between the fluctuation curves of the indicators during the heading process in the target cluster and the fluctuation amplitude of the fluctuation curves of the indicators; determining the first interaction index value between each indicator during the heading process according to the second interaction index value between the indicators during the heading process in each target cluster.

[0006] Optionally, determining the target cluster in which there is a positive correlation or negative correlation between the indicators according to the average value of the second DTW distances of the index data of all indicators in the third cluster includes: determining that there is a positive correlation between all indicators in the third cluster when the average value of the second DTW distances is less than the second threshold; determining that there is a negative correlation between all indicators in the third cluster when the average value of the second DTW distances is not less than the second threshold.

[0007] Optionally, when the target cluster is the third cluster with a positive correlation, determining the second interaction index value between the indicators in the heading process of the target cluster according to the third DTW distance between the fluctuation curves of the indicators in the heading process of the target cluster and the fluctuation amplitudes of the fluctuation curves of the indicators includes: calculating the first product between the first quantity of all indicators in the third cluster and the second quantity of influencing indicators that have a positive correlation with the indicators in the third cluster; calculating the third DTW distance between the fluctuation curve of the indicator in the target cluster and the fluctuation curve of the influencing indicator that has a positive correlation with the indicator, and normalizing the third DTW distance to obtain the first normalized distance; calculating the absolute value of the first difference between the fluctuation amplitude of the fluctuation curve of the indicator in the target cluster and the fluctuation amplitude of the fluctuation curve of the influencing indicator that has a positive correlation with the indicator, and calculating the reciprocal of the first sum value between the absolute value of the first difference and a preset value; superimposing the second product between the first normalized distance of each indicator in the third cluster and the reciprocal of the first sum value to obtain the first superimposed value; determining the first ratio between the first superimposed value and the first product as the second interaction index value.

[0008] Optionally, when the target cluster is the third cluster with a negative correlation, determining the second interaction index value between the indicators in the heading process of the target cluster according to the third DTW distance between the fluctuation curves of the indicators in the heading process of the target cluster and the fluctuation amplitudes of the fluctuation curves of the indicators includes: calculating the third product between the third quantity of all indicators in the third cluster and the fourth quantity of influencing indicators that have a negative correlation with the indicators in the third cluster; calculating the third DTW distance between the fluctuation curve of the indicator in the target cluster and the fluctuation curve of the influencing indicator that has a negative correlation with the indicator, and normalizing the third DTW distance to obtain the second normalized distance; calculating the reciprocal of the second sum value between the second normalized distance and a preset value; calculating the absolute value of the second difference between the fluctuation amplitude of the fluctuation curve of the indicator in the target cluster and the fluctuation amplitude of the fluctuation curve of the influencing indicator that has a negative correlation with the indicator, and calculating the reciprocal of the third sum value between the absolute value of the second difference and a preset value; superimposing the fourth product between the reciprocal of the second sum value of each indicator in the third cluster and the reciprocal of the third sum value to obtain the second superimposed value; determining the second ratio between the second superimposed value and the third product as the second interaction index value.

[0009] Optionally, determining the first interaction index value between the indicators in the heading process according to the second interaction index value between the indicators in the heading process of each target cluster includes: determining the average value of the second interaction index value between the indicators in the heading process of each target cluster as the first interaction index value.

[0010] Optionally, cluster the heading data according to the first interaction index value among the indicators of the heading data in the first cluster during different heading processes, and obtain a second cluster, including: determining the matching degree value between the first clusters of the heading data in different heading processes according to the indicators in the first cluster of the heading data during different heading processes; using the matching degree value to determine the matching quantity of the first clusters of the heading data in different heading processes; determining the similarity degree value between the heading data in different heading processes according to the matching quantity, the quantity of the first clusters of the heading data in different heading processes, and the first interaction index value among the indicators in different heading processes; clustering the heading data based on the similarity degree value between the heading data in different heading processes to obtain a second cluster.

[0011] Optionally, determining the matching degree value between the first clusters of the heading data in different heading processes according to the indicators in the first cluster of the heading data during different heading processes includes: determining the fifth quantity of the same indicators in the first cluster of the heading data in different heading processes, and the sixth quantity of the indicators in the first cluster of the heading data during different heading processes; determining the third ratio between the fifth quantity and the sixth quantity as the matching degree value.

[0012] Optionally, determining the anomaly degree value of each indicator of the underwater robot according to the anomaly score value of each indicator, the number of the second cluster where the indicator data of the indicator is located, and the quantity of the second cluster includes: calculating the fourth ratio between the number of the second cluster where the indicator data of the indicator is located and the quantity of the second cluster; determining the fourth product between the fourth ratio and the anomaly score value of the indicator as the anomaly degree value of the indicator.

[0013] In a second aspect, an embodiment of the present invention provides an underwater robot heading data anomaly monitoring system, including: a processor and a memory; wherein, the memory is used for storing a computer program that can run on the processor; the processor is used for executing the program stored on the memory to implement the steps of the underwater robot heading data anomaly monitoring method as mentioned in the first aspect.

[0014] The present invention has the following beneficial effects: First, obtain the heading data of the underwater robot within a predetermined time period, where the heading data includes index data of multiple indicators; then cluster the heading data to obtain multiple first clusters, and each first cluster includes the heading data of the same section during multiple heading processes of the underwater robot; secondly, determine the first interaction index value between each indicator during the heading process according to the index data of multiple indicators in the heading data within each first cluster; then cluster the heading data according to the similarity degree value between the heading data in the first cluster during different heading processes to obtain a second cluster; and calculate the anomaly score value of the anomaly data of the index data of each indicator in the second cluster; finally, determine the anomaly degree value of each indicator of the underwater robot according to the anomaly score value of each indicator, the number of the second cluster where the index data of the indicator is located, and the number of the second cluster, and determine the indicator with the anomaly degree value greater than the first threshold as the anomaly indicator.

[0015] In this way, the embodiment of the present invention can determine the interaction index value between various indicators during the operation of the underwater robot by analyzing the heading data of the underwater robot when working in the same area. Then, according to the interaction index value between the indicators of different heading data, determine the similarity between the heading data during different heading processes, so as to cluster the heading data. Then calculate the anomaly score value of the anomaly data of the index data of each indicator in the clustering cluster, and calculate the anomaly degree value of each indicator based on this. Thus, the monitoring of the abnormal indicators is achieved. In this way, it is possible to avoid the situation where the normal fluctuation of the heading data caused by the change of the underwater environment is misjudged as abnormal, improve the detection accuracy of the abnormal detection of the underwater robot, and further improve the control accuracy of the underwater robot. Brief Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] Figure 1 It is a flowchart of a method for monitoring abnormal heading data of an underwater robot provided by an embodiment of the present invention.

[0018] Figure 2 It is a schematic structural diagram of a device for monitoring abnormal heading data of an underwater robot provided by an embodiment of the present invention.

[0019] Figure 3 It is a schematic structural diagram of a system for monitoring abnormal heading data of an underwater robot provided by an embodiment of the present invention. Detailed Embodiments

[0020] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following specifically describes, in conjunction with the accompanying drawings and preferred embodiments, a method and system for monitoring abnormal heading data of an underwater robot according to the present invention, including its specific implementation manner, structure, features, and effects. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0022] The following specifically describes the specific solution of a method and system for monitoring abnormal heading data of an underwater robot provided by the present invention in conjunction with the accompanying drawings.

[0023] Embodiment 1: Please refer to Figure 1 , which shows a flowchart of a method for monitoring abnormal heading data of an underwater robot provided by an embodiment of the present invention, including: S101, obtaining the heading data of the underwater robot within a predetermined time period, where the heading data includes index data of multiple indicators.

[0024] Specifically, in an embodiment of the present invention, the heading data of the underwater robot can be obtained from the terminal database. Among them, the heading data of the underwater robot includes index data such as heading angle, angular velocity, acceleration, and speed. The index data of each index in these heading data together constitute the heading data set of the underwater robot, providing necessary information for the navigation, control, and task execution of the underwater robot. In an embodiment of the present invention, for each underwater robot, it is assumed that the heading data of M navigation processes is obtained in total.

[0025] S102, clustering the heading data to obtain a plurality of first clusters, where each first cluster includes the heading data of the same section of the underwater robot during multiple heading processes.

[0026] Specifically, when the underwater robot performs tasks in the same section, the corresponding heading data each time is not very different. Therefore, first, the heading data with the same section in the heading data of these M heading processes is clustered into one category, that is, the heading data with the same section is clustered to obtain the first cluster, and the heading data in each first cluster is the heading data in the same section. In an embodiment of the present invention, subsequent analysis is performed on the heading data in the same category. In an embodiment of the present invention, for the heading data in the same category, it is assumed that there are a total of R times of heading data.

[0027] S103. Determine the first interaction index value between each index during the course of determining the heading based on the index data of multiple indices in the heading data of each first cluster.

[0028] Specifically, during each navigation of the underwater robot, the fluctuations between different indices in the heading data will affect each other. For example, the fluctuation of the heading angle will affect indices such as acceleration and speed. In the embodiments of the present invention, the heading angle is recorded as the reference index, and acceleration and speed are recorded as the influencing indices of the reference index, that is, the indices that have a correlation with the heading angle.

[0029] Further, when determining the first interaction index value between each index during the course of determining the heading, as an optional embodiment of the present invention, first obtain the fluctuation curves of the index data of each index in each first cluster during multiple heading processes; Then, cluster the index data of each index according to the first DTW distance between the fluctuation curves to obtain multiple third clusters; then determine the target cluster with a positive or negative correlation between the indices according to the average value of the second DTW distances of the index data of all the indices in the third cluster; secondly, determine the second interaction index value between each index during the heading process in the target cluster according to the third DTW distance between the fluctuation curves of each index during the heading process in the target cluster and the fluctuation amplitude of the fluctuation curves of each index; finally, determine the first interaction index value between each index during the heading process according to the second interaction index value between each index during the heading process in each target cluster.

[0030] Specifically, in the embodiments of the present invention, obtain the fluctuation curves of the index data of each index in each first cluster during multiple heading processes, then calculate the DTW distance between the fluctuation curves of each index during each heading process, and normalize the calculated DTW distance. Then use the K-means clustering algorithm to cluster all the indices based on the normalized DTW distance, where the number of clustering clusters can be determined by the silhouette coefficient method. After clustering, there are N third clusters in total. It should be noted that the calculation method of the DTW distance and the specific implementation method of determining the number of clustering clusters by the silhouette coefficient method can refer to the prior art, and the embodiments of the present invention will not elaborate here.

[0031] Further, after classifying multiple third clusters, calculate the average value of the second DTW distances of the index data of all the indicators in the third cluster. When determining the correlation relationship between the indicators based on this average value, as an optional embodiment of the present invention, when the average value of the second DTW distances is less than the second threshold, it is determined that there is a positive correlation relationship between all the indicators in the third cluster; when the average value of the second DTW distances is not less than the second threshold, it is determined that there is a negative correlation relationship between all the indicators in the third cluster. Among them, the second threshold can be determined according to the actual situation, and in the embodiment of the present invention, the value here is 0.5. If the average value of the second DTW distances of the index data of all the indicators in the third cluster is less than 0.5, it means that there is a positive correlation relationship between all the indicators in this third cluster, otherwise it means that there is a negative correlation relationship between all the indicators in this third cluster.

[0032] Further, when determining the second interaction index value between the indicators in the course of the heading in the target cluster, if the target cluster is a third cluster with a positive correlation relationship, as an optional embodiment of the present invention, first calculate the first product between the first quantity of all the indicators in the third cluster and the second quantity of the influencing indicators that have a positive correlation relationship with the indicators in the third cluster; then calculate the third DTW distance between the fluctuation curves of the indicators in the target cluster and the fluctuation curves of the influencing indicators that are positively correlated with the indicators, and normalize the third DTW distance to obtain the first normalized distance; secondly, calculate the absolute value of the first difference between the fluctuation amplitudes of the fluctuation curves of the indicators in the target cluster and the fluctuation amplitudes of the fluctuation curves of the influencing indicators that are positively correlated with the indicators, and calculate the reciprocal of the first sum value between the absolute value of the first difference and the preset value; and superimpose the second products between the first normalized distance of each indicator in the third cluster and the reciprocal of the first sum value to obtain the first superimposed value; finally, determine the first ratio between the first superimposed value and the first product as the second interaction index value.

[0033] Specifically, taking the heading data of the r-th course of the heading as an example in the embodiment of the present invention, clustering the heading data of the r-th course of the heading according to the above embodiment of the present invention to obtain multiple third clusters. For the index data of each indicator in the n-th third cluster (assuming that there is a positive correlation relationship among the indicators in the n-th third cluster), the second interaction index value between its indicators can be calculated by the following formula:

[0034] In the above formula, represents the second interaction index value between the indicators in the n-th cluster in the heading data of the r-th course of the heading. represents the number of reference indicators in the n-th cluster. J represents the number of indicators affected by the i-th reference indicator, that is, the second quantity of the influencing indicators that have a positive correlation relationship with the reference indicators in the third cluster. represents a normalization function, which is used to perform normalization processing. represents the third DTW distance between represents the fluctuation curve of the i-th reference index in the n-th cluster of the heading data of the r-th heading process. represents the fluctuation curve of the j-th index affected by the i-th reference index in the n-th cluster of the heading data of the r-th heading process. represents the fluctuation amplitude of the fluctuation curve of the i-th reference index in the n-th cluster of the heading data in the r-th heading process. represents the fluctuation amplitude of the fluctuation curve of the j-th index affected by the i-th reference index in the n-th cluster of the heading data in the r-th heading process. The larger the value of , the more similar the change trends between the two indicators. At the same time the smaller the value of

[0035]

[0036] Specifically, taking the heading data of the r-th heading process as an example in the embodiments of the present invention, clustering is performed on the heading data of the r-th heading process according to the above embodiments of the present invention to obtain a plurality of third clusters. For the index data of each index in the n-th third cluster (assuming that there is a negative correlation between the indexes in the n-th third cluster), the second interaction index value between its indexes can be calculated using the following formula:

[0037] In the above formula, represents the second interaction index value between various indicators in the nth cluster of the heading data during the rth heading process. represents the number of reference indicators in the nth cluster. represents the number of indicators affected by the ith reference indicator, that is, the fourth quantity of the influencing indicators that have a negative correlation with the reference indicators in the third cluster. represents a normalization function, which is used to perform normalization processing. represents and the third DTW distance between. represents the fluctuation curve of the ith reference indicator in the nth cluster of the heading data during the rth heading process. represents the fluctuation curve of the jth indicator affected by the ith reference indicator in the nth cluster of the heading data during the rth heading process. represents the fluctuation amplitude of the fluctuation curve of the ith reference indicator in the nth cluster of the heading data during the rth heading process. represents the fluctuation amplitude of the fluctuation curve of the jth indicator affected by the ith reference indicator in the nth cluster of the heading data during the rth heading process.

[0038] Furthermore, when determining the first interaction index value between various indicators during the heading process, the average value of the second interaction index values between various indicators in the heading process of each target cluster is determined as the first interaction index value.

[0039] Specifically, the embodiment of the present invention uses the following formula to calculate the first interaction index value between various indicators in the heading data of the rth heading process:

[0040] In the above formula, represents the first interaction index value between various indicators in the heading data during the rth heading process. N represents the number of clustering clusters formed by all indicators. represents the second interaction index value between various indicators in the nth cluster of the heading data during the rth heading process. It should be noted that if there is a negative correlation between various indicators, then in the above formula is replaced by .

[0041] S104, cluster the heading data according to the first interaction index values between various indicators in the heading data of the first cluster during different heading processes to obtain the second cluster.

[0042] Specifically, in the above embodiments of the present invention, the interaction index values between the various indicators in each heading data in the nth cluster are calculated. Since these multiple heading data are obtained when working in the same area, the fluctuation differences of the various indicators in the heading data of different times should be relatively small, and the distribution of the positive correlation indicators and negative correlation indicators should also be the same. Therefore, by analyzing the similarity degree of the first interaction index values between the various indicators in the heading data of different times in the nth cluster and the similarity of the index clustering results, the similarity degree value of the heading data of different times is determined, and the heading data is clustered based on this to obtain the second cluster.

[0043] Further, when clustering the heading data according to the first interaction index values between the various indicators in the heading data of the first cluster during different heading processes to obtain the second cluster, as an optional embodiment of the present invention, first, the matching degree value between the first clusters of the heading data of different heading processes is determined according to the indicators in the first cluster of the heading data during different heading processes; then, the matching quantity of the first clusters of the heading data during different heading processes is determined by using the matching degree value; then, according to the matching quantity, the quantity of the first clusters of the heading data during different heading processes, and the first interaction index values between the various indicators during different heading processes, the similarity degree value between the heading data of different heading processes is determined; finally, the heading data is clustered based on the similarity degree value between the heading data of different heading processes to obtain the second cluster.

[0044] Specifically, in the embodiments of the present invention, since the clustering results of the various indicators in the heading data during different heading processes in the nth cluster may be different, it is first necessary to match the clustering clusters in the clustering results of the various indicators in the heading data of different times. As an optional embodiment of the present invention, first, the fifth quantity of the same indicators in the first cluster of the heading data during different heading processes and the sixth quantity of the indicators in the first cluster of the heading data during different heading processes are determined; then, the third ratio between the fifth quantity and the sixth quantity is determined as the matching degree value.

[0045] More specifically, in the embodiments of the present invention, taking the heading data of the pth time and the heading data of the qth time in the nth cluster as an example, first, their clustering clusters are matched. Among them, the mathematical formula for the matching degree value between the ath cluster in the heading data of the pth heading process and the bth cluster in the heading data of the qth heading process is:

[0046] In the above formula, represents the matching degree value between the ath cluster in the heading data of the pth heading process and the bth cluster in the heading data of the qth heading process. Denotes the sixth quantity of the metrics in the ath cluster in the heading data of the qth heading process. Denotes the fifth quantity of the same metrics in the ath cluster in the heading data of the pth heading process and the bth cluster in the heading data of the qth heading process.

[0047] Where, if is greater than a predetermined threshold value, such as the predetermined threshold value is 0.5. If is greater than 0.5, it indicates a successful match. Otherwise, it indicates an unsuccessful match. If there are multiple clusters in the ath cluster of the pth heading data and the qth heading data that match, then the cluster with the largest is recorded as the matching cluster. Each cluster in the pth heading data in the nth cluster and each cluster in the qth heading data are matched according to the above method. Suppose a total of

[0048] matching clusters are matched after the matching is completed.

[0048] Further, when determining the similarity degree value, in an embodiment of the present invention, taking the heading data of the pth time and the heading data of the qth time in the nth cluster as an example, the following formula is specifically used to calculate the similarity degree value between the heading data of different heading processes:

[0049] In the above formula, denotes the similarity degree value between the heading data of the pth heading process and the heading data of the qth heading process. denotes the number of clustering clusters of the metrics in the heading data of the pth heading process. denotes the number of clustering clusters of the metrics in the heading data of the qth heading process. denotes the number of matching clustering clusters of the metrics in the heading data of the pth heading process and the clustering clusters of the metrics in the heading data of the qth heading process. denotes the first interaction index value between the metrics in the pth heading data. denotes the first interaction index value between the metrics in the qth heading data. The smaller the value of and the larger the value of , it indicates that the number of clustering clusters of the metrics in the pth heading data and the qth heading data is more similar, and at the same time, the metric data of the metrics in each cluster of the clustering result is also more similar, that is, it indicates that the clustering results of the metrics in the two heading data are more similar. At the same time, The smaller the value of , it indicates that the interaction indexes between the metrics in the pth heading data and the qth heading data are also similar. Therefore, The larger the value of

[0050] Further, according to the above embodiments of the present invention, the similarity degree value between the p-th course data and the q-th course data can be calculated. Then, the k-means method is used to cluster all M times of course data based on this similarity degree value, where the distance between different times of course data is quantified by to quantify, and the number of clustering clusters is determined by the silhouette coefficient method. Among them, the use of the k-means method for clustering and the use of the silhouette coefficient method to determine the number of clustering clusters can refer to the prior art, and the embodiments of the present invention will not be elaborated herein. Suppose there are a total of R second clusters after clustering is completed, and then they are sorted in ascending order according to the number of course data within the second cluster. Since the larger the number of course data within the cluster indicates that when the underwater robot works in this area, the change of the course data shows more characteristics of this cluster. Therefore, when performing abnormal data analysis later, the weight value of the course data within this cluster is higher.

[0051] S105. Calculate the abnormal score value of the abnormal data of the index data of each index in the second cluster.

[0052] Specifically, in the embodiments of the present invention, the LOF method can be used to calculate the abnormal data of the time series data of each index of each course data within each second cluster. The mean value of the LOF scores of all calculated abnormal data is recorded as the abnormal score value of the abnormal data of each index. It should be noted that the use of the LOF method to calculate abnormal data can refer to the prior art, and the embodiments of the present invention will not be elaborated herein.

[0053] S106. Determine the abnormal degree value of each index of the underwater robot according to the abnormal score value of each index, the number of the second cluster where the index data of the index is located, and the number of the second cluster, and determine the index with the abnormal degree value greater than the first threshold as the abnormal index.

[0054] Specifically, the first threshold can be determined according to the actual situation, and the value in the embodiments of the present invention is 0.7. The index with the abnormal degree value greater than 0.7 is recorded as the abnormal index, indicating that the index has an abnormality during the operation of the underwater robot, and professionals need to check the parts and functions corresponding to this index of the underwater robot to ensure the safe operation of the underwater robot.

[0055] Further, when calculating the abnormal degree value of each index of the underwater robot, as an optional embodiment of the present invention, first calculate the fourth ratio between the number of the second cluster where the index data of the index is located and the number of the second cluster; then determine the fourth product between the fourth ratio and the abnormal score value of the index as the abnormal degree value of the index.

[0056] Specifically, the embodiments of the present invention use the following formula to calculate the abnormal degree value:

[0057] In the formula, represents the abnormal degree value of the nth index during the operation of the underwater robot. represents the number of clustering clusters of all heading data. represents the number of the cluster where the mth heading data is located after sorting, and this number is a natural number greater than or equal to 1. represents the abnormal score value of the nth index in the mth heading data. The larger the value of the greater the abnormal degree value of the nth index during the operation of the underwater robot.

[0058] In the embodiment of the present invention, the interaction index value between various indexes during the operation of the underwater robot can be determined by analyzing the heading data of the underwater robot when working in the same area. Then, according to the interaction index value between the indexes of different heading data, the similarity between the heading data in different heading processes is determined, so as to cluster the heading data. Then, the abnormal score value of the abnormal data of the index data of each index in the clustering cluster is calculated, and the abnormal degree value of each index is calculated based on this. Thus, the monitoring of abnormal indexes is achieved. In this way, the normal fluctuation of the heading data caused by the change of the underwater environment can be avoided from being misjudged as abnormal, the detection accuracy of the abnormal detection of the underwater robot is improved, and thus the control accuracy of the underwater robot is improved.

[0059] Embodiment 2: Corresponding to the underwater robot heading data abnormal monitoring method provided in the above embodiment, based on the same technical concept, the embodiment of the present invention also provides an underwater robot heading data abnormal monitoring device. This underwater robot heading data abnormal monitoring system is used to execute the above underwater robot heading data abnormal monitoring method. Figure 2 is a schematic structural diagram of an underwater robot heading data abnormal monitoring device provided by an embodiment of the present invention, as Figure 2As shown in the figure. The device 200 includes: an acquisition module 201, configured to acquire the heading data of the underwater robot within a predetermined time period, where the heading data includes the index data of multiple indicators; a clustering module 202, configured to cluster the heading data to obtain multiple first clusters, and each first cluster includes the heading data of the same section during multiple heading processes of the underwater robot; a determination module 203, configured to determine the first interaction index value between each indicator during the heading process according to the index data of multiple indicators in the heading data within each first cluster; the clustering module 202 is further configured to cluster the heading data according to the first interaction index value between each indicator in the heading data in the first cluster during different heading processes to obtain a second cluster; a calculation module 204, configured to calculate the anomaly score value of the anomaly data of the index data of each indicator in the second cluster; the determination module 203 is further configured to determine the anomaly degree value of each indicator of the underwater robot according to the anomaly score value of each indicator, the number of the second cluster where the index data of the indicator is located, and the number of the second cluster, and determine the indicator with the anomaly degree value greater than the first threshold as an anomaly indicator.

[0060] Embodiment 3: Corresponding to the underwater robot heading data anomaly monitoring method provided in the above embodiment, based on the same technical concept, an embodiment of the present invention further provides an underwater robot heading data anomaly monitoring system, and this underwater robot heading data anomaly monitoring system is used to execute the above underwater robot heading data anomaly monitoring method. Figure 3 The structural schematic diagram of an underwater robot heading data anomaly monitoring system provided by an embodiment of the present invention is shown in Figure 3 the figure. The underwater robot heading data anomaly monitoring system may have relatively large differences due to different configurations or performances, and may include one or more processors 301 and a memory 302. The memory 302 is used to store a computer program that can run on the processor 301, and the processor 301 is used to execute the program stored in the memory 302 to implement each step in the method embodiment above. Figure 1 Among them, the memory 302 can be a short-term storage or a persistent storage. The application program stored in the memory 302 may include one or more modules (not shown in the figure), and each module may include a series of computer-executable instructions for the underwater robot heading data anomaly monitoring system.

[0061] Furthermore, the processor 301 can be set to communicate with the memory 302 and execute a series of computer-executable instructions in the memory 302 on the underwater robot heading data anomaly monitoring system. The underwater robot heading data anomaly monitoring system may further include one or more power supplies 303, one or more wired or wireless network interfaces 304, one or more input / output interfaces 305, and one or more keyboards 306.

[0062] Specifically, in this embodiment, the underwater robot heading data anomaly monitoring system includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the bus. The memory is used to store computer programs. The processor is used to execute the programs stored on the memory to implement the above Figure 1 steps in the method embodiments and has the beneficial effects of the above method embodiments. To avoid repetition, the embodiments of the present invention will not be described in detail here.

[0063] It should be noted that the underwater robot heading data anomaly monitoring system provided by the embodiments of the present invention and the underwater robot heading data anomaly monitoring method provided by the embodiments of the present invention are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned underwater robot heading data anomaly monitoring method and has the same or similar beneficial effects. The repeated parts will not be described again.

[0064] It should be noted that the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0065] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

Claims

1. A method for monitoring abnormal heading data of an underwater robot, characterized in that: The underwater robot heading data abnormality monitoring method comprises: Acquiring heading data of the underwater robot within a predetermined time period, wherein the heading data includes indicator data of multiple indicators; Clustering the heading data to obtain a plurality of first clusters, each of the first clusters including heading data of the same section during multiple heading processes of the underwater robot; Determine a first interaction index value between the indicators in the heading process according to the index data of the multiple indicators in the heading data in each of the first clusters; Clustering the heading data according to first interaction index values ​​between various indexes of the heading data in the first cluster in different heading processes to obtain a second cluster; Calculate anomaly scores of abnormal data of indicator data of each indicator in the second cluster; The abnormality degree value of each indicator of the underwater robot is determined according to the abnormality score value of each indicator, the number of the second cluster where the indicator data of the indicator is located, and the number of the second clusters, and the indicator whose abnormality degree value is greater than the first threshold is determined as an abnormal indicator.

2. The method for monitoring abnormal heading data of an underwater robot according to claim 1, characterized in that: The step of determining the first interaction index value between the indicators in the heading process according to the index data of the multiple indicators in the heading data in each of the first clusters includes: Obtaining a fluctuation curve of the indicator data of each indicator in each of the first clusters during multiple heading processes; Clustering the indicator data of each of the indicators according to the first DTW distances between the fluctuation curves to obtain a plurality of third clusters; Determine a target cluster in which there is a positive correlation or a negative correlation between indicators according to an average value of the second DTW distances of the indicator data of all indicators in the third cluster; Determine a second interaction index value between the indicators of the heading process in the target cluster according to a third DTW distance between the fluctuation curves of the indicators of the heading process in the target cluster and a fluctuation amplitude of the fluctuation curve of each indicator; The first interaction index value between the indicators in the heading process is determined according to the second interaction index value between the indicators in the heading process in each target cluster.

3. The method for monitoring abnormal heading data of an underwater robot according to claim 2, characterized in that: The step of determining the target cluster in which the indicators have a positive correlation or a negative correlation according to the average value of the second DTW distance of the indicator data of all the indicators in the third cluster comprises: When the average value of the second DTW distance is less than a second threshold, determining that there is a positive correlation between all indicators in the third cluster; When the average value of the second DTW distance is not less than the second threshold, it is determined that there is a negative correlation between all indicators in the third cluster.

4. The method for monitoring abnormal heading data of an underwater robot according to claim 2, characterized in that: When the target cluster is a third cluster with a positive correlation, determining the second interaction index value between the indicators of the heading process in the target cluster according to the third DTW distance between the fluctuation curves of the indicators of the heading process in the target cluster and the fluctuation amplitude of the fluctuation curve of each indicator includes: Calculating a first product between a first number of all indicators in the third cluster and a second number of influencing indicators that are positively correlated with the indicators in the third cluster; Calculating a third DTW distance between a fluctuation curve of the indicator in the target cluster and a fluctuation curve of an influencing indicator that is positively correlated with the indicator, and normalizing the third DTW distance to obtain a first normalized distance; Calculating an absolute value of a first difference between a fluctuation amplitude of a fluctuation curve of the indicator in the target cluster and a fluctuation amplitude of a fluctuation curve of an influencing indicator that is positively correlated with the indicator, and calculating a reciprocal of a first sum between the absolute value of the first difference and the preset value; Superimposing a second product of the first normalized distance of each of the indicators in the third cluster and the reciprocal of the first sum value to obtain a first superimposed value; A first ratio between the first superposition value and the first product is determined as the second interaction index value.

5. The method for monitoring abnormal heading data of an underwater robot according to claim 2, characterized in that: When the target cluster is a third cluster with a negative correlation, determining the second interaction index value between the indicators of the heading process in the target cluster according to the third DTW distance between the fluctuation curves of the indicators of the heading process in the target cluster and the fluctuation amplitude of the fluctuation curve of each indicator includes: calculating a third product between a third number of all indicators in the third cluster and a fourth number of influencing indicators that are negatively correlated with the indicators in the third cluster; Calculating a third DTW distance between a fluctuation curve of the indicator in the target cluster and a fluctuation curve of an influencing indicator that is negatively correlated with the indicator, and normalizing the third DTW distance to obtain a second normalized distance; Calculating the reciprocal of a second sum between the second normalized distance and a preset value; Calculating an absolute value of a second difference between a fluctuation amplitude of a fluctuation curve of the indicator in the target cluster and a fluctuation amplitude of a fluctuation curve of an influencing indicator that is negatively correlated with the indicator, and calculating a reciprocal of a third sum between the absolute value of the second difference and the preset value; Superimposing a fourth product of the reciprocal of the second sum of each of the indicators in the third cluster and the reciprocal of the third sum to obtain a second superimposed value; A second ratio between the second superposition value and the third product is determined as the second interaction index value.

6. The method for monitoring abnormal heading data of an underwater robot according to claim 2, characterized in that: Determining the first interaction index value between the indicators in the heading process according to the second interaction index value between the indicators in the heading process in each target cluster comprises: An average value of the second interaction index values ​​between the indicators of the heading process in each of the target clusters is determined as the first interaction index value.

7. The method for monitoring abnormal heading data of an underwater robot according to claim 1, characterized in that: The second cluster obtained by clustering the heading data according to the first interaction index value between the indexes of the heading data in the first cluster in different heading processes includes: Determining the matching degree values ​​between the first clusters of the heading data of different heading processes according to the indexes in the first clusters of the heading data of different heading processes; Determining the matching number of the first cluster of heading data in different heading processes by using the matching degree value; Determine the similarity value between the heading data of different heading processes according to the matching number, the number of first clusters of heading data in different heading processes and the first interaction index value between the indexes in different heading processes; The heading data are clustered based on similarity values ​​between the heading data of different heading processes to obtain a second cluster.

8. The method for monitoring abnormal heading data of an underwater robot according to claim 7, characterized in that: The step of determining the matching degree values ​​between the first clusters of the heading data of different heading processes according to the indexes in the first clusters of the heading data of different heading processes comprises: determining a fifth number of the same indicator in the first cluster of the heading data in different heading processes, and a sixth number of the indicator in the first cluster of the heading data in different heading processes; A third ratio between the fifth number and the sixth number is determined as the matching degree value.

9. The method for monitoring abnormal heading data of an underwater robot according to claim 1, characterized in that: Determining the abnormality level value of each indicator of the underwater robot according to the abnormality score value of each indicator, the number of the second cluster where the indicator data of the indicator is located, and the number of the second clusters includes: calculating a fourth ratio between the number of the second cluster where the indicator data of the indicator is located and the number of the second cluster; A fourth product between the fourth ratio and the abnormality score value of the indicator is determined as the abnormality degree value of the indicator.

10. An underwater robot heading data abnormality monitoring system, characterized in that: The underwater robot heading data anomaly monitoring system includes: a processor and a memory; wherein the memory is used to store a computer program that can be run on the processor; the processor is used to execute the program stored in the memory to implement the steps of the underwater robot heading data anomaly monitoring method as described in any one of claims 1-9.

Images