A method of monitoring a game device for failure

Abstract

This invention relates to the field of intelligent device monitoring and fault detection, and particularly to a method for monitoring faults in gaming devices. The method includes: collecting operational data from the gaming device to construct a multi-dimensional time-series dataset; performing recursive cross-correlation analysis on the multi-dimensional time-series dataset, using multi-layered nested time window analysis to generate a nested time window matrix; introducing an interaction matrix, and processing the nested time window matrix through an adaptive interdependent network model to generate a comprehensive node state vector; calculating a predicted fault score based on the comprehensive node state vector; and monitoring gaming device faults based on the predicted fault score. This method solves the problems of easily generating false alarms or missed alarms when processing complex gaming device operational data, the inability to dynamically adapt to changes in the gaming device's state, the failure to fully utilize the potential correlations between various dimensions to improve fault detection accuracy, and the inability to adjust model parameters in a timely manner to adapt to newly emerging data characteristics.

Description

Technical Field

[0001] This invention relates to the field of intelligent device monitoring and fault detection, and in particular to a method for monitoring faults in gaming devices. Background Technology

[0002] In the development of modern gaming devices, with the improvement of hardware performance and the increasing complexity of software functions, the structure and operating environment of gaming devices have become increasingly complex. This complexity is not only reflected in the collaborative work of multiple components and modules within the gaming device, but also involves the interaction between the gaming device and external environments such as networks and servers. To provide a more immersive and smooth gaming experience, the utilization rate of hardware configurations such as CPU, GPU, and memory in gaming devices is constantly increasing, and the software also needs to run a large number of processes and real-time computing tasks. At the same time, the dependence of gaming devices on network connectivity is also gradually increasing; for example, online gaming, cloud gaming, and real-time updates all require gaming devices to have high-quality network communication capabilities.

[0003] With these increasing demands, gaming devices face greater challenges in stability and reliability during actual operation. High hardware loads can lead to overheating and resource contention; complex task management and multi-tasking in software can easily cause memory leaks and process anomalies; and network instability can result in increased latency and data loss, impacting user experience. Therefore, real-time monitoring of gaming device operation and timely identification and handling of potential faults are crucial to ensuring the normal operation of gaming devices.

[0004] However, existing game device fault monitoring methods have the following technical problems: they cannot fully capture potential problems in the operation of game devices; they are prone to false alarms or missed alarms when processing complex game device operation data; they cannot dynamically adapt to changes in the state of game devices, resulting in a significant reduction in the timeliness and accuracy of fault warnings; they fail to fully utilize the potential correlations between various dimensions to improve the accuracy of fault detection; and when dealing with dynamically changing game device operating environments, they lack adaptive capabilities and usually cannot adjust model parameters in a timely manner to adapt to newly emerging data characteristics. This makes them perform poorly in dealing with complex and dynamic game device fault scenarios and cannot meet the high precision and real-time requirements of practical applications. Summary of the Invention

[0005] This invention provides a method for monitoring faults in gaming devices to address the following issues: the inability to comprehensively capture potential problems during the operation of gaming devices; the tendency to generate false alarms or missed alarms when processing complex gaming device operation data; the inability to dynamically adapt to changes in the state of gaming devices, resulting in a significant reduction in the timeliness and accuracy of fault warnings; the failure to fully utilize the potential correlations between various dimensions to improve the accuracy of fault detection; and the lack of adaptive capabilities when dealing with dynamically changing gaming device operating environments, typically failing to adjust model parameters in a timely manner to adapt to newly emerging data characteristics. These shortcomings make the method ineffective in dealing with complex and dynamic gaming device fault scenarios, failing to meet the high precision and real-time requirements of practical applications.

[0006] The present invention provides a method for monitoring faults in gaming devices, specifically comprising the following technical solutions:

[0007] A method for monitoring malfunctions in gaming devices includes the following steps:

[0008] S1. Collect different operating data of gaming devices to construct a multidimensional time series dataset; perform cross-correlation recursive analysis between dimensions of the multidimensional time series dataset, and generate a nested time window matrix using multi-level nested time window analysis;

[0009] S2. Introduce an interaction matrix and process the nested time window matrix through a multi-layer adaptive interdependent network model to generate a comprehensive node state vector; calculate the predicted fault score based on the comprehensive node state vector; and monitor game device faults based on the predicted fault score.

[0010] Preferably, S1 specifically includes:

[0011] In the process of performing cross-correlation recursive analysis on multidimensional time series datasets, a cross-correlation function is defined to identify dimensional pairs with strong coupling relationships. Through multi-level recursive processing, the strong coupling relationships between each dimension are refined to obtain recursive results at different levels.

[0012] Preferably, S2 specifically includes:

[0013] The multi-layer adaptive interdependent network model introduces an adaptive interdependency mechanism into a multi-layer network structure. By constructing a multi-layer adaptive interdependent network structure and processing the nested time window matrix, the node state vector of each layer is obtained. The specific formula is as follows:

[0014]

[0015] Among them, h (l) (t) represents the node state vector of the l-th layer network at time t; h (l)(t-1) represents the node state vector of the l-th layer network at time t-1; This represents the self-updating weight matrix of the l-th layer; W represents the input weight matrix of the l-th layer; (l) (t) is the nested time window matrix of the l-th layer; This represents the sum of interdependent effects from layers other than layer l; s represents the index of a different network layer, used to traverse all other layers different from layer l; γ ls h represents the mutual dependency weights from layer s to layer l; (s) (t) represents the node state vector of the s-th layer network at time t; Ω ls This represents the interaction matrix from layer s to layer l; L is the number of layers in the multilayer adaptive interdependent network.

[0016] A comprehensive node state vector is generated by weighted summation of the node state vectors at each layer.

[0017] Preferably, S2 specifically includes:

[0018] The synthesized node state vector is converted into a scalar fault score through linear transformation.

[0019] Preferably, S2 specifically includes:

[0020] Design a loss function for a multi-layer adaptive interdependent network model, focusing on the interaction matrix Ω. ls Update.

[0021] Preferably, S2 specifically includes:

[0022] By introducing mean squared error, robustness term, and sparsity regularization term, and by weighting and summing these terms, the loss function of the multilayer adaptive interdependent network model is obtained.

[0023] Preferably, S2 specifically includes:

[0024] Based on the interaction matrix, an adaptive adjustment mechanism is introduced to enable the interaction matrix to dynamically adapt to changes in the state of nodes at each layer; the interaction matrix is ​​adjusted according to the loss function of the multi-layer adaptive interdependent network model through iterative optimization.

[0025] Preferably, S2 specifically includes:

[0026] After obtaining the predicted fault score, a threshold is set, and the predicted fault score is compared with the threshold to determine whether the current gaming device is in a faulty state.

[0027] The beneficial effects of the technical solution of the present invention are:

[0028] 1. This invention deploys multiple sensors to collect hardware signals, software signals, and network signals of gaming devices in real time, constructing a multi-dimensional time-series dataset. Through complex cross-correlation recursive analysis between dimensions and multi-level recursive processing, it can identify the potential correlations between multi-dimensional data, thereby achieving early warning and accurate detection of gaming device malfunctions.

[0029] 2. The introduction of the multi-layer adaptive interdependent network model can dynamically adjust the interdependence between layers according to the changes in the state of the game device. By adaptively adjusting the interaction matrix, the multi-layer adaptive interdependent network model can gradually strengthen or weaken the interaction between different layers, thereby better adapting to the complex operating environment of the game device and improving the accuracy of fault detection.

[0030] 3. The loss function of the multilayer adaptive interdependent network model of this invention takes into account the accuracy of fault detection, the robustness of prediction results, and the sparsity of the interaction matrix. Through mean square error, robustness term and sparsity regularization term, the accuracy of fault detection is improved while the false alarms and false negatives are effectively reduced, ensuring the reliability of fault early warning. Attached Figure Description

[0031] Figure 1 This is a flowchart of a game device fault monitoring method according to the present invention. Detailed Implementation

[0032] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0034] The following description, in conjunction with the accompanying drawings, details a specific scheme for a game device fault monitoring method provided by the present invention.

[0035] See attached document Figure 1 The diagram illustrates a flowchart of a game device fault monitoring method according to an embodiment of the present invention, which includes the following steps:

[0036] S1. Collect different operating data of gaming devices to construct a multidimensional time series dataset; perform cross-correlation recursive analysis between dimensions of the multidimensional time series dataset, and generate a nested time window matrix using multi-level nested time window analysis;

[0037] Multiple sensors are deployed to collect different operational data from the gaming device, including hardware signals, software signals, and network signals. Hardware signals primarily collect key hardware parameters such as CPU temperature, GPU utilization, and memory status. This hardware signal data is acquired in real-time by the sensors, forming a basic dataset of the gaming device's hardware health status. This dataset is used to monitor the basic operational status of the gaming device and ensure the early detection of hardware failures. Software signal collection focuses on operating system processes, log information, and memory utilization. Log analysis and process monitoring tools are used to acquire data streams in real-time, identifying potential software failures such as memory leaks and abnormal process termination, providing a basis for subsequent software-level fault detection. Network signal collection mainly collects data such as network latency, packet loss rate, and bandwidth usage. This network signal is used to monitor the network communication quality of the gaming device, ensuring stable communication with external servers or other devices and preventing failures caused by network problems.

[0038] The collected operational data from different gaming devices form a multidimensional time-series dataset: X(t) = {x1(t), x2(t), ..., x...} m Let X(t)} represent a multidimensional time series dataset at time t, where xt represents the dataset at time t. i (t) represents the data value of the i-th dimension of a multidimensional time series dataset at time t, where m is the total number of dimensions. A multidimensional time series dataset not only contains the current state of each dimension but also allows for the capture of interrelationships between dimensions through subsequent processing.

[0039] To reveal the complex relationships between data points in a multidimensional time-series dataset, a cross-correlation recursive analysis is performed on the multidimensional time-series dataset. This cross-correlation recursive analysis aims to identify potential relationships within the multidimensional time-series data, and the formula is as follows:

[0040]

[0041] Where, ρ i j(τ) is the cross-correlation function of the i-th and j-th dimensions of the multidimensional time series dataset at times t and t+τ; x j (t+τ) represents the data value of the i-th dimension of the multidimensional time-series dataset at time t+τ, where τ is the time delay; μ i and μ j σi and σj are the means of the i-th and j-th dimensions of the multidimensional time series dataset, respectively; i and σj Let ρi and ρi be the standard deviations of the i-th and j-th dimensions of the multidimensional time-series dataset, respectively. The cross-correlation between different dimensions of the multidimensional time-series dataset is calculated to analyze ρi. j (τ) changes with time delay τ. Find the peak of the cross-correlation function. The position of the peak reflects the moment of maximum correlation between the i-th dimension and the i-th dimension, indicating that the two dimensions have a strong coupling relationship under the time delay τ. Thus, we can identify the dimension pairs with a strong coupling relationship on the time axis.

[0042] After identifying strong coupling relationships, multi-level recursive processing is used to further refine the strong coupling relationships between each dimension, making the detection results more accurate. The multi-level recursive formula is as follows:

[0043]

[0044] Among them, R (k) (t) is the result of the k-th recursion; f is a non-linear activation function used to enhance the non-linear relationships that occur during the recursion process; It is the cross-correlation function in the (k-1)th level of recursion. Each recursive level in the multi-level recursive processing depends on the output of the previous recursive level, forming a recursive chain, which ensures a progressively in-depth analysis of the relationships between data in various dimensions.

[0045] After the cross-correlation recursive analysis is completed, multi-level nested time window analysis is introduced to further capture data patterns within the time range. The specific time range value is determined based on the time window offset setting and the time point, typically set according to the application scenario and the characteristics of the gaming device, depending on the possible cycle of faults in the specific application scenario. The formula for the nested time window matrix is:

[0046]

[0047] Among them, w l，n The offset of the l-th nested time window determines the position of the nth time point in the nested time window matrix relative to the current time t; W (l) (t) is the nested time window matrix of the l-th level, representing the recursive result matrix at time t, after time offset processing; R (k) (tw l，n ) is the result of the k-th recursion after the time window offset is w. l，n The output after time offset processing represents the time tw l，n The nested time window matrix reveals dynamic patterns at different time scales by capturing recursive results at multiple time points.

[0048] S2. Introduce an interaction matrix and process the nested time window matrix through a multi-layer adaptive interdependent network model to generate a comprehensive node state vector; calculate the predicted fault score based on the comprehensive node state vector; and monitor game device faults based on the predicted fault score.

[0049] The data in the nested time window matrix is ​​not isolated but processed through a multi-layer adaptive interdependent network model. This model introduces an adaptive interdependency mechanism into the multi-layer network structure, ensuring that the output of each layer depends not only on its input but also on the states of other layers, thus capturing more complex temporal and dimensional relationships. Specifically, when processing the nested time window matrix, the multi-layer adaptive interdependent network model constructs a multi-layer adaptive interdependent network structure. The node state of each layer consists of two parts: its own network node state evolution and the influence of interdependencies from other layers. The node state vector h of each layer l... (l) The update formula for (t) is:

[0050]

[0051] Among them, h (l) (t) represents the node state vector of the l-th layer network at time t; h (l) (t-1) represents the node state vector of the l-th layer network at time t-1; This represents the self-updating weight matrix of the l-th layer; This represents the input weight matrix of the l-th layer; This represents the sum of interdependent effects from other layers; s represents the index of a different network layer, used to traverse all other layers different from layer l; γ ls h represents the mutual dependency weights from layer s to layer l; (s) (t) represents the node state vector of the s-th layer network at time t; Ω ls Let L represent the interaction matrix from layer s to layer l; L is the number of layers in the multilayer adaptive interdependent network.

[0052] By combining the outputs of each layer, a comprehensive node state vector H(t) is generated by weighted summing of the node state vectors of all layers:

[0053]

[0054] Where H(t) is a comprehensive node state vector, representing the comprehensive node state evaluation of the multilayer adaptive interdependent network model at time t; α lThese are the weighting coefficients for each layer of the network, which are optimized during the training of the multi-layer adaptive interdependent network model to ensure that the comprehensive node state vector can fully reflect the information contribution of each layer of the network.

[0055] The output layer generates a predicted fault score based on the comprehensive node state vector H(t). The synthesized node state vector is converted into a scalar fault score through a linear transformation. The specific formula is as follows:

[0056]

[0057] Where, ω o This is the weight matrix of the output layer, responsible for mapping the high-dimensional synthesized node state vector H(t) to a scalar output; b o It is a bias term used to adjust the baseline level of the scalar output; It is an activation function, and you can choose a linear or non-linear function depending on the specific application, such as the sigmoid function.

[0058] In a multi-layer adaptive interdependent network model, in order to adjust the interaction matrix Ω ls A loss function needs to be designed to guide the optimization process. The design of the loss function is directly related to the optimization objective and performance of the multilayer adaptive interdependent network model. The design of the loss function is based on the following key factors: the accuracy of fault detection, the robustness of prediction results, and the sparsity of the interaction matrix.

[0059] Specifically, the main part of the loss function comes from the accuracy evaluation of fault detection by the multilayer adaptive interdependent network model. Assuming the actual fault state at time t is y(t), the predicted fault score of the multilayer adaptive interdependent network model is... Mean squared error is used to measure the difference between the predicted fault score and the actual fault state:

[0060]

[0061] Among them, L mse is the mean squared error; T is the total length of the time series; the mean squared error loss function described above ensures that the multilayer adaptive interdependent network model can predict the fault state as accurately as possible.

[0062] To avoid drastic fluctuations in the fault scores predicted by the multi-layer adaptive interdependent network model over time, a robustness term L is introduced into the loss function. smooth This is used to penalize excessive variations in fault scores predicted by a multilayer adaptive interdependent network model between adjacent time points. The specific expression is:

[0063]

[0064] The introduced robustness term helps the multilayer adaptive interdependent network model output a smoother predicted fault score, avoiding overfitting to short-term noisy data.

[0065] To make the interaction matrix Ω ls To make it more concise and interpretable, a sparsity regularization term L is introduced into the loss function. sparse The sparse regularization term L sparse Based on the interaction matrix Ω ls L1 norm:

[0066]

[0067] By introducing sparsity constraints, the loss function encourages most elements in the interaction matrix to approach zero, thus preserving only the most significant inter-layer dependencies. This helps improve the generalization ability of multilayer adaptive interdependent network models and simplifies the structure of the interaction matrix, making the interaction matrix and fault prediction scores easier to interpret.

[0068] The final comprehensive loss function is a weighted sum of the mean squared error, the robustness term, and the sparsity regularization term:

[0069]

[0070] in, β1 is the loss function of the multilayer adaptive interdependent network model; β2 and β3 are the weight coefficients of the mean squared error, robustness term and sparsity regularization term, respectively, which can be adjusted experimentally to optimize the performance of the multilayer adaptive interdependent network model.

[0071] During the optimization process, gradient descent or its variants, such as the Adam optimizer, are used to minimize the loss function and adaptively adjust the interaction matrix Ω. ls .

[0072] In a multi-layer adaptive interdependent network model, the interaction matrix Ω ls The interaction matrix Ω determines the interdependencies between different layers. ls The generation and adaptive adjustment process is as follows:

[0073] Construct an initial interaction matrix for each pair of layers l and s (where s ≠ l). The initial construction is based on the correlation calculation of node states between two layers. The correlation coefficient matrix of node states in the s-th layer and the l-th layer at time point t is calculated, and its elements are:

[0074]

[0075] in, This represents the r-th node in the l-th layer and the r-th node in the s-th layer. The correlation coefficient matrix between nodes; This represents the node state vector of the r-th node in the l-th layer at time t; Represents the s-th layer The node state vectors of nodes at time t; μl and μs are the mean values ​​of the node state vectors of the l-th and s-th layers, respectively; T is the total length of the time series. The correlation coefficient matrix describes the linear dependence of node states between the two layers.

[0076] The correlation coefficient matrix is ​​normalized to obtain the initial interaction matrix. The initial interaction matrix is ​​only a basic estimate. In order to enable the interaction matrix to dynamically adapt to the changes in the state of nodes in each layer, an adaptive adjustment mechanism is introduced. The interaction matrix is ​​adjusted according to the loss function during the training process of the multi-layer adaptive interdependent network model through iterative optimization.

[0077] Specifically, in each iteration, the loss function of the multi-layer adaptive interdependent network model is used... For the interaction matrix Ω ls To prevent excessive fluctuations or local optima in the interaction matrix during training of the multilayer adaptive interdependent network model, a stability constraint is introduced. A regularization term is added to the update process of the interaction matrix to balance the adjustment intensity. The specific formula is as follows:

[0078]

[0079] in, and These are the interaction matrices before and after the update, respectively; η is the learning rate. It is the gradient of the loss function with respect to the interaction matrix; is the regularization coefficient. The update rule enables the interaction matrix to gradually adapt to the complex dependencies between layers during the training of a multi-layer adaptive interdependent network model, progressively strengthening or weakening the interactions between different layers. After multiple rounds of iteration and adaptive adjustment, the final interaction matrix not only reflects the static correlations between layers but also captures the complex dependencies between the states of nodes in each layer during dynamic training.

[0080] After obtaining the predicted fault score Next, the predicted fault score is compared with a pre-set threshold based on expert experience to determine whether the gaming device is currently in a faulty state. Once a faulty state is determined, an early warning mechanism is triggered, maintenance personnel are notified, and a detailed fault diagnosis procedure is initiated.

[0081] In summary, a method for monitoring game device malfunctions has been developed.

[0082] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0084] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for monitoring faults in gaming devices, characterized in that, Includes the following steps: S1. Collect different operating data from gaming devices and construct a multi-dimensional time-series dataset; A recursive cross-correlation analysis between dimensions is performed on a multidimensional time-series dataset, and a nested time window matrix is ​​generated using multi-level nested time window analysis. ， in, For the first Time window offset for nested layers; It is the first The nested time window matrix of the layers represents the time... The recursive result matrix has been processed with time offset; It is the first The time window offset of the recursive result is The output after time offset processing represents the time... The recursive result on; S2. An interaction matrix is ​​introduced, and the nested time window matrix is ​​processed through a multi-layer adaptive interdependent network model. The multi-layer adaptive interdependent network model introduces an adaptive interdependence mechanism into the multi-layer network structure. By constructing a multi-layer adaptive interdependent network structure and processing the nested time window matrix, the node state vector of each layer of the network is obtained. By weighted summing of the node state vectors of each layer, a comprehensive node state vector is generated. Based on the comprehensive node state vector, a predicted fault score is calculated. Based on the predicted fault score, the game device fault is monitored.

2. The game device fault monitoring method according to claim 1, characterized in that, S1 specifically includes: In the process of performing cross-correlation recursive analysis on multidimensional time series datasets, a cross-correlation function is defined to identify dimensional pairs with strong coupling relationships. Through multi-level recursive processing, the strong coupling relationships between each dimension are refined to obtain recursive results at different levels.

3. The game device fault monitoring method according to claim 1, characterized in that, S2 specifically includes: The specific formula for the node state vector of each layer of the network is as follows: ， in, Indicates in Time of the first The node state vector of a layered network; Indicates in Time of the first The node state vector of a layered network; Indicates the first The self-updating weight matrix of the layer; Indicates the first The input weight matrix of the layer; It is the first Nested time window matrix of layers; Indicates from except the first The comprehensive impact of interdependencies between layers outside the main layer; Represents different network layer indices, used to traverse all layers related to the first layer. Other layers that are different from the previous ones; Indicates from the first layer to the first Layer interdependency weights; Indicates in Time of the first The node state vector of a layered network; Indicates from the first layer to the first The interaction matrix of the layers; It refers to the number of layers in a multi-layer adaptive interdependent network.

4. The game device fault monitoring method according to claim 3, characterized in that, S2 specifically includes: The synthesized node state vector is converted into a scalar fault score through linear transformation.

5. The game device fault monitoring method according to claim 1, characterized in that, S2 specifically includes: Design a loss function for a multi-layer adaptive interdependent network model, focusing on the interaction matrix. Update.

6. The game device fault monitoring method according to claim 5, characterized in that, S2 specifically includes: By introducing mean squared error, robustness term, and sparsity regularization term, and by weighting and summing these terms, the loss function of the multilayer adaptive interdependent network model is obtained.

7. The game device fault monitoring method according to claim 6, characterized in that, S2 specifically includes: Based on the interaction matrix, an adaptive adjustment mechanism is introduced to enable the interaction matrix to dynamically adapt to changes in the state of nodes at each layer; the interaction matrix is ​​adjusted according to the loss function of the multi-layer adaptive interdependent network model through iterative optimization.

8. The game device fault monitoring method according to claim 1, characterized in that, S2 specifically includes: After obtaining the predicted fault score, a threshold is set, and the predicted fault score is compared with the threshold to determine whether the current gaming device is in a faulty state.

Images