Integrated gateway operation monitoring method, device and equipment based on artificial intelligence

Abstract

The invention provides a fusion gateway operation monitoring method, device and equipment based on artificial intelligence, and relates to the technical field of artificial intelligence. The method comprises the following steps: firstly, acquiring hardware operation state data and hardware operation environment data of a target fusion gateway; secondly, respectively performing semantic feature mining operation on the hardware running state data and the hardware running environment data, and outputting hardware running state features corresponding to the hardware running state data and hardware running environment features corresponding to the hardware running environment data; performing focusing mining operation on the hardware operation state features based on the hardware operation environment features, and outputting state environment focusing features; and finally, evaluating the operation state based on the state environment focusing characteristics to obtain an operation state evaluation result so as to realize operation monitoring of the target fusion gateway. Based on the above content, the problem that the reliability of operation monitoring of the fusion gateway is relatively low in the prior art can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology. Specifically, it relates to a method, device, and equipment for monitoring the operation of a fusion gateway based on artificial intelligence. Background Art

[0002] With the rapid development of information technology and the continuous innovation of artificial intelligence, as a key device connecting multiple networks and systems, the fusion gateway has been widely used in various information processing and communication scenarios, such as the Internet of Things, smart home, industrial automation, and other fields. The main function of the fusion gateway is to effectively integrate devices with different communication protocols and data formats to achieve unified management and transmission of information.

[0003] For example, an IPTV fusion gateway is a device that integrates an optical modem and a set-top box in the same casing and can simultaneously achieve functions such as fiber optic Internet access (such as OTT service, Over-The-Top, which refers to the transmission of video, audio, and other media content over the Internet without the need for a traditional TV service provider's transmission pipeline) and watching TV (such as IPTV service, Internet Protocol Television, which uses the Internet protocol to transmit TV programs and video content. Different from traditional cable or satellite TV, IPTV uses the Internet as the transmission medium and allows users to watch TV content through a network connection).

[0004] However, in the prior art, there are still some deficiencies in the operation monitoring of the fusion gateway. For example, most of the current fusion gateway operation status monitoring methods rely on traditional threshold detection and regular inspection means, and it is difficult to obtain and analyze the hardware operation status in real time and accurately. Therefore, there is a problem that the reliability of the operation monitoring of the fusion gateway is relatively low. Summary of the Invention

[0005] In view of this, the purpose of the present application is to provide a method, device, and equipment for monitoring the operation of a fusion gateway based on artificial intelligence to improve the problem of relatively low reliability of the operation monitoring of the fusion gateway existing in the prior art.

[0006] To achieve the above purpose, the present application adopts the following technical solutions:

[0007] A method for monitoring the operation of a fusion gateway based on artificial intelligence, comprising:

[0008] Obtaining the hardware operation status data and hardware operation environment data of a target fusion gateway, where the hardware operation status data is used to reflect the operation status of at least one hardware device in the target fusion gateway, and the hardware operation environment data includes the environment data of the at least one hardware device, and the environment data at least includes temperature data and power supply data;

[0009] Through the semantic feature mining unit included in the operating status evaluation model, semantic feature mining operations are respectively performed on the hardware operating status data and the hardware operating environment data, and the hardware operating status features corresponding to the hardware operating status data and the hardware operating environment features corresponding to the hardware operating environment data are output. Among them, the operating status evaluation model belongs to a neural network model, and the operating status evaluation model further includes a focus mining unit and an operating status evaluation unit;

[0010] Through the focus mining unit, a focus mining operation is performed on the hardware operating status features based on the hardware operating environment features, and status environment focus features are output. Among them, the status environment focus features are used to reflect the semantic information associated with the hardware operating environment features mined from the hardware operating status features;

[0011] Through the operating status evaluation unit, the operating status of the target fusion gateway is evaluated based on the status environment focus features, and the operating status evaluation result corresponding to the target fusion gateway is obtained, so as to realize the operating monitoring of the target fusion gateway.

[0012] In a preferred selection of the present application, in the above-mentioned method for monitoring the operation of a fusion gateway based on artificial intelligence, the step of respectively performing semantic feature mining operations on the hardware operating status data and the hardware operating environment data through the semantic feature mining unit included in the operating status evaluation model, and outputting the hardware operating status features corresponding to the hardware operating status data and the hardware operating environment features corresponding to the hardware operating environment data includes:

[0013] Through the semantic embedding subunit in the semantic feature mining unit included in the operating status evaluation model, semantic embedding operations are respectively performed on the hardware operating status data and the hardware operating environment data, and the operating status embedding features corresponding to the hardware operating status data and the operating environment embedding features corresponding to the hardware operating environment data are output;

[0014] Through the first deep mining subunit in the semantic feature mining unit, a deep mining operation is performed on the operating status embedding features, and the hardware operating status features corresponding to the hardware operating status data are output;

[0015] Through the first deep mining subunit in the semantic feature mining unit, a deep mining operation is performed on the operating environment embedding features, and the hardware operating environment features corresponding to the hardware operating environment data are output.

[0016] In a preferred option of the present application, in the above-mentioned method for monitoring the operation of an AI-based fusion gateway, the step of deeply mining the operation state embedding features through the first deep mining subunit in the semantic feature mining unit and outputting the hardware operation state features corresponding to the hardware operation state data includes:

[0017] Load the operation state embedding features into the first deep mining subunit in the semantic feature mining unit, where the first deep mining subunit has a first mapping branch and a second mapping branch, and both mapping branches include linear mapping and non-linear mapping;

[0018] Perform self-attention processing on the operation state embedding features to form operation state attention features corresponding to the operation state embedding features;

[0019] Perform a mapping operation on the operation state attention features through the first mapping branch to form first operation state mapping features corresponding to the operation state attention features;

[0020] Perform a mapping operation on the operation state attention features through the second mapping branch to form second operation state mapping features corresponding to the operation state attention features;

[0021] Perform cross-attention processing on the first operation state mapping features and the second operation state mapping features to form the hardware operation state features corresponding to the hardware operation state data.

[0022] In a preferred option of the present application, in the above-mentioned method for monitoring the operation of an AI-based fusion gateway, the step of deeply mining the operation environment embedding features through the first deep mining subunit in the semantic feature mining unit and outputting the hardware operation environment features corresponding to the hardware operation environment data includes:

[0023] Load the operation environment embedding features into the first deep mining subunit in the semantic feature mining unit, where the first deep mining subunit has a third mapping branch and a fourth mapping branch, and both mapping branches include linear mapping and non-linear mapping;

[0024] Perform self-attention processing on the operation environment embedding features to form operation environment attention features corresponding to the operation environment embedding features;

[0025] Perform a mapping operation on the operation environment attention features through the third mapping branch to form first operation environment mapping features corresponding to the operation environment attention features;

[0026] Through the fourth mapping branch, perform a mapping operation on the running environment attention feature to form a second running environment mapping feature corresponding to the running environment attention feature;

[0027] Perform cross-attention processing on the first running environment mapping feature and the second running environment mapping feature to form a hardware running environment feature corresponding to the hardware running environment data.

[0028] In a preferred selection of the present application, in the above-mentioned method for monitoring the operation of an AI-based fusion gateway, the step of, through the focusing and mining unit, performing a focusing and mining operation on the hardware running state feature based on the hardware running environment feature and outputting a state environment focusing feature includes:

[0029] Through the first focusing branch in the focusing and mining unit, perform a first focusing and mining operation on the hardware running state feature based on the hardware running environment feature and output a first focusing feature;

[0030] Through the second focusing branch in the focusing and mining unit, perform a second focusing and mining operation on the hardware running state feature based on the hardware running environment feature and output a second focusing feature;

[0031] Perform a mean calculation on the first focusing feature, the second focusing feature, and the hardware running state feature, and output a state environment focusing feature.

[0032] In a preferred selection of the present application, in the above-mentioned method for monitoring the operation of an AI-based fusion gateway, the step of, through the first focusing branch in the focusing and mining unit, performing a first focusing and mining operation on the hardware running state feature based on the hardware running environment feature and outputting a first focusing feature includes:

[0033] Through the first spatial transformation matrix of the first focusing branch in the focusing and mining unit, perform a first spatial transformation operation on the hardware running environment feature, so that the hardware running environment feature is transformed from the current semantic space to the semantic space where the hardware running state feature is located, thereby obtaining a corresponding running environment transformation feature;

[0034] Determine the association parameter distribution between the running environment transformation feature and the hardware running state feature, and after normalizing the association parameter distribution, based on the normalized association parameter distribution, perform a weighting process on the hardware running state feature and output a first focusing feature.

[0035] In a preferred option of the present application, in the above-mentioned method for monitoring the operation of an AI-based fusion gateway, the step of performing a second focused mining operation on the hardware operation state features based on the hardware operation environment features through the second focused branch in the focused mining unit and outputting second focused features includes:

[0036] Performing a second space conversion operation on the hardware operation state features through the second space conversion matrix of the second focused branch in the focused mining unit, so that the hardware operation state features are converted from the current semantic space to the semantic space where the hardware operation environment features are located, thereby obtaining corresponding operation state conversion features;

[0037] Determining the associated parameter distribution between the hardware operation environment features and the operation state conversion features, and after normalizing the associated parameter distribution, performing a weighted processing on the operation state conversion features based on the normalized associated parameter distribution to output second focused features.

[0038] In a preferred option of the present application, in the above-mentioned method for monitoring the operation of an AI-based fusion gateway, it further includes:

[0039] Obtaining a hardware operation state sample and a hardware operation environment sample;

[0040] Performing semantic feature mining operations on the hardware operation state sample and the hardware operation environment sample respectively through the semantic feature mining unit included in the candidate operation state evaluation model, and outputting the hardware operation state features corresponding to the hardware operation state sample and the hardware operation environment features corresponding to the hardware operation environment sample;

[0041] Performing a focused mining operation on the hardware operation state features corresponding to the hardware operation state sample based on the hardware operation environment features corresponding to the hardware operation environment sample through the focused mining unit included in the candidate operation state evaluation model, and outputting a state environment focused feature sample;

[0042] Evaluating the operation state of the corresponding fusion gateway based on the state environment focused feature sample through the operation state evaluation unit included in the candidate operation state evaluation model, and obtaining corresponding operation state evaluation data;

[0043] Updating the model parameters of the candidate operation state evaluation model based on the error between the operation state evaluation data and the operation state label of the corresponding fusion gateway until the error converges, thereby obtaining a trained operation state evaluation model.

[0044] The present application also provides an AI-based fusion gateway operation monitoring device, including:

[0045] An operation data acquisition module, configured to acquire hardware operation status data and hardware operation environment data of a target fusion gateway, where the hardware operation status data is used to reflect the operation status of at least one hardware device in the target fusion gateway, and the hardware operation environment data includes environment data of the at least one hardware device, and the environment data at least includes temperature data and power supply data;

[0046] A semantic feature mining module, configured to perform semantic feature mining operations on the hardware operation status data and the hardware operation environment data respectively through a semantic feature mining unit included in an operation status evaluation model, and output a hardware operation status feature corresponding to the hardware operation status data and a hardware operation environment feature corresponding to the hardware operation environment data, where the operation status evaluation model belongs to a neural network model, and the operation status evaluation model further includes a focus mining unit and an operation status evaluation unit;

[0047] A focus mining module, configured to perform a focus mining operation on the hardware operation status feature based on the hardware operation environment feature through the focus mining unit, and output a status environment focus feature, where the status environment focus feature is used to reflect semantic information associated with the hardware operation environment feature mined from the hardware operation status feature;

[0048] An operation status evaluation module, configured to evaluate the operation status of the target fusion gateway based on the status environment focus feature through the operation status evaluation unit, and obtain an operation status evaluation result corresponding to the target fusion gateway, so as to implement operation monitoring of the target fusion gateway.

[0049] On the basis of the above, the present application further provides an electronic device, including:

[0050] A memory, configured to store a computer program;

[0051] A processor connected to the memory, configured to execute the computer program stored in the memory to implement the above-mentioned method for monitoring the operation of a fusion gateway based on artificial intelligence.

[0052] The method, device, and equipment for monitoring the operation of a fusion gateway based on artificial intelligence provided by this application first obtain the hardware operation status data and hardware operation environment data of the target fusion gateway; secondly, perform semantic feature mining operations on the hardware operation status data and hardware operation environment data respectively, and output the hardware operation status features corresponding to the hardware operation status data and the hardware operation environment features corresponding to the hardware operation environment data. Then, based on the hardware operation environment features, perform a focused mining operation on the hardware operation status features to output the status environment focused features; finally, evaluate the operation status based on the status environment focused features to obtain the operation status evaluation result, so as to realize the operation monitoring of the target fusion gateway. Based on the above content, on the one hand, the powerful learning ability of the neural network model is utilized, so that the reliability of the operation status evaluation can be improved to a certain extent. On the other hand, since after obtaining the hardware operation status features, a focused mining will be performed based on the hardware operation environment features, it is possible to further mine semantic features with higher representation accuracy, that is, the status environment focused features. Therefore, when performing status evaluation based on the status environment focused features, a relatively high reliability can be achieved, thereby improving the problem of relatively low reliability in the operation monitoring of the fusion gateway existing in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] To make the above objects, features, and advantages of this application more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, makes a detailed description as follows.

[0054] Figure 1 It is a block diagram of the electronic device provided by the embodiment of this application.

[0055] Figure 2 It is a schematic flowchart of the method for monitoring the operation of a fusion gateway based on artificial intelligence provided by the embodiment of this application.

[0056] Figure 3 It is a schematic diagram of semantic feature mining provided by the embodiment of this application.

[0057] Figure 4 It is a schematic diagram of focused mining provided by the embodiment of this application.

[0058] Figure 5 It is a block schematic diagram of the device for monitoring the operation of a fusion gateway based on artificial intelligence provided by the embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this application. Apparently, the described embodiments are only a part rather than all of the embodiments of this application. Usually, the components of the embodiments of this application described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0060] Therefore, the detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts fall within the scope of protection of this application.

[0061] As Figure 1 shown, the embodiments of this application provide an electronic device. Among them, the electronic device may include a memory, a processor, and a fusion gateway operation monitoring device based on artificial intelligence.

[0062] Specifically, the memory and the processor are directly or indirectly electrically connected to achieve data transmission or interaction. For example, the memory and the processor may be electrically connected through one or more communication buses or signal lines. The fusion gateway operation monitoring device based on artificial intelligence includes at least one software functional module stored in the memory in the form of software or firmware. The processor is used to execute the executable computer program stored in the memory, such as the software functional modules and computer programs included in the fusion gateway operation monitoring device based on artificial intelligence, to implement the fusion gateway operation monitoring method based on artificial intelligence provided by the embodiments of this application.

[0063] Optionally, the memory may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc.

[0064] Further, the processor may be a general - purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), a System on Chip (SoC), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0065] It can be understood that Figure 1 the structure shown is only schematic, and the electronic device may also include more or fewer components than those shown Figure 1 in it, or have a configuration different from that shown Figure 1 For example, it may also include a communication unit for information interaction with other devices.

[0066] Combined with Figure 2 , an embodiment of the present application also provides an artificial - intelligence - based fusion gateway operation monitoring method applicable to the above - mentioned electronic device. Among them, the method steps defined by the processes related to the artificial - intelligence - based fusion gateway operation monitoring method can be implemented by the electronic device. The following will elaborate on Figure 2 the specific process shown in detail.

[0067] Step S110, obtain the hardware operation status data and hardware operation environment data of the target fusion gateway.

[0068] In an embodiment of the present application, the electronic device may obtain the hardware operation status data and hardware operation environment data of the target fusion gateway. Among them, the hardware operation status data is used to reflect the operation status of at least one hardware device in the target fusion gateway, and the hardware operation environment data includes the environment data of the at least one hardware device, and this environment data at least includes temperature data and power supply data. Exemplarily, the hardware status data may include CPU usage rate (high CPU load may cause the device to respond slowly or crash), memory usage rate (memory leak or overload may cause the device performance to decline), disk read speed, disk write speed, disk usage rate (high disk utilization may mean a large load on the hard disk, which may affect performance), and disk bad - track condition, etc. In other implementation manners, it may also include the operation status of more or different hardware. In addition, the hardware operation environment data may refer to the temperature data and power supply data corresponding to the foregoing hardware devices. Among them, the power supply data may refer to voltage, current (fluctuations in voltage and current will also affect the status of the corresponding hardware), etc. In other implementation manners, it may also include other data. In addition, both the hardware operation status data and the hardware operation environment data can be represented in text form.

[0069] Step S120: Through the semantic feature mining unit included in the operating state evaluation model, perform semantic feature mining operations on the hardware operating state data and the hardware operating environment data respectively, and output the hardware operating state features corresponding to the hardware operating state data and the hardware operating environment features corresponding to the hardware operating environment data.

[0070] In the embodiment of the present application, after obtaining the hardware operating state data and the hardware operating environment data, the electronic device can perform semantic feature mining operations on the hardware operating state data and the hardware operating environment data respectively through the semantic feature mining unit included in the operating state evaluation model, and output the hardware operating state features corresponding to the hardware operating state data and the hardware operating environment features corresponding to the hardware operating environment data (exemplarily, in the embodiment of the present application, the manifestation form of each feature can be a vector, that is, the potential semantic information is mined and represented in the form of a vector). Among them, the operating state evaluation model belongs to a neural network model (which can be formed by learning based on corresponding samples and labels on the basis of an initial neural network model), and the operating state evaluation model further includes a focused mining unit and an operating state evaluation unit.

[0071] Step S130: Through the focused mining unit, perform a focused mining operation on the hardware operating state features based on the hardware operating environment features, and output state-environment focused features.

[0072] In the embodiment of the present application, after mining the hardware operating environment features and the hardware operating state features, the electronic device can perform a focused mining operation on the hardware operating state features based on the hardware operating environment features through the focused mining unit, and output state-environment focused features. Among them, the state-environment focused features are used to reflect the semantic information associated with the hardware operating environment features mined from the hardware operating state features. That is to say, since the semantic information in one dimension is associated with the semantic information in another dimension, it can be shown that the corresponding semantic information is important, so it can be mined, thereby realizing the focusing of important semantic information and improving the semantic representation ability of the state-environment focused features.

[0073] Step S140: Through the operating state evaluation unit, evaluate the operating state of the target fusion gateway based on the state-environment focused features, and obtain the operating state evaluation result corresponding to the target fusion gateway, so as to realize the operating monitoring of the target fusion gateway.

[0074] In an embodiment of the present application, after obtaining the state environment focused feature, the electronic device may, through the operating state evaluation unit, evaluate the operating state of the target fusion gateway based on the state environment focused feature to obtain an operating state evaluation result corresponding to the target fusion gateway, so as to implement the operating monitoring of the target fusion gateway. Exemplarily, the operating state evaluation unit may perform a fully connected process on the state environment focused feature to obtain a corresponding fully connected feature, where the size of the fully connected feature may be 1*n, and n may be the number of operating state types (such as 3 operating state types: good, average, and poor), and then, based on the softmax function, map and output the fully connected feature to obtain the probabilities of each operating state type. Finally, the operating state type with the highest probability may be used as the operating state evaluation result.

[0075] Based on the above content, on the one hand, the powerful learning ability of the neural network model is utilized, so that the reliability of the operating state evaluation can be improved to a certain extent. On the other hand, since after obtaining the hardware operating state feature, focused mining is further performed based on the hardware operating environment feature, semantic features with higher representation accuracy, that is, state environment focused features, can be further mined. Therefore, when performing state evaluation based on the state environment focused feature, a relatively high reliability can be achieved, thereby improving the problem of relatively low reliability of the operating monitoring of the fusion gateway existing in the prior art.

[0076] In an embodiment of the present application, it should also be noted that for step S120, the specific manner of separately performing semantic feature mining operations on the hardware operating state data and the hardware operating environment data is not limited and can be selected accordingly according to actual needs.

[0077] For example, in an alternative embodiment, in order to capture more detailed information during the semantic feature mining process, step S120 described above may further include step S121, step S122, and step S123, and the specific content of each step is as follows (in combination with Figure 3 ).

[0078] Step S121: Respectively perform semantic embedding operations on the hardware operating state data and the hardware operating environment data through the semantic embedding subunit in the semantic feature mining unit included in the operating state evaluation model, and output the operating state embedding feature corresponding to the hardware operating state data and the operating environment embedding feature corresponding to the hardware operating environment data.

[0079] In an embodiment of the present application, the semantic embedding subunit in the semantic feature mining unit included in the operating state evaluation model can perform semantic embedding operations on the hardware operating state data and the hardware operating environment data respectively, and output the operating state embedding features corresponding to the hardware operating state data and the operating environment embedding features corresponding to the hardware operating environment data. Among them, the semantic embedding subunit can be a word embedding model. In this way, the corresponding data can be segmented and embedded, and then the embedding vectors of each word can be combined such as concatenation or cascading, so as to form the corresponding operating state embedding features and operating environment embedding features.

[0080] Step S122: Perform a deep mining operation on the operating state embedding features through the first deep mining subunit in the semantic feature mining unit, and output the hardware operating state features corresponding to the hardware operating state data.

[0081] In an embodiment of the present application, after obtaining the operating state embedding features, a deep mining operation can be performed on the operating state embedding features through the first deep mining subunit in the semantic feature mining unit, and the hardware operating state features corresponding to the hardware operating state data can be output. In this way, the deep semantic information in the operating state embedding features can be captured, thereby improving the semantic representation ability of the hardware operating state features.

[0082] Step S123: Perform a deep mining operation on the operating environment embedding features through the first deep mining subunit in the semantic feature mining unit, and output the hardware operating environment features corresponding to the hardware operating environment data.

[0083] In an embodiment of the present application, after obtaining the operating environment features, a deep mining operation can be performed on the operating environment embedding features through the first deep mining subunit in the semantic feature mining unit, and the hardware operating environment features corresponding to the hardware operating environment data can be output. In this way, the deep semantic information in the operating environment embedding features can be captured, thereby improving the semantic representation ability of the hardware operating environment features.

[0084] It can be understood that in the above step S122, the specific manner of performing the deep mining operation on the operating state embedding features is not limited. For example, in an alternative embodiment, in order to improve the reliability of the deep mining operation, the above step S122 may include the following content:

[0085] First, the operating state embedding features can be loaded into the first deep mining subunit in the semantic feature mining unit, where the first deep mining subunit has a first mapping branch and a second mapping branch, and both mapping branches include linear mapping and non-linear mapping;

[0086] Secondly, self-attention processing can be performed on the running state embedding features to form running state attention features corresponding to the running state embedding features. Based on this, since the running state embedding features contain state semantic information of multiple hardware devices, and there is generally a correlation between the states of hardware devices, self-attention processing of the running state embedding features can associate and fuse the state semantic information of different hardware devices.

[0087] Then, through the first mapping branch, a mapping operation can be performed on the running state attention features to form first running state mapping features corresponding to the running state attention features. Exemplarily, the first mapping branch may include a first weight matrix, a first bias matrix, and a first non-linear mapping function. In this way, after multiplying the first weight matrix by the running state attention features to obtain corresponding weighted features, adding the weighted features to the first bias matrix, and then, based on the first non-linear mapping function (such as tanh (hyperbolic tangent function)), activating the added matrix to obtain the first running state mapping features. In this way, the richness of the captured semantic information can be improved to a certain extent.

[0088] Moreover, through the second mapping branch, a mapping operation can be performed on the running state attention features to form second running state mapping features corresponding to the running state attention features. Exemplarily, the second mapping branch may include a second weight matrix, a second bias matrix, and a second non-linear mapping function. In this way, after multiplying the second weight matrix by the running state attention features to obtain corresponding weighted features, adding the weighted features to the second bias matrix, and then, based on the second non-linear mapping function (such as tanh (hyperbolic tangent function)), activating the added matrix to obtain the second running state mapping features. Based on this, different mappings of the running state attention features can be realized, so as to capture different semantic information in the running state attention features.

[0089] Finally, cross-attention processing can be performed on the first running state mapping features and the second running state mapping features to form hardware running state features corresponding to the hardware running state data. Based on this, since the first running state mapping features and the second running state mapping features respectively represent different semantic information in the running state attention features, cross-attention processing can realize the association and fusion of different semantic information, that is, by performing different mappings on the features and based on these two mapping features for cross-attention, it can help the model capture information from multiple dimensions and improve the diversity, flexibility, and robustness of feature expression.

[0090] It can be understood that in the above step S123, the specific manner of deeply mining the running environment embedding features is not limited. For example, in an alternative embodiment, in order to improve the reliability of the deep mining operation, the above step S123 may include the following content:

[0091] First, the running environment embedding features can be loaded into the first deep mining subunit in the semantic feature mining unit. Among them, the first deep mining subunit has a third mapping branch and a fourth mapping branch, and both mapping branches include linear mapping and non-linear mapping;

[0092] Second, self-attention processing can be performed on the running environment embedding features to form running environment attention features corresponding to the running environment embedding features. Based on this, since the running environment embedding features contain environmental semantic information of multiple hardware devices, and there is generally a correlation between the environments of hardware devices, therefore, by performing self-attention processing on the running environment embedding features, the environmental semantic information of different hardware devices can be associated and fused;

[0093] Then, through the third mapping branch, mapping operation can be performed on the running environment attention features to form the first running environment mapping features corresponding to the running environment attention features. Exemplarily, the third mapping branch may include a third weight matrix, a third bias matrix, and a third non-linear mapping function. In this way, the third weight matrix and the running environment attention features can be multiplied to obtain the corresponding weighted features, and then the weighted features and the third bias matrix can be added. Then, based on the third non-linear mapping function (such as tanh (hyperbolic tangent function)), the added matrix can be activated to obtain the first running environment mapping features. In this way, the richness of the captured semantic information can be improved to a certain extent;

[0094] And, through the fourth mapping branch, mapping operation can be performed on the running environment attention features to form the second running environment mapping features corresponding to the running environment attention features. Exemplarily, the fourth mapping branch may include a fourth weight matrix, a fourth bias matrix, and a fourth non-linear mapping function. In this way, the fourth weight matrix and the running environment attention features can be multiplied to obtain the corresponding weighted features, and then the weighted features and the fourth bias matrix can be added. Then, based on the fourth non-linear mapping function (such as tanh (hyperbolic tangent function)), the added matrix can be activated to obtain the second running environment mapping features. In this way, the richness of the captured semantic information can be improved to a certain extent;

[0095] Finally, cross-attention processing can be performed on the first operating environment mapping feature and the second operating environment mapping feature to form a hardware operating environment feature corresponding to the hardware operating environment data. Based on this, since the first operating environment mapping feature and the second operating environment mapping feature respectively reflect different semantic information in the operating environment attention feature, cross-attention processing can achieve the associated fusion of different semantic information. That is, by performing different mappings on the features and performing cross-attention based on these two mapping features, the model can be helped to capture information from multiple dimensions, improving the diversity, flexibility, and robustness of feature expression.

[0096] In the embodiment of the present application, it should also be noted that for step S130, the specific manner of performing the focused mining operation on the hardware operating state feature based on the hardware operating environment feature is not limited and can be selected according to actual needs.

[0097] For example, in an alternative embodiment, in order to fully mine the associated information in the hardware operating environment feature and the hardware operating state feature through the focused mining operation, the above step S130 may further include step S131, step S132, and step S133. The specific content of each step is as follows (in combination with Figure 4 ).

[0098] Step S131: Through the first focusing branch in the focused mining unit, perform a first focused mining operation on the hardware operating state feature based on the hardware operating environment feature, and output a first focused feature.

[0099] In the embodiment of the present application, a first focused mining operation can be performed on the hardware operating state feature based on the hardware operating environment feature through the first focusing branch in the focused mining unit, and a first focused feature can be output.

[0100] Step S132: Through the second focusing branch in the focused mining unit, perform a second focused mining operation on the hardware operating state feature based on the hardware operating environment feature, and output a second focused feature.

[0101] In the embodiment of the present application, a second focused mining operation is performed on the hardware operating state feature based on the hardware operating environment feature through the second focusing branch in the focused mining unit, and a second focused feature is output. Exemplarily, the first focusing branch and the second focusing branch are different, that is, the first focused mining operation and the second focused mining operation are different, so that different focused mining operations can be performed on the hardware operating environment feature and the hardware operating state feature, thereby mining different associated semantic information, thus enriching the richness and diversity of semantic information.

[0102] Step S133: Calculate the mean of the first focus feature, the second focus feature, and the hardware operation state feature, and output a state environment focus feature.

[0103] In the embodiment of the present application, in order to avoid the problem that some important semantic information that is not related is lost due to excessive attention to associated semantic information in the focus mining operation, the mean of the first focus feature, the second focus feature, and the hardware operation state feature can be calculated, that is, the fusion of multiple semantic information is realized, so as to output a state environment focus feature with richer semantics.

[0104] It can be understood that in the above step S131, the specific manner of performing the first focus mining operation on the hardware operation state feature based on the hardware operation environment feature is not limited. For example, in an alternative embodiment, in order to enable reliable fusion of features in different semantic spaces during the focus mining process, the above step S131 may include:

[0105] First, the first spatial transformation matrix of the first focus branch in the focus mining unit can be used to perform a first spatial transformation operation on the hardware operation environment feature, so that the hardware operation environment feature is transformed from the current semantic space to the semantic space where the hardware operation state feature is located, thereby obtaining a corresponding operation environment transformation feature; exemplarily, the first spatial transformation matrix and the hardware operation environment feature can be multiplied to obtain the operation environment transformation feature;

[0106] Second, the association parameter distribution between the operation environment transformation feature and the hardware operation state feature can be determined (exemplarily, the dot product between the operation environment transformation feature and the transposed feature of the hardware operation state feature can be calculated), and after normalizing the association parameter distribution, the hardware operation state feature is weighted based on the normalized association parameter distribution, and the first focus feature is output; based on this, due to the dot product calculation, the similarity of each parameter between the operation environment transformation feature and the hardware operation state feature can be determined, and then, weighting can be performed based on this similarity, so that the weights corresponding to similar features are larger, so that key attention and characterization can be realized; in addition, since the hardware operation environment feature will be transformed from the current semantic space to the semantic space where the hardware operation state feature is located, subsequent association fusion can be performed in the same semantic space, thereby improving the fusion accuracy.

[0107] It can be understood that in the above step S132, the specific manner of performing the second focused mining operation on the hardware operation state features based on the hardware operation environment features is not limited. For example, in an alternative embodiment, in order to enable reliable fusion of features in different semantic spaces during the focused mining process, the above step S132 may include:

[0108] First, the second space conversion matrix of the second focused branch in the focused mining unit can be used to perform a second space conversion operation on the hardware operation state features, so that the hardware operation state features are converted from the current semantic space to the semantic space where the hardware operation environment features are located, thereby obtaining corresponding operation state conversion features; exemplarily, the second space conversion matrix and the hardware operation state features can be multiplied to obtain the operation state conversion features;

[0109] Secondly, determine the associated parameter distribution between the hardware operation environment features and the operation state conversion features (exemplarily, the dot product between the hardware operation environment features and the transposed features of the operation state conversion features can be calculated), and after normalizing the associated parameter distribution, based on the normalized associated parameter distribution, perform a weighting process on the operation state conversion features to output the second focused features; based on this, due to the dot product calculation, the similarity of each parameter between the hardware operation environment features and the operation state conversion features can be determined, and then, based on this similarity, a weighting process can be performed to make the weights of similar features larger. Therefore, key attention and characterization can be achieved; in addition, since the hardware operation state features will be converted from the current semantic space to the semantic space where the hardware operation environment features are located, subsequent associated fusion can be performed in the same semantic space, thereby improving its fusion accuracy.

[0110] Based on this, since the semantic features corresponding to the operation state have a direct characterization effect, while the semantic features corresponding to the operation environment have an auxiliary characterization effect, when performing the step of "calculating the mean of the first focused feature, the second focused feature, and the hardware operation state features and outputting the state environment focused feature", that is, when performing step S133, a weighted mean calculation can be performed, where the corresponding weighting coefficients can be sorted from largest to smallest as the hardware operation state features, the first focused feature, and the second focused feature; that is, the weighting coefficient corresponding to the hardware operation state features can be greater than the weighting coefficient corresponding to the first focused feature, and the weighting coefficient corresponding to the first focused feature can be greater than the weighting coefficient corresponding to the second focused feature. In this way, sufficient characterization of the original semantic information and the associated semantic information can be achieved.

[0111] In the embodiments of the present application, it should also be noted that in order to ensure that the operation state evaluation model has better evaluation ability, corresponding training can be performed in advance so that the mapping relationship between the samples and the corresponding labels can be learned. Therefore, the operation monitoring method for the fusion gateway based on artificial intelligence may further include the following content:

[0112] First, hardware operation state samples and hardware operation environment samples can be obtained, and the relevant explanations in step S110 can be referred to;

[0113] Second, through the semantic feature mining unit included in the candidate operation state evaluation model, semantic feature mining operations can be respectively performed on the hardware operation state samples and the hardware operation environment samples, and the hardware operation state features corresponding to the hardware operation state samples and the hardware operation environment features corresponding to the hardware operation environment samples can be output, and the relevant explanations in step S120 can be referred to;

[0114] Then, through the focusing mining unit included in the candidate operation state evaluation model, a focusing mining operation can be performed on the hardware operation state features corresponding to the hardware operation state samples based on the hardware operation environment features corresponding to the hardware operation environment samples, and a state environment focusing feature sample can be output, and the relevant explanations in step S130 can be referred to;

[0115] After that, through the operation state evaluation unit included in the candidate operation state evaluation model, the operation state of the corresponding fusion gateway can be evaluated based on the state environment focusing feature sample, and the corresponding operation state evaluation data can be obtained, and the relevant explanations in step S140 can be referred to;

[0116] Finally, based on the error between the operation state evaluation data and the operation state label of the corresponding fusion gateway (such as cross-entropy error, etc. For example, the operation state evaluation data can be the probability distribution of each operation state type, such as (0.1, 0.2, 0.7), and the operation state label can also be the probability distribution of each operation state type, such as (0, 0, 1). In this way, the cross-entropy between (0.1, 0.2, 0.7) and (0, 0, 1) can be calculated), the model parameters of the candidate operation state evaluation model can be updated (that is, the parameters are updated along the direction of reducing this error) until this error converges (such as this error is less than the preset error or the reduction amplitude of this error is less than the preset value), so as to obtain the trained operation state evaluation model, that is, learn the reliable mapping relationship between the samples and the labels, so that in subsequent applications, this mapping relationship can be used to achieve reliable evaluation.

[0117] Combined with Figure 5, an embodiment of the present application further provides an artificial intelligence-based fusion gateway operation monitoring device applicable to the above-mentioned electronic device. Among them, the artificial intelligence-based fusion gateway operation monitoring device may include an operation data acquisition module, a semantic feature mining module, a focused mining module, and an operation status evaluation module.

[0118] Specifically, the operation data acquisition module can be used to acquire the hardware operation status data and the hardware operation environment data of the target fusion gateway. Among them, the hardware operation status data is used to reflect the operation status of at least one hardware device in the target fusion gateway, and the hardware operation environment data includes the environment data of the at least one hardware device, and the environment data at least includes temperature data and power supply data. In an embodiment of the present application, the operation data acquisition module can be used to execute Figure 2 the step S110 shown. For the relevant content of the operation data acquisition module, reference can be made to the description of step S110 above.

[0119] Specifically, the semantic feature mining module can be used to perform semantic feature mining operations on the hardware operation status data and the hardware operation environment data respectively through the semantic feature mining unit included in the operation status evaluation model, and output the hardware operation status features corresponding to the hardware operation status data and the hardware operation environment features corresponding to the hardware operation environment data. Among them, the operation status evaluation model belongs to a neural network model, and the operation status evaluation model further includes a focused mining unit and an operation status evaluation unit. In an embodiment of the present application, the semantic feature mining module can be used to execute Figure 2 the step S120 shown. For the relevant content of the semantic feature mining module, reference can be made to the description of step S120 above.

[0120] Specifically, the focused mining module can be used to perform a focused mining operation on the hardware operation status features based on the hardware operation environment features through the focused mining unit, and output a status environment focused feature, where the status environment focused feature is used to reflect the semantic information associated with the hardware operation environment features mined from the hardware operation status features. In an embodiment of the present application, the focused mining module can be used to execute Figure 2 the step S130 shown. For the relevant content of the focused mining module, reference can be made to the description of step S130 above.

[0121] Specifically, the operation status evaluation module can be used to evaluate the operation status of the target fusion gateway based on the status environment focused features through the operation status evaluation unit, so as to obtain the operation status evaluation result corresponding to the target fusion gateway, so as to realize the operation monitoring of the target fusion gateway. In the embodiment of the present application, the operation status evaluation module can be used to execute Figure 2 The steps S140 shown. For the relevant content of the operation status evaluation module, reference can be made to the description of step S140 above.

[0122] In the embodiment of the present application, corresponding to the above-mentioned method for monitoring the operation of a fusion gateway based on artificial intelligence applied to the electronic device, a computer-readable storage medium is also provided. A computer program is stored in the computer-readable storage medium, and when the computer program runs, it executes each step of the method for monitoring the operation of a fusion gateway based on artificial intelligence.

[0123] Among them, the steps executed when the foregoing computer program runs will not be elaborated one by one here, and reference can be made to the foregoing explanation of the method for monitoring the operation of a fusion gateway based on artificial intelligence.

[0124] In summary, for the method, device and equipment for monitoring the operation of a fusion gateway based on artificial intelligence provided by the present application, first, obtain the hardware operation status data and hardware operation environment data of the target fusion gateway; secondly, perform semantic feature mining operations on the hardware operation status data and hardware operation environment data respectively, and output the hardware operation status features corresponding to the hardware operation status data and the hardware operation environment features corresponding to the hardware operation environment data. Then, perform a focused mining operation on the hardware operation status features based on the hardware operation environment features, and output the status environment focused features; finally, perform an operation status evaluation based on the status environment focused features to obtain an operation status evaluation result, so as to realize the operation monitoring of the target fusion gateway. Based on the above content, on the one hand, the powerful learning ability of the neural network model is utilized, so that the reliability of the operation status evaluation can be improved to a certain extent. On the other hand, since after obtaining the hardware operation status features, focused mining will be performed based on the hardware operation environment features, so that semantic features with higher representation accuracy, that is, status environment focused features, can be further mined. Therefore, when performing status evaluation based on the status environment focused features, a relatively high reliability can be achieved, thereby improving the problem of relatively low reliability of the operation monitoring of the fusion gateway existing in the prior art.

[0125] In several embodiments provided by the embodiments of the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device and method embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order from that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0126] In addition, each functional module in various embodiments of the present application may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.

[0127] If the above functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, an electronic device, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes. It should be noted that in this article, the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article, or device. Without further limitations, the element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article, or device including the said element.

[0128] The above are only the preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A method for monitoring the operation of a fusion gateway based on artificial intelligence, characterized in that: include: Acquire hardware operation status data and hardware operation environment data of the target converged gateway, wherein the hardware operation status data is used to reflect the operation status of at least one hardware device in the target converged gateway, and the hardware operation environment data includes environment data of the at least one hardware device, and the environment data includes at least temperature data and power supply data; Through the semantic feature mining unit included in the operation status evaluation model, the semantic feature mining operation is performed on the hardware operation status data and the hardware operation environment data respectively, and the hardware operation status features corresponding to the hardware operation status data and the hardware operation environment features corresponding to the hardware operation environment data are output, wherein the operation status evaluation model belongs to a neural network model, and the operation status evaluation model also includes a focus mining unit and an operation status evaluation unit; By means of the focused mining unit, a focused mining operation is performed on the hardware operation state feature based on the hardware operation environment feature, and a state environment focused feature is output, wherein the state environment focused feature is used to reflect semantic information associated with the hardware operation environment feature mined from the hardware operation state feature; The operating status evaluation unit evaluates the operating status of the target fusion gateway based on the state environment focus feature, and obtains the operating status evaluation result corresponding to the target fusion gateway, so as to realize the operation monitoring of the target fusion gateway.

2. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to claim 1 is characterized in that: The semantic feature mining unit included in the operation status evaluation model performs semantic feature mining operations on the hardware operation status data and the hardware operation environment data respectively, and outputs the hardware operation status features corresponding to the hardware operation status data and the hardware operation environment features corresponding to the hardware operation environment data, including: Through the semantic embedding subunit in the semantic feature mining unit included in the operation status evaluation model, the semantic embedding operation is performed on the hardware operation status data and the hardware operation environment data respectively, and the operation status embedding feature corresponding to the hardware operation status data and the operation environment embedding feature corresponding to the hardware operation environment data are output; Performing a deep mining operation on the operation status embedded feature through the first deep mining subunit in the semantic feature mining unit, and outputting a hardware operation status feature corresponding to the hardware operation status data; Through the first deep mining sub-unit in the semantic feature mining unit, a deep mining operation is performed on the operating environment embedded features to output the hardware operating environment features corresponding to the hardware operating environment data.

3. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to claim 2 is characterized in that: The step of performing a deep mining operation on the operation status embedded feature through the first deep mining subunit in the semantic feature mining unit to output the hardware operation status feature corresponding to the hardware operation status data includes: Loading the operating state embedding feature into a first deep mining subunit in the semantic feature mining unit, wherein the first deep mining subunit has a first mapping branch and a second mapping branch, and both mapping branches include a linear mapping and a nonlinear mapping; Performing self-attention processing on the running state embedded feature to form a running state attention feature corresponding to the running state embedded feature; Through the first mapping branch, a mapping operation is performed on the running state attention feature to form a first running state mapping feature corresponding to the running state attention feature; Through the second mapping branch, a mapping operation is performed on the running state attention feature to form a second running state mapping feature corresponding to the running state attention feature; Cross-attention processing is performed on the first operating state mapping feature and the second operating state mapping feature to form a hardware operating state feature corresponding to the hardware operating state data.

4. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to claim 2 is characterized in that: The step of performing a deep mining operation on the operating environment embedded features through the first deep mining subunit in the semantic feature mining unit to output the hardware operating environment features corresponding to the hardware operating environment data includes: Loading the operating environment embedding feature into a first deep mining subunit in the semantic feature mining unit, wherein the first deep mining subunit has a third mapping branch and a fourth mapping branch, and both mapping branches include a linear mapping and a nonlinear mapping; Performing self-attention processing on the operating environment embedded feature to form an operating environment attention feature corresponding to the operating environment embedded feature; Through the third mapping branch, a mapping operation is performed on the operating environment attention feature to form a first operating environment mapping feature corresponding to the operating environment attention feature; Through the fourth mapping branch, a mapping operation is performed on the operating environment attention feature to form a second operating environment mapping feature corresponding to the operating environment attention feature; Cross-attention processing is performed on the first operating environment mapping feature and the second operating environment mapping feature to form a hardware operating environment feature corresponding to the hardware operating environment data.

5. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to claim 1 is characterized in that: The step of performing a focused mining operation on the hardware operating state feature based on the hardware operating environment feature by the focused mining unit and outputting the state environment focused feature comprises: Through a first focusing branch in the focusing mining unit, a first focusing mining operation is performed on the hardware operating state feature based on the hardware operating environment feature, and a first focusing feature is output; Through a second focusing branch in the focusing mining unit, a second focusing mining operation is performed on the hardware operating state feature based on the hardware operating environment feature, and a second focusing feature is output; The first focusing feature, the second focusing feature and the hardware operating state feature are averaged and a state environment focusing feature is output.

6. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to claim 5 is characterized in that: The step of performing a first focused mining operation on the hardware operating state feature based on the hardware operating environment feature through the first focused branch in the focused mining unit to output a first focused feature includes: By using the first space conversion matrix of the first focusing branch in the focusing mining unit, a first space conversion operation is performed on the hardware operating environment feature, so that the hardware operating environment feature is converted from the current semantic space to the semantic space where the hardware operating state feature is located, thereby obtaining a corresponding operating environment conversion feature; Determine the associated parameter distribution between the operating environment conversion feature and the hardware operating status feature, and after normalizing the associated parameter distribution, perform weighted processing on the hardware operating status feature based on the normalized associated parameter distribution to output a first focused feature.

7. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to claim 5 is characterized in that: The step of performing a second focused mining operation on the hardware operating state feature based on the hardware operating environment feature through the second focused branch in the focused mining unit to output a second focused feature comprises: By using the second space conversion matrix of the second focusing branch in the focused mining unit, a second space conversion operation is performed on the hardware operation state feature, so that the hardware operation state feature is converted from the current semantic space to the semantic space where the hardware operation environment feature is located, thereby obtaining a corresponding operation state conversion feature; Determine the associated parameter distribution between the hardware operating environment feature and the operating state transition feature, and, after normalizing the associated parameter distribution, perform weighted processing on the operating state transition feature based on the normalized associated parameter distribution to output a second focused feature.

8. The method for monitoring the operation of a fusion gateway based on artificial intelligence according to any one of claims 1 to 7, characterized in that: Also includes: Obtain hardware operating status samples and hardware operating environment samples; Through the semantic feature mining unit included in the candidate operation status evaluation model, the semantic feature mining operation is performed on the hardware operation status sample and the hardware operation environment sample respectively, and the hardware operation status feature corresponding to the hardware operation status sample and the hardware operation environment feature corresponding to the hardware operation environment sample are output; By using the focused mining unit included in the candidate operating status evaluation model, a focused mining operation is performed on the hardware operating status features corresponding to the hardware operating status samples based on the hardware operating environment features corresponding to the hardware operating environment samples, and a state environment focused feature sample is output; By using the operation status evaluation unit included in the candidate operation status evaluation model, the operation status of the corresponding fusion gateway is evaluated based on the state environment focus feature sample to obtain corresponding operation status evaluation data; Based on the error between the running status evaluation data and the running status label of the corresponding fusion gateway, the model parameters of the candidate running status evaluation model are updated until the error converges, thereby obtaining a trained running status evaluation model.

9. An artificial intelligence-based fusion gateway operation monitoring device, characterized in that: include: An operation data acquisition module, used to acquire hardware operation status data and hardware operation environment data of the target converged gateway, wherein the hardware operation status data is used to reflect the operation status of at least one hardware device in the target converged gateway, and the hardware operation environment data includes environment data of the at least one hardware device, and the environment data includes at least temperature data and power supply data; A semantic feature mining module, configured to perform semantic feature mining operations on the hardware operating state data and the hardware operating environment data respectively through a semantic feature mining unit included in an operating state evaluation model, and output hardware operating state features corresponding to the hardware operating state data and hardware operating environment features corresponding to the hardware operating environment data, wherein the operating state evaluation model belongs to a neural network model, and the operating state evaluation model further includes a focus mining unit and an operating state evaluation unit; A focused mining module, configured to perform a focused mining operation on the hardware operating state feature based on the hardware operating environment feature through the focused mining unit, and output a state environment focused feature, wherein the state environment focused feature is used to reflect semantic information associated with the hardware operating environment feature mined from the hardware operating state feature; The operation status evaluation module is used to evaluate the operation status of the target fusion gateway based on the state environment focus feature through the operation status evaluation unit, and obtain the operation status evaluation result corresponding to the target fusion gateway to realize the operation monitoring of the target fusion gateway.

10. An electronic device, characterized in that: include: Memory for storing computer programs; A processor connected to the memory is used to execute a computer program stored in the memory to implement the artificial intelligence-based fusion gateway operation monitoring method described in any one of claims 1 to 8.

Images