Embedded system bus communication data monitoring and intelligent diagnosis method and platform

Abstract

The application relates to the technical field of Internet of Things, and provides a kind of embedded system bus communication data monitoring and intelligent diagnosis method and platform. The operating parameters of the embedded system are acquired and adaptive adjustment is performed to generate adjustment data;Through the mixed time sequence determination window length, the first feature extraction is performed on the adjustment data to generate time sequence segmentation data;From the time-frequency domain and the spatial domain, the second feature extraction is performed on the time sequence segmentation data to obtain low-dimensional space-time features;The low-dimensional space-time features are input into the mixed network obtained by pre-training to generate a fault probability distribution;According to the fault probability distribution, a reliability factor is determined, and a diagnosis result is output in combination with the fault probability distribution and the reliability factor. The technical scheme of the application forms low-dimensional space-time features by fusing multi-dimensional features of time-frequency domain and spatial domain, intelligently analyzes the fault probability distribution through the mixed network, and finally outputs a high-reliability diagnosis result in combination with the reliability factor, significantly improving the system operation stability and diagnosis accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things, in particular to a kind of embedded system bus communication data monitoring and intelligent diagnosis method and platform. BACKGROUND

[0002] Embedded system is widely used in automobile, industrial control and aerospace fields, and its bus communication is the key link for the cooperative work between devices, such as Controller Area Network (CAN) and Local Interconnect Network (LIN). With the increase of system complexity, the amount of bus data grows exponentially, and traditional manual monitoring cannot meet the real-time and accuracy requirements. For example, modern automobile bus can transmit tens of thousands of data per second, and any communication abnormality may cause system failure or even safety accidents.

[0003] Traditional bus monitoring relies on preset threshold or rule base, and can only detect known fault modes, with high false negative rate in complex scenarios. For example, occasional communication interruption of a certain automobile bus caused by electromagnetic interference cannot be captured by traditional methods, which may cause misjudgment or delay maintenance. Therefore, there is a problem of low precision in the prior art of embedded system bus monitoring. SUMMARY

[0004] The present application provides an embedded system bus communication data monitoring and intelligent diagnosis method and platform, which can at least partially solve the problem of low precision in embedded system bus monitoring.

[0005] Other characteristics and advantages of the present application will become apparent from the following detailed description, or will be learned partly through practice of the present application.

[0006] According to one aspect of the present application, an embedded system bus communication data monitoring and intelligent diagnosis method is provided, comprising: obtaining running parameters of an embedded system, adaptively adjusting the running parameters to generate adjustment data, the running parameters including physical parameters, data packet parameters and state parameters of the embedded system; performing first feature extraction on the adjustment data by mixed timing determination of the split window length to generate timing segmentation data; performing second feature extraction on the timing segmentation data from time-frequency domain and spatial domain to obtain low-dimensional space-time features; inputting the low-dimensional space-time features into a mixed network trained in advance to generate a fault probability distribution; determining a reliability factor according to the fault probability distribution, and outputting a diagnosis result combining the fault probability distribution and the reliability factor.

[0007] In the present application, based on the foregoing scheme, the running parameters of the embedded system are obtained, the running parameters are adaptively adjusted, and adjustment data is generated, including: obtaining historical parameters and real-time running parameters of the embedded system; based on the historical parameters and the running parameters, the running data is adaptively adjusted to generate adjustment data.

[0008] In the present application, based on the foregoing scheme, the running parameters of the embedded system are obtained, the running parameters are adaptively adjusted, and adjustment data is generated, including: obtaining historical parameters and real-time running parameters of the embedded system; based on the historical parameters and the running parameters, the running data is adaptively adjusted to generate adjustment data.

[0009] In the present application, based on the foregoing scheme, the running parameters of the embedded system are obtained, the running parameters are adaptively adjusted, and adjustment data is generated, including: obtaining historical parameters and real-time running parameters of the embedded system; based on the historical parameters and the running parameters, the running data is adaptively adjusted to generate adjustment data. For:

[0010]

[0011] Among them, respectively represent the mean and standard deviation of the historical parameters in the preset time, respectively represent the voltage value of the current sampling period and the current noise power, represents the reference noise power, represents the noise sensitivity coefficient.

[0012] In the present application, based on the foregoing scheme, the running parameters of the embedded system are obtained, the running parameters are adaptively adjusted, and adjustment data is generated, including: obtaining historical parameters and real-time running parameters of the embedded system; based on the historical parameters and the running parameters, the running data is adaptively adjusted to generate adjustment data.

[0013] In the present application, based on the foregoing scheme, the second feature extraction is performed on the time series segmentation data from the time-frequency domain and the spatial domain respectively to obtain low-dimensional space-time features, including: determining a spatial factor based on a preset scale parameter and a translation parameter, and extracting an energy spectrum feature of the time-frequency domain from the time series segmentation data based on the spatial factor; constructing a topological relationship based on the layout information of the bus node, and generating a spatial domain feature based on the topological relationship; and fusing the energy spectrum feature and the spatial domain feature to obtain the low-dimensional space-time features.

[0014] In the present application, based on the foregoing scheme, the low-dimensional space-time features are input into a hybrid network trained in advance to generate a fault probability distribution, including: inputting the low-dimensional space-time features into a long short-term memory network trained in advance to generate a time series feature; inputting the low-dimensional space-time features into a graph neural network trained in advance to generate a graph feature; fusing the time series feature and the graph feature to generate a fused feature; and mapping the fused feature to a fault probability distribution through a fully connected layer.

[0015] In the present application, based on the foregoing scheme, the reliable factor is determined according to the fault probability distribution, and the diagnostic result is output in combination with the fault probability distribution and the reliable factor, including: determining a probability mean and a probability variance according to the fault probability distribution; determining a reliable factor according to the probability mean and the probability variance; and outputting a diagnostic result in combination with the fault probability distribution and the reliable factor.

[0016] According to an aspect of the present application, an embedded system bus communication data monitoring and intelligent diagnosis platform is provided, including:

[0017] An acquisition unit is configured to acquire running parameters of an embedded system, perform adaptive adjustment on the running parameters, generate adjustment data, and the running parameters include physical parameters, data packet parameters and state parameters of the embedded system;

[0018] A segmentation unit is configured to perform first feature extraction on the adjustment data through a mixed time series determined segmentation window length to generate time series segmentation data;

[0019] A feature unit is configured to perform second feature extraction on the time series segmentation data from the time-frequency domain and the spatial domain respectively to obtain low-dimensional space-time features;

[0020] A probability unit is configured to input the low-dimensional space-time features into a hybrid network trained in advance to generate a fault probability distribution;

[0021] A diagnosis unit is configured to determine a reliable factor according to the fault probability distribution, and output a diagnostic result in combination with the fault probability distribution and the reliable factor.

[0022] According to one aspect of this application, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the embedded system bus communication data monitoring and intelligent diagnosis method as described in the above embodiments.

[0023] According to one aspect of this application, an electronic device is provided, comprising: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the embedded system bus communication data monitoring and intelligent diagnostic method as described in the above embodiments.

[0024] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the embedded system bus communication data monitoring and intelligent diagnostic method provided in the various optional implementations described above.

[0025] In this application's technical solution, the operating parameters of the embedded system are acquired, and the operating parameters are adaptively adjusted to generate adjustment data. A first feature extraction is performed on the adjustment data using a segmentation window length determined by hybrid time series, generating time-series segmented data. A second feature extraction is performed on the time-frequency domain and spatial domain respectively to obtain low-dimensional spatiotemporal features. These low-dimensional spatiotemporal features are input into a pre-trained hybrid network to generate a fault probability distribution. A reliability factor is determined based on the fault probability distribution, and a diagnostic result is output by combining the fault probability distribution and the reliability factor. This application's technical solution generates precise adjustment data by dynamically acquiring and adaptively adjusting the physical, data packet, and state parameters of the embedded system; extracts time-series features using a hybrid time-series segmentation window, and fuses multi-dimensional features from the time-frequency and spatial domains to form low-dimensional spatiotemporal features; generates a fault probability distribution through intelligent analysis using a hybrid network; and finally outputs a high-reliability diagnostic result by combining the reliability factor. This achieves intelligent processing from parameter monitoring to fault diagnosis, significantly improving system stability and diagnostic accuracy.

[0026] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0027] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0028] Figure 1 The flowchart illustrating an embedded system bus communication data monitoring and intelligent diagnostic method in one embodiment of this application is shown.

[0029] Figure 2 The flowchart illustrating the generation of low-dimensional spatiotemporal features is shown in one embodiment of this application.

[0030] Figure 3 The illustration shows a schematic diagram of an embedded system bus communication data monitoring and intelligent diagnostic platform in one embodiment of this application.

[0031] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0032] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art.

[0033] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0034] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0035] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0036] The implementation details of the technical solution of this application are described below:

[0037] Figure 1 A flowchart illustrating an embedded system bus communication data monitoring and intelligent diagnostic method according to an embodiment of this application is shown. (Refer to...) Figure 1 As shown, the embedded system bus communication data monitoring and intelligent diagnosis method includes at least steps S110 to S150, which are described in detail below:

[0038] S110, Obtain the operating parameters of the embedded system, adaptively adjust the operating parameters, and generate adjustment data. The operating parameters include the physical parameters, data packet parameters, and status parameters of the embedded system.

[0039] In one embodiment of this application, the operating parameters of the embedded system are acquired, including real-time acquisition of physical parameters (such as temperature, voltage, and power consumption), data packet parameters (such as transmission rate, packet size, and latency), and status parameters (such as operating mode, error log, and load). Then, the operating parameters are dynamically analyzed through an adaptive algorithm based on deep learning, and the parameter configuration is automatically adjusted according to the current system status and environmental conditions. Finally, optimized adjustment data is generated to improve system stability and efficiency. The entire process does not require manual intervention and is completed entirely within the embedded system.

[0040] In one embodiment of this application, obtaining the operating parameters of an embedded system, adaptively adjusting the operating parameters, and generating adjustment data includes:

[0041] Obtain historical and real-time operating parameters of the embedded system;

[0042] Based on the historical parameters and the operating parameters, the operating data is adaptively adjusted to generate adjusted data.

[0043] In one embodiment of this application, the physical parameters of the embedded system bus are acquired by real-time capture of voltage and current waveforms using smart sensors. A sensor network or Internet of Things (IoT) is built using these smart sensors, with the sensor network chip operating at a microsecond-level sampling rate. In the IoT, data is transmitted via an IoT chip, establishing an IoT control system and a fieldbus control system. Simultaneously, common-mode noise is suppressed using differential input to avoid signal distortion caused by environmental electromagnetic interference. The acquired raw signal passes through a dynamic gain adjustment module, which automatically adjusts the amplification factor based on the current noise intensity. When a sudden increase in noise is detected, the gain is reduced to prevent signal saturation; in low-noise environments, the gain is increased to capture weak signal characteristics. This process ensures that the physical signal reflects the true state of the bus while maintaining a sufficient signal-to-noise ratio for subsequent analysis by adapting to changes in environmental noise in real time.

[0044] In one embodiment of this application, the voltage value and current noise power of the current sampling period are determined based on the operating parameters, and the mean and standard deviation of historical parameters within a preset time period are determined based on historical parameters. Then, based on the mean and standard deviation of historical parameters within the preset time period, the voltage value of the current sampling period, and the current noise power, the operating data is adaptively adjusted to generate adjusted data. for:

[0045]

[0046] in, These represent the mean and standard deviation of the historical parameters within a preset time period, respectively. These represent the voltage value and the current noise power during the current sampling period, respectively. Indicates the reference noise power. This represents the noise sensitivity coefficient. The above process, by adjusting the dynamic gain, improves the signal-to-noise ratio in environments with strong interference.

[0047] In one embodiment of this application, a dedicated hardware parser is used to capture protocol layer data packet parameters, such as those of a CAN bus controller, to capture bus communication content in real time. During execution, the parser identifies the start flag, identifier, data segment, and checksum of the data packet and converts them into a structured data format. For example, in the CAN protocol, the capture process strictly follows bit timing rules to ensure accurate alignment of the sampling points for each data bit. The principle is to ensure the real-time performance and protocol compliance of data capture through hardware-level parsing, avoiding delays or errors that may be introduced by software parsing. A timestamp is appended to the captured data packets to align them with physical signals and system state data, providing a basis for subsequent correlation analysis.

[0048] In one embodiment of this application, the embedded system's status parameters, such as CPU load and memory utilization, are collected in real time using performance counters built into the embedded system. During execution, a monitoring module periodically reads hardware registers, such as the synchronous phasor measurement unit (PMU) of an Advanced RISC Machine (ARM) architecture, to count the number of instruction execution cycles and memory page table usage. For example, CPU load is calculated by determining the percentage of idle cycles, and memory utilization is calculated by the ratio of allocated memory pages to total memory pages. The principle is to use hardware-level performance indicators to directly reflect system resource consumption, avoiding the additional overhead that software-level statistics may introduce.

[0049] After acquiring the operating parameters, the physical parameters, data packet parameters, and status parameters are synchronously stored to form a multi-dimensional dataset, providing a causal correlation basis for diagnosing bus communication anomalies. In this embodiment, the three-stage data acquisition uses a hardware-level time synchronization mechanism, such as the IEEE 1588 protocol, to ensure timing consistency, aligning the physical parameters, data packet parameters, and status parameters to the same time base. After acquisition, lightweight preprocessing is performed to finally output a structured multimodal dataset, providing complete input for subsequent time-series segmentation and feature extraction steps.

[0050] The above process, by dynamically integrating historical parameters and real-time operating parameters, intelligently adjusts the system's physical parameters, data packet parameters, and status parameters, enabling the adjusted data to more accurately reflect the system's current operating status. This provides a more adaptable data source for subsequent feature extraction, enhancing the real-time performance and accuracy of monitoring and diagnosis.

[0051] S120, using the segmentation window length determined by the mixed time series, the first feature is extracted from the adjusted data to generate time series segmentation data.

[0052] In one embodiment of this application, the segmentation window length is first dynamically determined based on hybrid time series analysis. The window range is flexibly defined by fusing the feature change trends of different time dimensions. Then, the time series information in the adjusted data is scanned in multiple dimensions and pattern recognition is performed to extract representative time series feature segments. Finally, these feature segments are integrated into structured time series segmentation data according to time correlation to fully preserve the dynamic evolution law of the system's operating state.

[0053] In one embodiment of this application, by using a segmentation window length determined through hybrid temporal sequencing, a first feature extraction is performed on the adjusted data to generate temporal segmentation data, including:

[0054] The window length for segmenting the data is determined by the frame interval of the bus protocol and the response time in the adjusted data.

[0055] Based on the window length, the adjusted data is subjected to first feature extraction to generate time-series segmentation data.

[0056] In one embodiment of the application, the starting point of a communication event is identified by parsing protocol layer data packets, such as the flag bits of the CAN protocol. For example, when a data packet start flag is detected, a feature extraction operation is immediately triggered. In the absence of protocol events, a sliding window is used to scan physical signals and system status data; optionally, the window length is preset to 80% of the protocol frame interval to capture potential abnormal patterns.

[0057] In one embodiment of the application, the window parameters are determined by the weights and entropy values ​​corresponding to each protocol field in the bus protocol. for:

[0058]

[0059] in, These represent the weight and entropy value corresponding to each protocol field, respectively, where i and k represent the identifier and total number of protocol fields, respectively.

[0060] Then, based on the frame interval of the bus protocol and the response time in the adjusted data, the window length L for segmenting the data is determined as follows:

[0061]

[0062] in, This indicates the frame interval of the bus protocol, such as the standard frame interval of the CAN protocol as defined in the protocol specification, which is 120μs; This represents the system response time, which can be calculated from the operating parameters of the embedded system. For example, when the CPU load exceeds 70%, the system response time increases to 50μs.

[0063] In one embodiment of the application, based on the adjusted window length L, synchronized time segments are extracted from the physical parameters, data packet parameters, and status parameters of the embedded system in the adjusted data; this is known as first feature extraction, resulting in time-series segmented data. During time segment extraction, data continuity is ensured through timestamp alignment and sequence identification.

[0064] Through the extraction process described above, physical parameters, data packet parameters, and state parameters are assigned a unified time base, ensuring time alignment of data across different modes. For example, with IEEE 1588 synchronization, the start and end time errors of segments are controlled within 1 μs. Simultaneously, each segment is assigned a unique sequence identifier, arranged according to the original data stream, preserving temporal causality. For instance, a segment containing both data packet transmission and a sudden increase in CPU load will be marked as part of the same sequence for subsequent analysis to correlate communication and system state anomalies.

[0065] In this embodiment, real-time monitoring of protocol events and sliding window states is triggered; threshold calculation dynamically adjusts the window length according to the protocol and system; fragment extraction segments data according to the calculation results and synchronizes them; the results are output to generate a structured time-series feature set. For example, when the bus malfunctions due to interference, the protocol event triggers segmentation immediately, extends the window length according to the current CPU load (high), extracts a complete fragment containing the error packet and system load, and finally outputs it for in-depth analysis to ensure that feature extraction is both accurate and efficient.

[0066] The above process dynamically determines the segmentation window length by combining the bus protocol frame interval and data response time, and performs time-series slicing on the adjusted data to generate time-series segmented data. This process can accurately capture key time window features during system operation, avoid feature omissions or redundancy caused by fixed windows, and improve the time-series targeting of feature extraction.

[0067] S130, perform second feature extraction on the time-series segmented data in the time-frequency domain and spatial domain respectively to obtain low-dimensional spatiotemporal features.

[0068] In one embodiment of this application, time-frequency domain analysis is used to capture the dynamic change patterns of time-series segmented data in the time and frequency dimensions, such as periodic fluctuations or abrupt changes. At the same time, spatial domain analysis is combined to mine the distribution patterns and correlation characteristics of the data in the spatial structure. Finally, the two types of features are fused and compressed into low-dimensional spatiotemporal features to comprehensively reflect the essential laws of the system state in both spatiotemporal dimensions.

[0069] like Figure 2 As shown, in one embodiment of this application, a second feature extraction is performed on the time-series segmented data in both the time-frequency domain and the spatial domain to obtain low-dimensional spatiotemporal features, including:

[0070] S210, Based on preset scale parameters and translation parameters, determine spatial factors, and extract time-frequency domain energy spectrum features from the time-series segmented data based on the spatial factors;

[0071] S220, Based on the deployment information of bus nodes, construct topological relationships, and generate spatial domain features based on the topological relationships;

[0072] S230, the energy spectrum features and the spatial domain features are fused to obtain low-dimensional spatiotemporal features.

[0073] In one embodiment of this application, the spatial factor is determined based on preset scale parameters and translation parameters. :

[0074]

[0075] Where a and b represent preset scale parameters and translation parameters, respectively. The scale parameter is used to control the frequency resolution, and the translation parameter is used to control the position. Let represent the wavelet function, and t represent the current time. The above process observes low-frequency, long-period changes on a large scale, such as a slow increase in bus load, and high-frequency, short-time changes on a small scale, such as sudden communication pulses. Through multi-scale decomposition, the energy distribution characteristics of the signal in time and frequency are extracted, preserving the details of the signal's instantaneous changes. This fully characterizes the dynamic characteristics of the bus signal, avoiding the loss of time-domain information inherent in traditional Fourier transforms.

[0076] In one embodiment of this application, time-frequency domain energy spectrum features are extracted from the time-series segmented data based on the spatial factor. for:

[0077]

[0078] in, This represents the time-series segmentation data corresponding to time t.

[0079] In one embodiment of this application, a topological relationship is constructed based on the deployment information of bus nodes, and spatial domain features are generated based on the topological relationship. Specifically, devices on the bus, such as sensors, are abstracted as nodes, and the physical communication links between devices are abstracted as edges, thus constructing a bus node communication graph. Then, a graph convolutional network is used to analyze the graph structure. The graph convolutional network learns the position and role of each node in the communication graph, such as a central node or an edge node, by aggregating information from neighboring nodes. Similar to identifying key figures in social network analysis, graph convolution can identify core devices that communicate frequently on the bus. By capturing the topological dependencies between bus nodes, spatial domain features are obtained, revealing the impact of spatial distribution on communication patterns.

[0080] In one embodiment of this application, the energy spectrum features in the time-frequency domain and the spatial domain features are fused to obtain low-dimensional spatiotemporal features. The energy spectrum features obtained from time-frequency domain analysis are integrated with the spatial domain features obtained from spatial domain analysis to obtain low-dimensional spatiotemporal features. The time-frequency features reflect when and how the signal changes, while the spatial features reflect who is communicating with whom. After fusion, the low-dimensional spatiotemporal features include both signal dynamics, such as burst communication, and communication paths, such as a device frequently sending data to multiple nodes. Through the above process, a more comprehensive feature representation is generated, avoiding the information loss associated with single-dimensional analysis.

[0081] The above process extracts energy spectrum features in the time-frequency domain using scaling and translation parameters to capture the dynamic changes of the signal in the time and frequency dimensions; in the spatial domain, it generates spatial features based on the bus node topology to reflect the association patterns between nodes. After fusion and dimensionality reduction, an efficient and comprehensive feature representation is generated that preserves the time-frequency characteristics of the signal while integrating the spatial distribution information of the nodes, providing strong support for subsequent tasks and achieving the collaborative expression of multi-dimensional features.

[0082] S140, The low-dimensional spatiotemporal features are input into the pre-trained hybrid network to generate a fault probability distribution.

[0083] In one embodiment of this application, low-dimensional spatiotemporal features are input into a pre-trained hybrid network. This network integrates multiple neural network structures based on machine learning, such as convolutional layers to capture local spatial patterns and recurrent layers to analyze temporal dependencies. It performs multi-dimensional analysis and integration of features and finally outputs the probability distribution of each potential fault type of the system, which intuitively reflects the likelihood of different fault modes occurring.

[0084] In one embodiment of this application, the low-dimensional spatiotemporal features are input into a pre-trained hybrid network to generate a fault probability distribution, including:

[0085] The low-dimensional spatiotemporal features are input into a pre-trained long short-term memory network to generate temporal features;

[0086] The low-dimensional spatiotemporal features are input into a pre-trained graph neural network to generate graph features;

[0087] The temporal features and the graph features are fused to generate fused features;

[0088] The fused features are mapped to a fault probability distribution through a fully connected layer.

[0089] In one embodiment of this application, the low-dimensional spatiotemporal features are input into a pre-trained Long Short-Term Memory (LSTM) network to generate temporal features. In this embodiment, the gating mechanism of the LSTM network captures the temporal variation patterns of signals, such as voltage fluctuations and communication frequency changes, thus addressing the gradient vanishing problem in long sequences. For example, abnormal fluctuations in bus voltage sequences (such as persistent high voltage) can be remembered by the LSTM network over a long period, identifying slowly developing faults such as battery aging.

[0090] In one embodiment of this application, the low-dimensional spatiotemporal features are input into a pre-trained graph neural network to generate graph features. In this embodiment, a graph convolutional neural network (GCN) is used to aggregate neighbor node information and analyze the connection relationships between nodes, such as whether a device frequently sends data to multiple nodes. If a computer control module frequently sends abnormal data to multiple nodes, the GCN can identify it as the central abnormal node and indicate a potential fault.

[0091] In one embodiment of this application, the temporal features and the graph features are fused to generate fused features. In this embodiment, the temporal features of LSTM and the graph features of GCN are directly concatenated, and then dimensionality is reduced through a fully connected layer and mapped to fault categories. The features extracted by LSTM and GCN are fused by concatenation or weighted averaging. For example, the temporal features generated by LSTM reflect when and what changes occur, while the spatial features generated by GCN reflect which features are related to which features; the two are merged into a comprehensive feature vector. The fused features contain both the temporal evolution pattern of the signal and the spatial distribution pattern of the nodes, forming a more comprehensive fault representation. Combining temporal changes and spatial distribution improves the fine-grained fault classification capability.

[0092] In one embodiment of this application, the fused features are mapped to a fault probability distribution through a fully connected layer. The fused features are then mapped to a fault type probability distribution through the fully connected layer. Based on learned patterns, such as specific temporal fluctuations corresponding to battery aging or specific spatial connectivity anomalies corresponding to short circuits, the model outputs the probability of various fault types. For example, if the fused features show that a node has abnormal temporal fluctuations and is associated with multiple nodes, the model may predict that the node has intermittent communication failures. The prediction results are presented in probabilistic form, intuitively reflecting the degree of matching between the current data and various fault types.

[0093] The above process, through the construction of an LSTM-GCN hybrid model, achieves spatiotemporal joint diagnosis of bus communication faults. Long Short-Term Memory (LSTM) networks are used to extract long-term dependencies of temporal features, graph neural networks are used to mine spatial correlation patterns between nodes, and a hybrid loss function is combined for optimization. The temporal and graph features are then fused and mapped to a fault probability distribution, ultimately outputting a high-precision fault classification result. This process combines temporal dynamics with spatial topology, making the fault probability distribution more closely reflect the actual operating rules of the system. It outperforms single-architecture models in terms of fine-grained classification, noise resistance, and real-time performance, improving the accuracy of fault prediction and making it suitable for intelligent diagnosis of complex bus systems.

[0094] S150, determine the reliability factor based on the fault probability distribution, and output the diagnostic result by combining the fault probability distribution and the reliability factor.

[0095] In one embodiment of this application, the reliability factor level is assessed based on the concentration and stability of the failure probability distribution. For example, a concentrated distribution of high-probability failures corresponds to a high reliability factor, while a dispersed distribution corresponds to a low reliability factor. Subsequently, the failure probability distribution and the reliability factor are correlated. In scenarios with a high reliability factor, the specific failure type and probability are directly output. When the reliability factor is low, it is supplemented with explanations that further verification or consideration of multiple factors is required. Finally, a comprehensive diagnostic result that balances probabilistic reliability and diagnostic accuracy is formed.

[0096] In one embodiment of this application, determining a reliability factor based on the failure probability distribution and outputting a diagnostic result by combining the failure probability distribution and the reliability factor includes:

[0097] Determine the probability mean and probability variance based on the fault probability distribution;

[0098] The reliability factor is determined based on the probability mean and probability variance.

[0099] Based on the fault probability distribution and the reliability factor, the diagnostic results are output.

[0100] In one embodiment of this application, based on the failure probability distribution, the probability mean and probability variance are determined by calculating the mean and variance, respectively. Then, a reliability factor is determined based on the probability mean and probability variance. for:

[0101]

[0102] in, Let represent the probability mean and probability variance of fault category c, where c and C represent the fault category identifier and total number, respectively. The confidence level of the model in the prediction results is assessed by using the mean and variance of the probability distribution. A concentrated probability distribution indicates a high reliability factor; a dispersed distribution indicates a low reliability factor. By measuring the degree of concentration of the probability distribution, misjudgments that might be caused by a single probability value are avoided.

[0103] In one embodiment of this application, a diagnostic result is output by combining the fault probability distribution and the reliability factor. Specifically, the category with the highest fault probability distribution is selected as the diagnostic conclusion, and the reliability of the result is evaluated in conjunction with the reliability factor. For example, if the reliability factor is less than a preset threshold, manual review or further inspection is triggered.

[0104] This embodiment generates diagnostic conclusions and reliability factors by inferring the uncertainty of the quantified model's predictions. This step, based on fine-grained fault classification, further addresses decision-making blind spots where the model may err but has a high probability of error, improving the reliability and interpretability of the diagnostic system. The introduction of reliability factors enables the system to distinguish between confirmed and suspected faults, providing crucial support for intelligent operation and maintenance of onboard buses.

[0105] The above process calculates the probability mean and variance based on the fault probability distribution, quantifies the reliability index of fault judgment, and generates a reliability factor in conjunction with the probability distribution. When finally outputting the diagnostic results, the reliability factor and the probability distribution work synergistically, providing both an assessment of the likelihood of a fault and a degree of confidence in the judgment, thus ensuring both accuracy and interpretability of the diagnostic results and enhancing decision support capabilities.

[0106] In this application's technical solution, the operating parameters of the embedded system are acquired, and the operating parameters are adaptively adjusted to generate adjustment data. A first feature extraction is performed on the adjustment data using a segmentation window length determined by hybrid time series, generating time-series segmented data. A second feature extraction is performed on the time-frequency domain and spatial domain respectively to obtain low-dimensional spatiotemporal features. These low-dimensional spatiotemporal features are input into a pre-trained hybrid network to generate a fault probability distribution. A reliability factor is determined based on the fault probability distribution, and a diagnostic result is output by combining the fault probability distribution and the reliability factor. This application's technical solution generates precise adjustment data by dynamically acquiring and adaptively adjusting the physical, data packet, and state parameters of the embedded system; extracts time-series features using a hybrid time-series segmentation window, and fuses multi-dimensional features from the time-frequency and spatial domains to form low-dimensional spatiotemporal features; generates a fault probability distribution through intelligent analysis using a hybrid network; and finally outputs a high-reliability diagnostic result by combining the reliability factor. This achieves intelligent processing from parameter monitoring to fault diagnosis, significantly improving system stability and diagnostic accuracy.

[0107] The following describes embodiments of the embedded system bus communication data monitoring and intelligent diagnostic platform of this application, which can be used to execute the embedded system bus communication data monitoring and intelligent diagnostic method in the above embodiments of this application. It is understood that the embedded system bus communication data monitoring and intelligent diagnostic platform can be a computer program (including program code) running on a computer device; for example, the embedded system bus communication data monitoring and intelligent diagnostic platform is an application software. The embedded system bus communication data monitoring and intelligent diagnostic platform can be used to execute the corresponding steps in the method provided in the embodiments of this application. For details not disclosed in the embodiments of the embedded system bus communication data monitoring and intelligent diagnostic platform of this application, please refer to the embodiments of the embedded system bus communication data monitoring and intelligent diagnostic method described above in this application.

[0108] Figure 3 A block diagram of an embedded system bus communication data monitoring and intelligent diagnostic platform according to an embodiment of this application is shown.

[0109] Reference Figure 3 As shown, an embedded system bus communication data monitoring and intelligent diagnostic platform according to an embodiment of this application includes:

[0110] The acquisition unit 310 is used to acquire the operating parameters of the embedded system, adaptively adjust the operating parameters, and generate adjustment data. The operating parameters include the physical parameters, data packet parameters, and status parameters of the embedded system.

[0111] The segmentation unit 320 is used to perform first feature extraction on the adjusted data by using a segmentation window length determined by the mixed time series, and generate time series segmentation data.

[0112] Feature unit 330 is used to extract second features from the time-frequency domain and the spatial domain respectively to obtain low-dimensional spatiotemporal features;

[0113] Probability unit 340 is used to input the low-dimensional spatiotemporal features into a pre-trained hybrid network to generate a fault probability distribution;

[0114] The diagnostic unit 350 is used to determine a reliability factor based on the fault probability distribution and output a diagnostic result by combining the fault probability distribution and the reliability factor.

[0115] In this application, based on the aforementioned scheme, the step of obtaining the operating parameters of the embedded system, adaptively adjusting the operating parameters, and generating adjustment data includes: obtaining historical parameters and real-time operating parameters of the embedded system; and adaptively adjusting the operating data based on the historical parameters and the operating parameters to generate adjustment data.

[0116] In this application, based on the aforementioned scheme, the step of adaptively adjusting the operating data based on the historical parameters and the operating parameters to generate adjusted data includes: determining the voltage value and current noise power of the current sampling period based on the operating parameters; determining the mean and standard deviation over a preset time period based on the historical parameters; and adaptively adjusting the operating data based on the voltage value of the current sampling period, the current noise power, and the mean and standard deviation over the preset time period to generate adjusted data.

[0117] In this application, based on the aforementioned scheme, the step of adaptively adjusting the operating data according to the voltage value of the current sampling period, the current noise power, and the mean and standard deviation of historical parameters within a preset time period to generate adjustment data includes: adaptively adjusting the operating data according to the voltage value of the current sampling period, the current noise power, and the mean and standard deviation of historical parameters within a preset time period to generate adjustment data. for:

[0118]

[0119] in, These represent the mean and standard deviation of the historical parameters within a preset time period, respectively. These represent the voltage value and the current noise power during the current sampling period, respectively. Indicates the reference noise power. This represents the noise sensitivity coefficient.

[0120] In this application, based on the aforementioned scheme, the step of extracting a first feature from the adjusted data and generating time-series segmented data by determining the segmentation window length through hybrid timing includes: determining the window length for segmenting the data by using the frame interval of the bus protocol and the response time in the adjusted data; and extracting a first feature from the adjusted data based on the window length to generate time-series segmented data.

[0121] In this application, based on the aforementioned scheme, the step of extracting second features from the time-frequency domain and spatial domain of the time-series segmented data to obtain low-dimensional spatiotemporal features includes: determining spatial factors based on preset scale parameters and translation parameters; extracting energy spectrum features in the time-frequency domain from the time-series segmented data based on the spatial factors; constructing topological relationships based on the layout information of bus nodes; generating spatial domain features based on the topological relationships; and fusing the energy spectrum features and the spatial domain features to obtain low-dimensional spatiotemporal features.

[0122] In this application, based on the aforementioned scheme, the step of inputting the low-dimensional spatiotemporal features into a pre-trained hybrid network to generate a fault probability distribution includes: inputting the low-dimensional spatiotemporal features into a pre-trained long short-term memory network to generate temporal features; inputting the low-dimensional spatiotemporal features into a pre-trained graph neural network to generate graph features; fusing the temporal features and the graph features to generate fused features; and mapping the fused features to the fault probability distribution through a fully connected layer.

[0123] In this application, based on the aforementioned scheme, the step of determining the reliability factor according to the fault probability distribution and outputting a diagnostic result by combining the fault probability distribution and the reliability factor includes: determining the probability mean and probability variance according to the fault probability distribution; determining the reliability factor according to the probability mean and probability variance; and outputting a diagnostic result by combining the fault probability distribution and the reliability factor.

[0124] In this application's technical solution, the operating parameters of the embedded system are acquired, and the operating parameters are adaptively adjusted to generate adjustment data. A first feature extraction is performed on the adjustment data using a segmentation window length determined by hybrid time series, generating time-series segmented data. A second feature extraction is performed on the time-frequency domain and spatial domain respectively to obtain low-dimensional spatiotemporal features. These low-dimensional spatiotemporal features are input into a pre-trained hybrid network to generate a fault probability distribution. A reliability factor is determined based on the fault probability distribution, and a diagnostic result is output by combining the fault probability distribution and the reliability factor. This application's technical solution generates precise adjustment data by dynamically acquiring and adaptively adjusting the physical, data packet, and state parameters of the embedded system; extracts time-series features using a hybrid time-series segmentation window, and fuses multi-dimensional features from the time-frequency and spatial domains to form low-dimensional spatiotemporal features; generates a fault probability distribution through intelligent analysis using a hybrid network; and finally outputs a high-reliability diagnostic result by combining the reliability factor. This achieves intelligent processing from parameter monitoring to fault diagnosis, significantly improving system stability and diagnostic accuracy.

[0125] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.

[0126] It should be noted that the computer system of the electronic device in this embodiment is only an example and should not impose any limitations on the function and scope of use of the embodiments of this application.

[0127] In this embodiment, the computer system includes a central processing unit 401, which can perform various appropriate actions and processes based on a program stored in the read-only memory 402 or a program loaded from the storage section 408 into the random access memory 403, such as executing the embedded system bus communication data monitoring and intelligent diagnostic method described in the above embodiment. The random access memory 403 also stores various programs and data required for system operation. The central processing unit 401, the read-only memory 402, and the random access memory 403 are interconnected via a bus 404. An input / output interface 405 is also connected to the bus 404.

[0128] The following components are connected to the input / output interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the input / output interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 410 as needed so that computer programs read from it can be installed into the storage section 408 as needed.

[0129] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit 401, it performs various functions defined in the system of this application.

[0130] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0131] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0132] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0133] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the various alternative implementations described above.

[0134] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to implement the embedded system bus communication data monitoring and intelligent diagnosis method described in the above embodiments.

[0135] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0136] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this application.

[0137] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0138] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for monitoring and intelligently diagnosing bus communication data in an embedded system, characterized in that, include: The operating parameters of the embedded system are obtained, and the operating parameters are adaptively adjusted to generate adjustment data. The operating parameters include the physical parameters, data packet parameters, and status parameters of the embedded system. By using the segmentation window length determined by the mixed temporal sequence, the first feature is extracted from the adjusted data to generate temporal segmentation data; The second feature is extracted from the time-frequency domain and the spatial domain respectively to obtain low-dimensional spatiotemporal features; The low-dimensional spatiotemporal features are input into a pre-trained hybrid network to generate a fault probability distribution; Based on the failure probability distribution, a reliability factor is determined, and a diagnostic result is output by combining the failure probability distribution and the reliability factor. Specifically, by using the segmentation window length determined through hybrid temporal sequencing, the adjusted data undergoes first feature extraction to generate temporal segmentation data, including: The window length for segmenting the data is determined by the frame interval of the bus protocol and the response time in the adjusted data. Based on the window length, the adjusted data is subjected to first feature extraction to generate time-series segmentation data; The process of inputting the low-dimensional spatiotemporal features into a pre-trained hybrid network to generate a fault probability distribution includes: The low-dimensional spatiotemporal features are input into a pre-trained long short-term memory network to generate temporal features; The low-dimensional spatiotemporal features are input into a pre-trained graph neural network to generate graph features; The temporal features and the graph features are fused to generate fused features; The fused features are mapped to a fault probability distribution through a fully connected layer.

2. The embedded system bus communication data monitoring and intelligent diagnosis method according to claim 1, characterized in that, Obtain the operating parameters of the embedded system, adaptively adjust the operating parameters, and generate adjustment data, including: Obtain historical and real-time operating parameters of the embedded system; Based on the historical parameters and the operating parameters, the operating parameters are adaptively adjusted to generate adjustment data.

3. The embedded system bus communication data monitoring and intelligent diagnosis method according to claim 2, characterized in that, Based on the historical parameters and the operating parameters, the operating parameters are adaptively adjusted to generate adjustment data, including: The voltage value and current noise power for the current sampling period are determined based on the operating parameters. The mean and standard deviation within a preset time period are determined based on historical parameters; Based on the voltage value of the current sampling period, the current noise power, and the mean and standard deviation within a preset time period, the operating parameters are adaptively adjusted to generate adjustment data.

4. The embedded system bus communication data monitoring and intelligent diagnosis method according to claim 1, characterized in that, The second feature is extracted from the time-series segmented data in both the time-frequency domain and the spatial domain to obtain low-dimensional spatiotemporal features, including: Based on preset scale and translation parameters, spatial factors are determined, and time-frequency domain energy spectrum features are extracted from the time-series segmented data based on the spatial factors. Based on the deployment information of bus nodes, a topology relationship is constructed, and spatial domain features are generated based on the topology relationship; By fusing the energy spectrum features and the spatial domain features, low-dimensional spatiotemporal features are obtained.

5. The embedded system bus communication data monitoring and intelligent diagnosis method according to claim 1, characterized in that, Based on the failure probability distribution, a reliability factor is determined, and a diagnostic result is output by combining the failure probability distribution and the reliability factor, including: Determine the probability mean and probability variance based on the fault probability distribution; The reliability factor is determined based on the probability mean and probability variance. Based on the fault probability distribution and the reliability factor, the diagnostic results are output.

6. An embedded system bus communication data monitoring and intelligent diagnostic platform, characterized in that, include: An acquisition unit is used to acquire the operating parameters of the embedded system, adaptively adjust the operating parameters, and generate adjustment data. The operating parameters include the physical parameters, data packet parameters, and status parameters of the embedded system. A segmentation unit is used to extract the first feature from the adjusted data by using a segmentation window length determined by the mixed temporal sequence, thereby generating temporal segmentation data; The feature unit is used to extract second features from the time-frequency domain and the spatial domain respectively to obtain low-dimensional spatiotemporal features; A probability unit is used to input the low-dimensional spatiotemporal features into a pre-trained hybrid network to generate a fault probability distribution. A diagnostic unit is used to determine a reliability factor based on the fault probability distribution and output a diagnostic result by combining the fault probability distribution and the reliability factor. Specifically, by using the segmentation window length determined through hybrid temporal sequencing, the adjusted data undergoes first feature extraction to generate temporal segmentation data, including: The window length for segmenting the data is determined by the frame interval of the bus protocol and the response time in the adjusted data. Based on the window length, the adjusted data is subjected to first feature extraction to generate time-series segmentation data; The process of inputting the low-dimensional spatiotemporal features into a pre-trained hybrid network to generate a fault probability distribution includes: The low-dimensional spatiotemporal features are input into a pre-trained long short-term memory network to generate temporal features; The low-dimensional spatiotemporal features are input into a pre-trained graph neural network to generate graph features; The temporal features and the graph features are fused to generate fused features; The fused features are mapped to a fault probability distribution through a fully connected layer.

7. The embedded system bus communication data monitoring and intelligent diagnostic platform according to claim 6, characterized in that, Obtain the operating parameters of the embedded system, adaptively adjust the operating parameters, and generate adjustment data, including: Obtain historical and real-time operating parameters of the embedded system; Based on the historical parameters and the operating parameters, the operating parameters are adaptively adjusted to generate adjustment data.

Images