Method and system for efficiently acquiring data of HPLC (High Performance Liquid Chromatography) dual-mode communication module based on edge calculation

Abstract

The invention provides an efficient data acquisition method and system for an HPLC (High Performance Liquid Chromatography) dual-mode communication module based on edge calculation, and relates to the technical field of HPLC dual-mode communication data acquisition. The method comprises the steps of collecting multi-dimensional resource data of an HPLC dual-mode communication module in real time, performing feature extraction by adopting a local linear embedding algorithm, constructing a resource state evaluation model based on an extreme learning machine, evaluating a resource load state of the communication module in real time, and establishing a multi-target optimization model. Solving by adopting a non-dominated sorting genetic algorithm III to obtain an acquisition strategy solution set, selecting an optimal acquisition strategy, converting the optimal acquisition strategy into a control instruction, and adjusting data acquisition parameters in combination with predictive control and a feedback correction mechanism. According to the method, the problems of resource waste, single index and the like of an existing scheme are effectively solved by performing adaptive denoising, feature extraction and resource state accurate evaluation on the multi-dimensional data and combining multi-objective optimization of data space-time association mining and acquisition efficiency, energy consumption and reliability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of HPLC dual-mode communication data acquisition, and in particular to an HPLC dual-mode communication module data efficient acquisition method and system based on edge computing. BACKGROUND

[0002] Under the background that HPLC (High-speed Power Line Communication) dual-mode communication technology is widely applied in intelligent power grids, industrial Internet of Things, new-generation mobile communication core networks and access network construction, networking and other mobile telecommunication services, the efficiency, stability and low energy consumption of data acquisition are increasingly highlighted. The current data acquisition scheme of HPLC dual-mode communication module generally has the problem of insufficient multi-dimensional resource coordination: on the one hand, the traditional acquisition method mostly adopts a fixed parameter strategy, without considering the dynamic changes of real-time computing power, memory and communication link state of the module, for example, when the CPU load rate suddenly rises or the link bandwidth fluctuates, the fixed acquisition frequency is easy to cause data loss or resource waste; on the other hand, the existing scheme lacks comprehensive optimization of data acquisition efficiency, communication energy consumption and transmission reliability, and often single-mindedly pursues a certain index, such as blindly increasing the acquisition frequency to improve the acquisition efficiency, ignoring the risk of increasing energy consumption and link congestion.

[0003] At the same time, although the application of edge computing technology in distributed data processing has gradually become popular, the existing HPLC acquisition system fails to fully utilize the distributed processing capability of edge nodes, and the data preprocessing and strategy adjustment mostly rely on the cloud, resulting in high response delay and difficulty in meeting real-time requirements. In addition, the traditional data denoising and feature extraction method has poor adaptability to multi-dimensional resource data, and is easily affected by environmental factors such as power line noise and wireless interference, causing feature distortion, and thus leading to inaccurate resource state evaluation and lack of reliable basis for acquisition strategy adjustment. These problems together result in low data acquisition efficiency, high energy consumption and insufficient reliability of the existing HPLC dual-mode communication module, which is difficult to adapt to complex and changeable communication scenarios, and restricts the further promotion of HPLC technology in high-precision and high-real-time application scenarios. SUMMARY

[0004] The present application provides an HPLC dual-mode communication module data efficient acquisition method and system based on edge computing to solve the defects in the prior art.

[0005] On the one hand, the present application provides an HPLC dual-mode communication module data efficient acquisition method based on edge computing, comprising: real-time acquisition of multi-dimensional resource data of the HPLC dual-mode communication module, the multi-dimensional resource data including computing power data, memory resource data and communication link data.

[0006] Adaptive median filtering algorithm is used for denoising multi-dimensional resource data, and local linear embedding algorithm is used for feature extraction to obtain computing power features, memory features and link features.

[0007] A resource state evaluation model based on extreme learning machine is constructed, the resource load state of the communication module is evaluated in real time according to the computing power features, the memory features and the link features, and the data collection strategy is adjusted according to the preset threshold rule.

[0008] A federal filtering algorithm is used to fuse multi-dimensional resource data, and a Gated Recurrent Unit network is used to construct a spatio-temporal correlation model of multi-dimensional resource data.

[0009] A multi-objective optimization model of comprehensive data collection efficiency, communication energy consumption and data transmission reliability is established according to the spatio-temporal correlation model, a non-dominated sorting genetic algorithm III is used to solve the multi-objective optimization model, a collection strategy solution set is obtained, and the optimal collection strategy of the current communication scene is selected by an analytic hierarchy process.

[0010] The distributed processing capacity of the edge computing node is utilized, the optimal collection strategy is converted into a control instruction through an HPLC dual-mode communication protocol, and the data collection parameters are adjusted in combination with a predictive control and a feedback correction mechanism.

[0011] According to the HPLC dual-mode communication module data efficient collection method based on edge computing provided by the application, the computing power data includes CPU load rate, GPU computing capacity value and computing power utilization rate.

[0012] According to the HPLC dual-mode communication module data efficient collection method based on edge computing provided by the application, the process of feature extraction by using the local linear embedding algorithm includes: A multi-dimensional resource data matrix is constructed, for each feature in the multi-dimensional resource data matrix, a local neighborhood similarity matrix is calculated, and a standardized data set and a local covariance matrix are obtained.

[0013] The local covariance matrix is singular value decomposed, the feature vectors corresponding to the first m largest singular values are selected as embedding features according to the cumulative contribution rate of the singular values, and the computing power features, the memory features and the link features are obtained.

[0014] According to the HPLC dual-mode communication module data efficient collection method based on edge computing provided by the application, the process of constructing a resource state evaluation model based on an extreme learning machine includes: Historical running multi-dimensional resource data samples of the HPLC dual-mode communication module are collected, and the resource load state labels corresponding to each data sample are labeled.

[0015] The radial basis kernel function is defined as an activation function of the extreme learning machine, and an optimization objective function of the extreme learning machine is constructed.

[0016] The output weight vector is solved by Moore-Penrose generalized inverse to obtain the resource state evaluation model.

[0017] According to the HPLC dual-mode communication module data efficient acquisition method based on edge computing provided by the application, the process of constructing the space-time correlation relationship model of multi-dimensional resource data by using the gated recurrent unit network comprises: The multi-dimensional resource data fused through the federated filtering is sorted according to time sequence to obtain an input sequence.

[0018] The structure of the gated recurrent unit network is set, including an input layer, a hidden layer and an output layer. The input layer is used for receiving the input sequence. The hidden layer comprises a plurality of gated recurrent units, and the space-time dependence in the data is captured by controlling the flow of information through the update gate and the reset gate. The output layer adopts a full connection layer, and outputs a hidden state vector of the data at each time.

[0019] The gated recurrent unit network is trained to analyze the correlation strength of the multi-dimensional resource data in the time dimension and the device node dimension.

[0020] The hidden state vector is taken as node features, and the space-time correlation coefficient between nodes is taken as edge weights to construct the space-time correlation relationship model of the multi-dimensional resource data.

[0021] According to the HPLC dual-mode communication module data efficient acquisition method based on edge computing provided by the application, the process of establishing a multi-objective optimization model of comprehensive data acquisition efficiency, communication energy consumption and data transmission reliability according to the space-time correlation relationship model comprises: The data acquisition efficiency is defined as the ratio of the amount of data effectively acquired in unit time to the theoretical maximum acquisition amount.

[0022] The communication energy consumption is defined as the total power consumption of the HPLC dual-mode communication module, and is calculated according to the transmission power, the acquisition period and the communication duration.

[0023] The data transmission reliability model is established according to the data packet loss rate, the transmission delay and the data integrity check result.

[0024] The comprehensive objective function is established as the multi-objective optimization model according to the data acquisition efficiency, the communication energy consumption and the data transmission reliability.

[0025] According to the HPLC dual-mode communication module data efficient acquisition method based on edge computing provided by the application, the non-dominated sorting genetic algorithm III is used to solve the multi-objective optimization model to obtain a collection strategy solution set. A certain number of individuals are randomly generated as an initial population, each individual corresponding to a set of acquisition parameter vectors, and the population size is set to For each individual, the objective function value thereof is calculated according to the multi-objective optimization model.

[0026] The population is stratified by using a reference point-based non-dominated sorting method, and the crowding distance and reference point correlation degree of each individual are calculated.

[0027] Excellent individuals are selected into the next generation population according to the sorting results and correlation degrees, and a selection ratio is set.

[0028] The selected individuals are subjected to simulated binary crossover and polynomial mutation operations, a crossover probability and a mutation probability are set, and new individuals are generated.

[0029] The newly generated individuals are merged with the original population, and non-dominated sorting and correlation degree calculation are performed again, and individuals are selected to form the next generation population.

[0030] When the number of iterations reaches a preset maximum value or the population convergence precision meets a preset threshold, the iteration is stopped, and the non-dominated solutions in the current population are taken as the acquisition strategy solution set.

[0031] According to the HPLC dual-mode communication module data efficient acquisition method based on edge computing provided by the application, the process of selecting the optimal acquisition strategy of the current communication scene by the analytic hierarchy process includes: A hierarchical structure model is established, the target layer is to select the optimal acquisition strategy, the criterion layer includes data acquisition efficiency, communication energy consumption and data transmission reliability, and the scheme layer is all strategies in the acquisition strategy solution set.

[0032] A judgment matrix is constructed, the importance of each factor in the criterion layer is compared by using the 1-9 scale method according to the demand of the current communication scene, and the criterion layer judgment matrix and the judgment matrix of the scheme layer relative to each criterion are obtained.

[0033] The maximum eigenvalue of the judgment matrix and the corresponding eigenvector are calculated, and the eigenvector is normalized to obtain a weight vector.

[0034] Consistency check is performed, the consistency index and the consistency ratio are calculated, and when the consistency ratio meets the preset requirement, the judgment matrix meets the consistency requirement.

[0035] The comprehensive weight score of each acquisition strategy is calculated, and the acquisition strategy with the highest comprehensive weight score is selected as the optimal acquisition strategy of the current communication scene.

[0036] The application provides a high-efficiency data acquisition method for an HPLC dual-mode communication module based on edge computing, and a process for adjusting data acquisition parameters by combining a predictive control mechanism and a feedback correction mechanism comprises the following steps: An optimal acquisition strategy is received by the edge computing node, and key acquisition parameters including an acquisition frequency, a data block size and a transmission bandwidth allocation ratio are extracted.

[0037] According to the hardware characteristics and historical operation data of the HPLC dual-mode communication module, an acquisition parameter prediction model based on an ARIMA model is established.

[0038] The optimal acquisition parameters of the next period are predicted by using the acquisition parameter prediction model to obtain predicted values.

[0039] Actual effect indicators of the current data acquisition including actual acquisition efficiency, actual energy consumption and actual reliability are acquired in real time through a feedback channel.

[0040] The deviation between the predicted values and actual values is calculated, and a proportional-integral correction algorithm is used to correct the predicted parameters.

[0041] The corrected acquisition parameters are converted into control instructions and sent to the HPLC dual-mode communication module to realize dynamic adjustment of the data acquisition parameters.

[0042] On the other hand, the application also provides a high-efficiency data acquisition system for an HPLC dual-mode communication module based on edge computing, which comprises a data acquisition module, a feature extraction module, a strategy generation module, an association construction module, a strategy optimization module and an acquisition control module.

[0043] The data acquisition module is used for acquiring multi-dimensional resource data of the HPLC dual-mode communication module in real time, and the multi-dimensional resource data includes computing power data, memory resource data and communication link data.

[0044] The feature extraction module is used for denoising the multi-dimensional resource data by using an adaptive median filtering algorithm, and extracting features by using a local linear embedding algorithm to obtain computing power features, memory features and link features.

[0045] The strategy generation module is used for constructing a resource state evaluation model based on an extreme learning machine, evaluating the resource load state of the communication module in real time according to the computing power features, the memory features and the link features, and triggering data acquisition strategy adjustment according to a preset threshold rule according to the resource load state.

[0046] The association construction module is used for fusing the multi-dimensional resource data by using a federated filtering algorithm, and constructing a spatio-temporal association relationship model of the multi-dimensional resource data by using a gated recurrent unit network.

[0047] The policy optimization module is used for establishing a multi-objective optimization model of comprehensive data collection efficiency, communication energy consumption and data transmission reliability according to the space-time correlation model, solving the multi-objective optimization model by using a non-dominated sorting genetic algorithm III, and obtaining a collection strategy solution set, and selecting an optimal collection strategy of the current communication scene by using an analytic hierarchy process.

[0048] The collection and regulation module is used for converting the optimal collection strategy into a control instruction by using a distributed processing capability of the edge computing node through an HPLC dual-mode communication protocol, and adjusting data collection parameters by combining a predictive control and a feedback correction mechanism.

[0049] The application provides an HPLC dual-mode communication module data efficient collection method and system based on edge computing, which can eliminate the influence of environmental interference on data quality by collecting multi-dimensional data such as real-time computing power, memory and link in combination with adaptive denoising and feature extraction technology, can accurately judge the module load state in real time by using a limit learning mechanism to build an evaluation model, can provide a reliable basis for collection strategy adjustment, and can avoid the problems of resource waste or data loss caused by fixed parameters. By constructing a multi-objective optimization and intelligent decision mechanism, the collection efficiency, communication energy consumption and transmission reliability are comprehensively considered, data is fused through federated filtering, a space-time correlation relationship is mined through a gated recurrent unit network, a multi-objective optimization model is solved by using a non-dominated sorting genetic algorithm III, different scene requirements are adapted by combining an analytic hierarchy process, and an optimal collection strategy is selected, so that the coordinated optimization of core indicators is realized, and the limitation of single indicator optimization in the traditional scheme is broken through. By using the distributed processing advantage of edge computing, the collection strategy generation and parameter adjustment are sunk to the edge node, the dependence on the cloud is greatly reduced, the response delay is shortened, and the collection parameters are dynamically adjusted in combination with the predictive control and the feedback correction mechanism, so that stable collection performance can be maintained under the fluctuation of module resources or environmental interference. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort.

[0051] Fig. 1 is a process schematic diagram of the HPLC dual-mode communication module data efficient collection method based on edge computing provided by the embodiments of the application; Fig. 2 is a process schematic diagram of solving the multi-objective optimization model by using a non-dominated sorting genetic algorithm III in the embodiments of the application; Fig. 3The application provides a flow structure intention of a high-efficiency data acquisition system of an HPLC dual-mode communication module based on edge computing. DETAILED DESCRIPTION

[0052] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be clearly and completely described below with reference to the drawings in the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0053] The technical solutions in the present application will be described below with reference to the drawings in the present application. Figs. 1-3 The application discloses an efficient data acquisition method and system of an HPLC dual-mode communication module based on edge computing.

[0054] The application provides an efficient data acquisition method and system of an HPLC dual-mode communication module based on edge computing. The execution subject can be an efficient data acquisition method of an HPLC dual-mode communication module based on edge computing. The method comprises the following steps. Real-time acquisition of multi-dimensional resource data of the HPLC dual-mode communication module, wherein the multi-dimensional resource data comprises computing power data, memory resource data and communication link data.

[0055] The computing power data comprises CPU load rate, GPU computing power value and computing power utilization rate. The memory resource data comprises memory remaining capacity, memory read-write rate and memory occupancy rate. The communication link data comprises link bandwidth, transmission delay and data packet loss rate.

[0056] Adaptive median filtering algorithm is used for denoising of the multi-dimensional resource data, and local linear embedding algorithm is used for feature extraction to obtain computing power features, memory features and link features. The process comprises the following steps. Construction of a multi-dimensional resource data matrix. For each feature in the multi-dimensional resource data matrix, a local neighborhood similarity matrix is calculated to obtain a standardized data set and a local covariance matrix.

[0057] Singular value decomposition is performed on the local covariance matrix. According to the cumulative contribution rate of the singular values, the feature vectors corresponding to the first m largest singular values are selected as embedding features to obtain the computing power features, the memory features and the link features. The objective function of feature extraction is as follows: In the formula, xi is the i-th original multi-dimensional resource data sample, is the low-dimensional embedding vector corresponding to the i-th original multi-dimensional resource data sample, is the low-dimensional embedding vector corresponding to the i-th original multi-dimensional resource data sample, W is a projection matrix, is the local neighborhood similarity of the samples i and j, is a regularization parameter, is the total number of samples of the original multi-dimensional resource data.

[0058] The resource state evaluation model based on extreme learning machine is constructed, and according to the computing power feature, the memory feature and the link feature, the resource load state of the communication module is evaluated in real time, and according to the resource load state, the data collection strategy adjustment is triggered according to the preset threshold rule.

[0059] The process of constructing the resource state evaluation model based on extreme learning machine includes: Collect historical running multi-dimensional resource data samples of the HPLC dual-mode communication module, suppose that the collected historical data has a total of samples, each sample contains p features, which is represented as a feature matrix wherein is the feature vector of the i-th sample, and the corresponding resource load state label is The radial basis kernel function is defined as the activation function of the extreme learning machine, and the formula is represented as: In the formula, is the output value of the activation function, is the center vector of the k-th hidden layer node, is the kernel width parameter, is the input feature vector.

[0060] The optimization objective function of the extreme learning machine is constructed, and the formula is represented as: In the formula, is the output weight vector, G is the number of hidden layer nodes, b is the bias term, C is the penalty coefficient, is the slack variable, represents the number of samples, represents the output weight of the k-th hidden layer node, represents the activation output of the i-th sample at the k-th hidden layer node, represents the input feature vector of the i-th historical sample, represents the resource load state label of the i-th historical sample.

[0061] The output weight vector is solved by Moore-Penrose generalized inverse, and the formula is represented as: In the formula, represents the optimal output weight vector, H is the hidden layer output matrix, Let H be the Moore-Penrose generalized inverse, and let e be an all-1 vector, to obtain the resource status assessment model.

[0062] The federated filtering algorithm is used to fuse multi-dimensional resource data, and a spatiotemporal correlation model of the multi-dimensional resource data is constructed using a gated recurrent unit network. The process includes: The multi-dimensional resource data, after being fused through federated filtering, is sorted by time series to obtain the input sequence. in Indicates the first A multi-dimensional resource data vector at any given time, where N is the length of the time series.

[0063] The structure of the gated recurrent unit network is defined, including an input layer, a hidden layer, and an output layer. The input layer receives the input sequence. The hidden layer contains multiple gated recurrent units, which control the flow of information by updating and resetting the gates, capturing the spatiotemporal dependencies in the data. The output layer is a fully connected layer that outputs the hidden state vector of the data at each time step.

[0064] Define the state update formula for the gated loop unit: In the formula, To reset the door, To update the door, In the candidate hidden state, Currently in a hidden state. This is the weight matrix. For bias vectors, It is the sigmoid activation function. For element-wise multiplication, Let be the multi-dimensional resource data vector at time t. For the first The hidden state at any given moment.

[0065] The gated recurrent unit network is trained, and the correlation strength of multi-dimensional resource data in the time dimension and device node dimension is analyzed.

[0066] By using hidden state vectors as node features and spatiotemporal correlation coefficients between nodes as edge weights, a spatiotemporal correlation model for multi-dimensional resource data is constructed.

[0067] According to the space-time correlation model, a multi-objective optimization model of data collection efficiency, communication energy consumption and data transmission reliability is established. The non-dominated sorting genetic algorithm III is used to solve the multi-objective optimization model, and the optimal collection strategy of the current communication scene is selected by the analytic hierarchy process.

[0068] The process of establishing the multi-objective optimization model of data collection efficiency, communication energy consumption and data transmission reliability according to the space-time correlation model includes: The data collection efficiency is defined as the ratio of the effective data collection amount per unit time to the theoretical maximum collection amount, and the formula is: In the formula, is the data collection efficiency, is the effective data collection amount per unit time, is the theoretical maximum data collection amount per unit time.

[0069] The communication energy consumption is defined as the total power consumption of the HPLC dual-mode communication module, and the formula is calculated according to the transmission power, collection period and communication duration: In the formula, E is the communication energy consumption, is the transmission power, is the collection period, is the single communication duration, is the idle power, and T is the statistical time window.

[0070] According to the data packet loss rate, transmission delay and data integrity check result, a data transmission reliability model is established, and the formula is: In the formula, R is the data transmission reliability, is the data packet loss rate, is the actual transmission delay, is the allowed maximum delay, is the data integrity check pass rate, is the weight coefficient and satisfies According to the data collection efficiency, communication energy consumption and data transmission reliability, a comprehensive objective function is established as a multi-objective optimization model, and the formula is: In the formula, is the multi-objective optimization target vector, u is the collection parameter vector, including the collection frequency, data block size and transmission bandwidth allocation ratio, is the negative value of the collection efficiency, is the communication energy consumption, Reliability complementary value.

[0071] The process of solving the multi-objective optimization model by using the non-dominated sorting genetic algorithm III to obtain the solution set of the collection strategy includes: A certain number of individuals are randomly generated as the initial population, each individual corresponds to a set of collection parameter vectors u, and the population size is set to For each individual, the objective function value is calculated according to the multi-objective optimization model The non-dominated sorting method based on reference points is used to stratify the population, and the crowding distance and reference point correlation of each individual are calculated.

[0072] According to the sorting results and correlation, select excellent individuals into the next generation population, and the selection ratio is set to Simulated binary crossover and polynomial mutation operations are performed on the selected individuals, the crossover probability is set to , the mutation probability is set to , and new individuals are generated.

[0073] The newly generated individuals are combined with the original population, and non-dominated sorting and correlation calculation are performed again to select individuals to form the next generation population.

[0074] When the number of iterations reaches the preset maximum value or the population convergence precision meets the preset threshold, stop iteration, and the non-dominated solutions in the current population are used as the solution set of the collection strategy.

[0075] The process of selecting the optimal collection strategy for the current communication scenario by using the analytic hierarchy process includes: A hierarchical model is established, the target layer is to select the optimal collection strategy, the criterion layer includes data collection efficiency, communication energy consumption, and data transmission reliability, and the scheme layer is all strategies in the solution set of the collection strategy.

[0076] Construct a judgment matrix, according to the requirements of the current communication scenario, use the 1-9 scale method to compare the importance of each factor in the criterion layer, and obtain the criterion layer judgment matrix A and the judgment matrix Calculate the maximum eigenvalue and the corresponding eigenvector of the judgment matrix, and normalize the eigenvector to obtain the weight vector w.

[0077] Conduct consistency check, calculate the consistency index and the consistency ratio , where RI is the average random consistency index, is the order of the judgment matrix, and when When the judgment matrix meets the consistency requirement.

[0078] The comprehensive weight score of each acquisition strategy is calculated, and the formula is: In the formula, is the comprehensive score of the jth acquisition strategy, is the weight of the kth criterion, is the normalized weight of the jth strategy with respect to the kth criterion. The acquisition strategy with the highest comprehensive weight score is selected as the optimal acquisition strategy for the current communication scenario.

[0079] The distributed processing capability of the edge computing node is utilized to convert the optimal acquisition strategy into control instructions through the HPLC dual-mode communication protocol, and the data acquisition parameters are adjusted by combining the predictive control and feedback correction mechanism. The process includes:

[0080] The optimal acquisition strategy is received by the edge computing node, and the key acquisition parameters are extracted, including the acquisition frequency, data block size, and transmission bandwidth allocation ratio.

[0081] According to the hardware characteristics and historical running data of the HPLC dual-mode communication module, the hardware characteristics include hardware-related parameters such as CPU model and core number, CPU benchmark frequency, memory-related hardware parameters such as total physical memory capacity, memory read / write rate upper limit, communication link-related hardware parameters such as HPLC dual-mode communication frequency range, and storage interface-related hardware parameters such as local cache or storage chip capacity. Historical acquisition parameter data and historical resource load data are included. An acquisition parameter prediction model based on an ARIMA model is established, and the formula is: In the formula, is an autoregressive operator, is a moving average operator, d is a difference order, and L is a lag operator, is a white noise sequence at time t, is a d-order difference of the acquisition parameter sequence. The optimal acquisition parameter for the next period is predicted using the acquisition parameter prediction model, and the predicted value is obtained.

[0082] The actual effect indicators of the current data acquisition are collected in real time through the feedback channel, including the actual acquisition efficiency , actual energy consumption , and actual reliability ​​The deviation of the predicted value from the actual value is calculated, and a proportional integral correction algorithm is used to correct the prediction parameters, which is expressed by the formula: In the formula, is the correction amount, is the proportional coefficient, is the integral coefficient, is the reference index vector, is the actual index vector, is the final collection parameter after correction, is the predicted value of the collection parameter in the next period.

[0083] The corrected collection parameter is converted into a control instruction and sent to the HPLC dual-mode communication module to realize dynamic adjustment of the data collection parameter.

[0084] In summary, the embodiment provides an HPLC dual-mode communication module data efficient collection method based on edge computing. By collecting real-time computing power, memory, and link multi-dimensional data, combining adaptive denoising and feature extraction technology, the influence of environmental interference on data quality is eliminated. An evaluation model is built using extreme learning machines to accurately determine the module load state in real time, providing a reliable basis for collection strategy adjustment and avoiding resource waste or data loss caused by fixed parameters. By building a multi-objective optimization and intelligent decision-making mechanism, the collection efficiency, communication energy consumption, and transmission reliability are comprehensively considered. Through federated filtering to fuse data, a gating recurrent unit network is used to mine the spatio-temporal correlation, and a non-dominated sorting genetic algorithm III is used to solve the multi-objective optimization model. Combined with the analytic hierarchy process, different scene requirements are adapted to filter out the optimal collection strategy, realize the coordinated optimization of core indicators, and break through the limitations of single indicator optimization in traditional solutions. Through the distributed processing advantage of edge computing, the collection strategy generation and parameter adjustment are sunk to the edge node, greatly reducing the dependence on the cloud and shortening the response delay. At the same time, combined with the prediction control and feedback correction mechanism, the collection parameters are dynamically adjusted to ensure stable collection performance under module resource fluctuations or environmental interference.

[0085] Based on the same overall inventive concept, the present application also protects an HPLC dual-mode communication module data efficient collection system based on edge computing. The HPLC dual-mode communication module data efficient collection system based on edge computing provided by the present application is described below. The HPLC dual-mode communication module data efficient collection system based on edge computing described below can be mutually corresponding and referred to with the HPLC dual-mode communication module data efficient collection method based on edge computing described above.

[0086] The HPLC dual-mode communication module data efficient acquisition system based on edge computing comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor comprises a data acquisition module, a feature extraction module, a strategy generation module, an association construction module, a strategy optimization module and an acquisition regulation module.

[0087] The data acquisition module is used for acquiring multi-dimensional resource data of the HPLC dual-mode communication module in real time, and the multi-dimensional resource data comprises computing power data, memory resource data and communication link data.

[0088] The feature extraction module is used for denoising the multi-dimensional resource data by using an adaptive median filtering algorithm, and performing feature extraction by using a local linear embedding algorithm to obtain computing power features, memory features and link features.

[0089] The strategy generation module is used for constructing a resource state evaluation model based on an extreme learning machine, evaluating a resource load state of the communication module in real time according to the computing power features, the memory features and the link features, and triggering data acquisition strategy adjustment according to a preset threshold rule according to the resource load state.

[0090] The association construction module is used for fusing the multi-dimensional resource data by using a federated filtering algorithm, and constructing a spatio-temporal association relationship model of the multi-dimensional resource data by using a gated recurrent unit network.

[0091] The strategy optimization module is used for establishing a multi-objective optimization model of comprehensive data acquisition efficiency, communication energy consumption and data transmission reliability according to the spatio-temporal association relationship model, solving the multi-objective optimization model by using a non-dominated sorting genetic algorithm III to obtain an acquisition strategy solution set, and selecting an optimal acquisition strategy of a current communication scene by using an analytic hierarchy process.

[0092] The acquisition regulation module is used for converting the optimal acquisition strategy into a control instruction by using an HPLC dual-mode communication protocol by utilizing distributed processing capability of an edge computing node, and adjusting data acquisition parameters in combination with a predictive control and feedback correction mechanism.

[0093] Those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary universal hardware platforms according to the description of the above embodiments, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, and the computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the method described in each embodiment or some parts of the embodiment.

[0094] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An efficient data acquisition method for HPLC dual-mode communication modules based on edge computing, characterized in that, The application relates to a method for optimizing data collection strategy of HPLC dual-mode communication module. Real-time acquisition of multi-dimensional resource data of an HPLC dual-mode communication module, wherein the multi-dimensional resource data comprises computing power data, memory resource data and communication link data; Adaptive median filtering algorithm is used to denoise the multi-dimensional resource data, and local linear embedding algorithm is used for feature extraction to obtain computing power features, memory features and link features; A resource state evaluation model based on extreme learning machine is constructed, the resource load state of the communication module is evaluated in real time according to the computing power features, memory features and link features, and the data collection strategy is adjusted according to the resource load state and a preset threshold rule; Federal filtering algorithm is used to fuse the multi-dimensional resource data, and a spatio-temporal correlation model of the multi-dimensional resource data is constructed by using a gated recurrent unit network; A multi-objective optimization model of comprehensive data collection efficiency, communication energy consumption and data transmission reliability is established according to the spatio-temporal correlation model, a non-dominated sorting genetic algorithm III is used to solve the multi-objective optimization model, an acquisition strategy solution set is obtained, and the optimal acquisition strategy of the current communication scene is selected by using an analytic hierarchy process; The optimal acquisition strategy is converted into a control instruction through an HPLC dual-mode communication protocol by using the distributed processing capacity of an edge computing node, and the data collection parameters are adjusted by combining a predictive control and a feedback correction mechanism.

2. The edge computing based HPLC dual-mode communication module data efficient acquisition method according to claim 1, characterized in that, The computing power data comprises CPU load rate, GPU computing capacity value and computing power utilization rate; The memory resource data comprises memory remaining capacity, memory read-write rate and memory occupancy rate; and the communication link data comprises link bandwidth, transmission delay and data packet loss rate.

3. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of feature extraction by using the local linear embedding algorithm comprises the following steps: A multi-dimensional resource data matrix is constructed, the local neighborhood similarity matrix of each feature in the multi-dimensional resource data matrix is calculated, and a standardized data set and a local covariance matrix are obtained; The local covariance matrix is singular value decomposed, the feature vectors corresponding to the first m largest singular values are selected as embedding features according to the cumulative contribution rate of the singular values, and computing power features, memory features and link features are obtained.

4. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of constructing the resource state evaluation model based on the extreme learning machine comprises the following steps: Historical running multi-dimensional resource data samples of the HPLC dual-mode communication module are collected, and resource load state labels corresponding to each data sample are labeled; A radial basis kernel function is defined as an activation function of the extreme learning machine, and an optimization objective function of the extreme learning machine is constructed; An output weight vector is solved by using Moore-Penrose generalized inverse, and a resource state evaluation model is obtained.

5. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of constructing the spatio-temporal correlation model of the multi-dimensional resource data by using the gated recurrent unit network comprises the following steps: The multi-dimensional resource data fused by the federal filtering is sorted according to time sequence, and an input sequence is obtained; The structure of the gated recurrent unit network is set, including an input layer, a hidden layer and an output layer; the input layer is used for receiving the input sequence; the hidden layer comprises a plurality of gated recurrent units, and the time-space dependent relationship in the data is captured by updating the gate and the reset gate to control the flow of information; the output layer adopts a full connection layer, and a hidden state vector of each time data is outputted; and The gate cycle unit network is trained to analyze the correlation strength of multi-dimensional resource data in the time dimension and the device node dimension. A space-time correlation relationship model of multi-dimensional resource data is constructed by taking the hidden state vector as a node feature and the space-time correlation coefficient between nodes as an edge weight.

6. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of establishing a multi-objective optimization model of comprehensive data collection efficiency, communication energy consumption and data transmission reliability according to the space-time correlation relationship model comprises: The data collection efficiency is defined as the ratio of the amount of data effectively collected in a unit time to the theoretical maximum collection amount; The communication energy consumption is defined as the total power consumption of the HPLC dual-mode communication module, which is calculated according to the transmission power, collection period and communication duration; A data transmission reliability model is established according to the data packet loss rate, transmission delay and data integrity check result; A comprehensive objective function is established according to the data collection efficiency, communication energy consumption and data transmission reliability as a multi-objective optimization model.

7. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of solving the multi-objective optimization model by using the non-dominated sorting genetic algorithm III to obtain a collection strategy solution set comprises: A certain number of individuals are randomly generated as an initial population, each individual corresponding to a set of acquisition parameter vectors, the population size being set to For each individual, the objective function value is calculated according to the multi-objective optimization model; The population is stratified by using the reference point-based non-dominated sorting method, and the crowding distance and reference point correlation degree of each individual are calculated; According to the sorting result and the correlation degree, excellent individuals are selected into the next generation population, and the selection ratio is set; The selected individuals are subjected to simulated binary crossover and polynomial mutation operations, and the crossover probability and mutation probability are set to generate new individuals; The newly generated individuals are combined with the original population, and non-dominated sorting and closeness calculation are performed again, and the next generation population is selected by Individuals When the number of iterations reaches the preset maximum value or the population convergence precision meets the preset threshold, the iteration is stopped, and the non-dominated solutions in the current population are taken as the collection strategy solution set.

8. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of selecting the optimal collection strategy for the current communication scenario by using the analytic hierarchy process comprises: A hierarchical model is established, the target layer is to select the optimal collection strategy, the criterion layer includes data collection efficiency, communication energy consumption and data transmission reliability, and the scheme layer is all strategies in the collection strategy solution set; A judgment matrix is constructed, the importance of each factor in the criterion layer is compared by using the 1-9 scale method according to the requirements of the current communication scenario, and the criterion layer judgment matrix and the judgment matrix of the scheme layer relative to each criterion are obtained; The maximum eigenvalue of the judgment matrix and the corresponding eigenvector are calculated, and the eigenvector is normalized to obtain the weight vector; Consistency test is performed, the consistency index and consistency ratio are calculated, and when the consistency ratio meets the preset requirement, the judgment matrix meets the consistency requirement; The comprehensive weight score of each collection strategy is calculated, and the collection strategy with the highest comprehensive weight score is selected as the optimal collection strategy for the current communication scenario.

9. The edge computing based HPLC dual-mode communication module data efficient acquisition method of claim 1, wherein, The process of adjusting the data collection parameters by combining the predictive control and feedback correction mechanism comprises: The optimal collection strategy is received by the edge computing node, and the key collection parameters including the collection frequency, data block size and transmission bandwidth allocation ratio are extracted; A collection parameter prediction model based on the ARIMA model is established according to the hardware characteristics and historical operation data of the HPLC dual-mode communication module; The optimal collection parameters for the next period are predicted by using the collection parameter prediction model to obtain the predicted values; Real-time acquisition of the actual effect index of current data acquisition through the feedback channel, including actual acquisition efficiency, actual energy consumption and actual reliability; Calculate the deviation between the predicted value and the actual value, and correct the prediction parameters using a proportional-integral correction algorithm; Convert the corrected acquisition parameters into control instructions and send them to the HPLC dual-mode communication module to realize dynamic adjustment of data acquisition parameters.

10. An edge computing based HPLC dual-mode communication module data efficient acquisition system comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the edge computing-based HPLC dual-mode communication module data efficient acquisition method of any one of claims 1-9 when executing the program.

Images