Industrial internet edge multi-source heterogeneous data collaborative acquisition method and device

Abstract

The invention provides an industrial internet edge multi-source heterogeneous data collaborative collection method and device. Fusing the time sequence change gradient of the information transmission quantity of each data source in the edge nodes into the data value density of the corresponding edge nodes by using the cross-source data difference of each edge node in the industrial internet and the service function weight of each data source; determining the network resource constraint of the cloud center end through the transmission delay characteristics between the cloud center end and each edge node and the fluctuation characteristics of the network bandwidth in the cloud center end; and according to all the data value densities, network resource constraints are taken as constraint conditions of each edge node in the industrial internet, the total data value transmitted to the cloud center end is maximized under the constraint conditions, and personalized optimization is performed on the data acquisition frequency of each edge node in the industrial internet. Based on the above scheme, acquisition frequency optimization under edge node data value and network resource delay cooperative constraint can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of edge computing, and more particularly, to a multi-source heterogeneous data collaborative collection method and device for industrial internet edge. BACKGROUND

[0002] Industrial internet is an ecological system of deep integration of industrial system and internet technology, which realizes the comprehensive interconnection of man, machine and object through intelligent sensing, internet of things, big data and artificial intelligence. The core of industrial internet is to build a data-driven intelligent analysis and optimization closed loop. Through deep collection and aggregation of industrial full-factor, full-industry chain and full-value chain data, real-time perception, accurate execution and intelligent decision-making of the production process are realized.

[0003] In the traditional industrial network environment, the allocation strategy (such as bandwidth scheduling) of network resources only depends on the instant indicators of the underlying network state. Although the allocation method based on pure network dimension can guarantee the connectivity and basic performance of the link, it completely ignores a key dimension, the actual value of the upper-layer business data, which causes high-value real-time control instructions and low-value daily status monitoring data to compete for the same bandwidth resources, resulting in delayed response of critical business and a large amount of network capacity occupied by non-critical data, thereby causing the disconnection between resource allocation and business demand, and further causing inefficient use of transmission resources, which cannot guarantee the deterministic transmission of high-value data in complex industrial scenarios, thereby limiting the improvement of the overall system efficiency. Therefore, how to realize the collection frequency optimization under the collaborative constraint of edge node data value and network resource delay has become a difficult problem in the industry. SUMMARY

[0004] The present application provides a multi-source heterogeneous data collaborative collection method and device for industrial internet edge, which can realize the collection frequency optimization under the collaborative constraint of edge node data value and network resource delay, thereby improving the value efficiency of industrial internet data transmission.

[0005] In a first aspect, the present application provides a multi-source heterogeneous data collaborative collection method for industrial internet edge, comprising: Obtaining multi-source data transmission information between a cloud center and each edge node in the industrial internet, and monitoring device state information of the cloud center; For each edge node in the industrial internet, calculating the time sequence change gradient of the transmission amount of each data source information in the edge node through the multi-source data transmission information of the edge node, and using the cross-source data difference in the edge node and the business function weight of each data source to fuse all the time sequence change gradients into the data value density of the edge node, and then obtaining the data value density of each edge node in the industrial internet; extract transmission delay features between the cloud center and each edge node from the device state information, delay constrain transmission resources between the cloud center and each edge node through each transmission delay feature and fluctuation features of network bandwidth in the cloud center, and obtain network resource constraints of the cloud center; According to all data value densities and taking the network resource constraints as constraint conditions of each edge node in the industrial internet, maximize total data value transmitted to the cloud center, and perform personalized optimization on data collection frequencies of each edge node in the industrial internet.

[0006] In some embodiments, calculating the time sequence change gradient of the information transmission amount of each data source in the edge node through the multi-source data transmission information of the edge node specifically includes: For each data source in the edge node, obtaining the information transmission amount of the data source at each sampling point from the multi-source data transmission information of the edge node; Determining the time sequence change gradient of the data source information transmission amount through all information transmission amounts, and obtaining the time sequence change gradient of the information transmission amount of each data source in the edge node.

[0007] In some embodiments, fusing all time sequence change gradients into the data value density of the edge node using the cross-source data difference in the edge node and the business function weight of each data source specifically includes: Determining the cross-source data difference in the edge node, and obtaining the business function weight of each data source; Fusing all time sequence change gradients into the value density initial value of the edge node through each business function weight; Differentially adjusting the value density initial value based on the cross-source data difference, and obtaining the data value density of the edge node.

[0008] In some embodiments, extracting transmission delay features between the cloud center and each edge node from the device state information specifically includes: For each edge node, obtaining multiple echo request data packets between the cloud center and the edge node from the device state information; Extracting round-trip delay values of each echo request data packet; Determining the transmission delay features between the cloud center and the edge node through all round-trip delay values, and obtaining the transmission delay features between the cloud center and each edge node.

[0009] In some embodiments, delay constraining transmission resources between the cloud center and each edge node through each transmission delay feature and fluctuation features of network bandwidth in the cloud center, and obtaining network resource constraints of the cloud center specifically includes: Extracting fluctuation features of network bandwidth in the cloud center; determine delay constraint amount of transmission resource between the cloud center and each edge node according to the fluctuation feature and each transmission delay feature; determine network resource constraint of the cloud center through all the delay constraint amount.

[0010] In some embodiments, the data acquisition frequency of each edge node in the industrial internet is individually optimized by maximizing the total data value transmitted to the cloud center according to all the data value densities and taking the network resource constraint as a constraint condition of each edge node in the industrial internet, which specifically includes: initialize a node constraint model based on linear programming; take all the data value densities as value decision parameters of the node constraint model; maximize the total data value transmitted to the cloud center as a constraint target of the node constraint model; take the network resource constraint as a resource constraint condition of each edge node in the node constraint model; use the node constraint model to individually optimize the data acquisition frequency of each edge node in the industrial internet.

[0011] In some embodiments, the edge node is an embedded gateway based on ARM Cortex-A7 architecture.

[0012] In a second aspect, the present application provides an industrial internet edge multi-source heterogeneous data collaborative acquisition device, which includes: an acquisition module, configured to acquire multi-source data transmission information between a cloud center and each edge node in an industrial internet, and monitor device state information of the cloud center; a processing module, configured to calculate time sequence change gradient of each data source information transmission amount in each edge node in the industrial internet through multi-source data transmission information of the edge node, fuse all the time sequence change gradients into data value density of the edge node using cross-source data difference in the edge node and business function weight of each data source, and further obtain data value density of each edge node in the industrial internet; the processing module is further configured to extract transmission delay feature between the cloud center and each edge node from the device state information, delay constrain transmission resource between the cloud center and each edge node through each transmission delay feature and fluctuation feature of network bandwidth in the cloud center, and obtain network resource constraint of the cloud center; an execution module, configured to individually optimize data acquisition frequency of each edge node in the industrial internet by maximizing the total data value transmitted to the cloud center according to all the data value densities and taking the network resource constraint as a constraint condition of each edge node in the industrial internet.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the above-mentioned multi-source heterogeneous data collaborative collection method of industrial internet edge.

[0014] In a fourth aspect, the present application provides a computer readable storage medium, which stores instructions or codes, when the instructions or codes are run on a computer, the computer executes the above-mentioned multi-source heterogeneous data collaborative collection method of industrial internet edge.

[0015] The technical scheme provided by the embodiments of the present application has the following beneficial effects: In the multi-source heterogeneous data collaborative collection method and device of industrial internet edge provided by the present application, the multi-source data transmission information between the cloud center end and each edge node in the industrial internet is obtained, and the device state information of the cloud center end is monitored; for each edge node in the industrial internet, the time sequence change gradient of the transmission amount of each data source information in the edge node is calculated through the multi-source data transmission information of the edge node, the data value density of the edge node is fused from all the time sequence change gradients using the cross-source data difference in the edge node and the business function weight of each data source, and then the data value density of each edge node in the industrial internet is obtained; the transmission delay characteristics between the cloud center end and each edge node are extracted from the device state information, the transmission resources between the cloud center end and each edge node are subjected to delay constraint through each transmission delay characteristic and the fluctuation characteristics of the network bandwidth in the cloud center end, and the network resource constraint of the cloud center end is obtained; according to all the data value densities, and taking the network resource constraint as the constraint condition of each edge node in the industrial internet, the total data value transmitted to the cloud center end is maximized, and the data collection frequency of each edge node in the industrial internet is individually optimized.

[0016] It can be seen that, in the present application, according to all the data value density, and taking the network resource constraint as the constraint condition of each edge node in the industrial internet, the total data value transmitted to the cloud center is maximized, and the data acquisition frequency of each edge node in the industrial internet is personalized optimized; first, the data value density is determined, that is, the data quantization standard for business value is obtained, the data value density which can dynamically reflect the importance and timeliness value of data business is determined, which can make the industrial internet device break through the traditional extensive management mode based on only the size of data volume, accurately identify the higher value data block of the edge node in the industrial internet device at the current time in the upper layer application, for example, a sensor data stream with a sudden sharp increase in transmission volume, which has a large change gradient, if supplemented with a higher business function weight, it will obtain a high value density score, and then provide a clear value orientation for subsequent optimization decision, so that the industrial internet device can distinguish between high value data and ordinary data, and lay the foundation for improving the value efficiency; then, the network resource constraint is determined, that is, the accurate boundary of the cloud service capability is obtained, so as to anchor the value transmission demand on the fluctuation characteristics of the physical reality, quantify the maximum capability of the cloud to receive data from each edge node within a certain time period through the constraint model, clearly define the transmission boundary of data value, improve the feasibility of the optimization scheme, for example, the industrial internet device can identify the network delay surge or bandwidth drop of the current edge node, so as to clearly inform the optimization algorithm that it is not feasible to ask for too much data from the edge node at present, and then the abstract network state can be converted into a concrete mathematical constraint condition which can be directly processed by the optimization algorithm, ensuring that any optimization decision on the acquisition frequency is made within the range allowed by the actual network physical condition, and the feasibility of the scheme is ensured; in summary, based on the above scheme, the acquisition frequency optimization under the coordination constraint of edge node data value and network resource delay can be realized, so as to improve the value efficiency of industrial internet data transmission. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 is an exemplary flowchart of the multi-source heterogeneous data collaborative acquisition method of the industrial internet edge according to some embodiments of the present application; Figure 2 is a flowchart for determining network resource constraints according to some embodiments of the present application; Figure 3 is a structural schematic diagram of collaborative collection of multi-source heterogeneous data of an industrial internet edge according to some embodiments of the present application; Figure 4 is a structural schematic diagram of a computer device for implementing a method for collaborative collection of multi-source heterogeneous data of an industrial internet edge according to some embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in combination with the drawings of the specification and specific embodiments.

[0020] Reference Figure 1 The figure is an exemplary flowchart of a method for collaborative collection of multi-source heterogeneous data of an industrial internet edge according to some embodiments of the present application, which mainly includes the following steps: In step 101, the multi-source data transmission information between the cloud center and each edge node in the industrial internet is obtained, and the device state information of the cloud center is monitored.

[0021] It should be noted that in the present application, the edge node is an embedded gateway based on ARM Cortex-A7 architecture; the multi-source data transmission information refers to a set of transmission parameters generated in the process of transmitting data from various edge devices to the cloud center in the industrial internet; and the device state information refers to the real-time running parameters of the cloud center server itself and its network interface.

[0022] In specific implementation, first, the multi-source data transmission information between the cloud center and each edge node in the industrial internet is obtained, which is achieved in the following manner, i.e., by deploying a dedicated data collection agent program on the cloud center, which can listen to and parse network data packets from all edge nodes at each sampling point (default 1s once), and by counting the source address, target address, size and timestamp of each data packet, the uplink data flow, data transmission duration and data transmission rate of each edge node in unit time are automatically calculated, so that the set of all uplink data flow, data transmission duration and data transmission rate is the multi-source data transmission information between the cloud center and each edge node in the industrial internet; then, it is completed by calling the operating system interface and hardware management interface of the cloud center server, the cloud center queries the central processor usage, memory occupancy, disk read / write speed and network interface card bandwidth utilization and connection state of the server at regular intervals, so that the set of central processor usage, memory occupancy, disk read / write speed and network interface card bandwidth utilization and connection state is the device state information of the cloud center.

[0023] In step 102, for each edge node in the industrial internet, the time sequence change gradient of the information transmission amount of each data source in the edge node is calculated by the multi-source data transmission information of the edge node, the time sequence change gradient of all is fused into the data value density of the edge node by using the cross-source data difference in the edge node and the business function weight of each data source, and then the data value density of each edge node in the industrial internet is obtained.

[0024] In some embodiments, the time sequence change gradient of the information transmission amount of each data source in the edge node calculated by the multi-source data transmission information of the edge node can be realized by the following steps: For each data source in the edge node, the information transmission amount of the data source at each sampling point is obtained from the multi-source data transmission information of the edge node; The time sequence change gradient of the information transmission amount of the data source is determined by all the information transmission amounts, and then the time sequence change gradient of the information transmission amount of each data source in the edge node is obtained.

[0025] It should be noted that in this application, the time sequence change gradient represents the rate of change of the information transmission amount of a data source with time; in specific implementation, first, for each data source in the edge node, the information transmission amount of the data source at each sampling point is obtained from the multi-source data transmission information of the edge node, which can be realized by the following way, that is, for each data source in the edge node, all data packets in the multi-source data transmission information of the edge node are obtained, so as to analyze the size and data source of each data packet, all data packets of the data source are screened out from all data packets, and the total size of all data packets of the data source at each sampling point is counted as the information transmission amount of the corresponding sampling point, that is, the information transmission amount of the data source at each sampling point is obtained, which is the size of the data transmitted from a data source to the edge node at a single sampling time point; then, the time sequence change gradient of the information transmission amount of the data source is determined by all the information transmission amounts, and then the time sequence change gradient of the information transmission amount of each data source in the edge node is obtained, which can be realized by the following way, that is, the mean value of the difference between the information transmission amounts of adjacent sampling points is calculated as the time sequence change gradient of the information transmission amount of the data source, and the time sequence change gradient of the information transmission amount of each data source in the edge node is obtained by the above-mentioned way.

[0026] In some embodiments, the time sequence change gradient of all is fused into the data value density of the edge node by using the cross-source data difference in the edge node and the business function weight of each data source can be realized by the following steps: The cross-source data difference in the edge node is determined, and the business function weight of each data source is obtained; All the time sequence change gradients are fused into the value density initial value of the edge node by the business function weight; differentially adjust the value density initial value based on the cross-source data difference, to obtain the data value density of the edge node.

[0027] It should be noted that in the present application, the data value density is a quantitative index for representing the overall edge node comprehensive data value level; the cross-source data difference is a quantitative index for measuring the inconsistency of data behaviors between multiple different data sources connected in the same edge node; the business function weight is a numerical coefficient representing the importance of each data source; and the value density initial value is a quantitative value reflecting the data value of all data sources in the edge node.

[0028] In specific implementation, first, the cross-source data difference in the edge node is determined, and the business function weight of each data source can be realized in the following manner, that is, all time sequence change gradients of all data sources under the edge node within a specified time period (default 24h) are collected, and the standard deviation of all time sequence change gradients is calculated as the cross-source data difference; the business function weight corresponding to each data source is read from the preset configuration database; then, all time sequence change gradients are fused into the value density initial value of the edge node through each business function weight, which can be realized in the following manner, that is, each business function weight is taken as a weight, and the weighted sum of all time sequence change gradients is calculated as the value density initial value of the edge node; finally, the value density initial value is differentially adjusted based on the cross-source data difference, to obtain the data value density of the edge node, which can be realized in the following manner, that is, the product of the cross-source data difference and the value density initial value is taken as the data value density of the edge node.

[0029] In step 103, transmission delay features between the cloud center and each edge node are extracted from the device state information, and transmission resources between the cloud center and each edge node are delay-constrained through each transmission delay feature and the fluctuation feature of the network bandwidth in the cloud center, to obtain the network resource constraint of the cloud center.

[0030] In some embodiments, the transmission delay features between the cloud center and each edge node can be extracted from the device state information in the following steps: For each edge node, a plurality of echo request data packets between the cloud center and the edge node are obtained from the device state information; The round-trip delay values of each echo request data packet are extracted; The transmission delay features between the cloud center and the edge node are determined through all round-trip delay values, and the transmission delay features between the cloud center and each edge node are obtained.

[0031] It should be noted that in the present application, the transmission delay characteristic is a representative value capable of representing the overall condition and fluctuation characteristic of the network delay; the echo request data packet refers to a special network data packet sent by the cloud center to the target edge node for detecting the network connectivity and performance; and the round-trip delay value refers to the total length of time experienced by an echo request data packet from the time when the cloud center sends the data packet to the time when the data packet is returned by the target edge node and finally received by the cloud center.

[0032] In a specific implementation, first, for each edge node, the echo request data packets between the cloud center and the edge node are obtained from the device state information in the following manner: a network detection scheduler is built in the cloud center, which sends small data packets to the IP address of each edge node to be monitored in a preset time interval. The data packets are specially used for network detection, and the protocol type is set to the echo request type in the Internet Control Message Protocol (ICMP), and the core function is to be answered and returned by the target node. Then, the round-trip delay value of each echo request data packet is obtained in the following manner: the cloud center records the accurate sending time stamp when sending each echo request data packet, and records the receiving time stamp when receiving the corresponding answer data packet. The difference between the receiving time stamp and the sending time stamp is the round-trip delay value of this detection, which is in milliseconds. All successful answer detection packets will generate a corresponding delay value, and the round-trip delay value of each echo request data packet is obtained. Finally, the transmission delay characteristic between the cloud center and the edge node is determined by all the round-trip delay values, and the transmission delay characteristic between the cloud center and each edge node is obtained in the following manner: the 95th percentile of all round-trip delay values is taken as the transmission delay characteristic between the cloud center and the edge node. Through the above manner, the transmission delay characteristic between the cloud center and each edge node is obtained.

[0033] In some embodiments, the transmission resources between the cloud center and each edge node are delay-constrained by the transmission delay characteristic and the fluctuation characteristic of the network bandwidth in the cloud center, and the network resource constraint of the cloud center is obtained, as shown in the following formula: Figure 2 The figure is a flowchart for determining the network resource constraint in some embodiments of the present application. The network resource constraint in the present embodiment can be achieved in the following steps: In step 1031, the fluctuation characteristic of the network bandwidth in the cloud center is extracted. In step 1032, the delay constraint amount of the transmission resources between the cloud center and each edge node is determined according to the fluctuation characteristic and the transmission delay characteristic. In step 1033, the network resource constraint of the cloud center end is determined by all delay constraint amounts.

[0034] It should be noted that in the present application, the network resource constraint is a global constraint condition set composed of all link delay constraint amounts; the fluctuation feature is a statistical index for quantifying the stability of the uplink network bandwidth of the cloud center end; and the delay constraint amount is an upper limit index of resource occupation of the network link between the cloud center end and each edge node.

[0035] In specific implementation, first, the fluctuation feature of the network bandwidth of the cloud center end can be realized by the following manner, that is, the cloud center end continuously monitors the real-time bandwidth data of its uplink network interface, records the bandwidth sampling value of the cloud center end every fixed time interval (default 1s), and takes the standard deviation of all bandwidth sampling values in a specified time period (default 24h) as the fluctuation feature of the network bandwidth of the cloud center end; then, the delay constraint amount of the transmission resource between the cloud center end and each edge node can be realized by the following manner according to the fluctuation feature and each transmission delay feature, that is, for each edge node, the transmission delay feature of the edge node is multiplied by the fluctuation feature of the network bandwidth, and then multiplied by an empirical reference constant, and the calculation result is the delay constraint amount of the edge node; through the above calculation process, for the link with high delay or when the entire network fluctuates greatly, the resource usage amount must be more strictly limited, and therefore the calculated delay constraint amount value will be smaller, indicating tighter constraint; and through the above manner, the delay constraint amount of the transmission resource between the cloud center end and each edge node can be obtained; finally, the network resource constraint of the cloud center end can be realized by the following manner through all delay constraint amounts, that is, the set of all delay constraint amounts is taken as the network resource constraint of the cloud center end.

[0036] In step 104, according to all data value densities and taking the network resource constraint as the constraint condition of each edge node in the industrial internet, the total data value transmitted to the cloud center end is maximized, and the data acquisition frequency of each edge node in the industrial internet is individually optimized.

[0037] In some embodiments, according to all data value densities and taking the network resource constraint as the constraint condition of each edge node in the industrial internet, the total data value transmitted to the cloud center end is maximized, and the data acquisition frequency of each edge node in the industrial internet is individually optimized, which can be realized by the following steps: Initialize a node constraint model based on linear programming; Take all data value densities as value decision parameters of the node constraint model; Take the maximization of the total data value transmitted to the cloud center end as the constraint target of the node constraint model; The network resource constraint is taken as a resource constraint condition of each edge node in the node constraint model. The node constraint model is used to individually constrain the data collection frequency of each edge node in the industrial internet.

[0038] It should be noted that in the present application, the node constraint model is a mathematical optimization framework based on linear programming theory. The node constraint model finds the optimal resource allocation scheme under the given system resource limit. The node constraint model defines the data collection frequency of each edge node as a decision variable, and takes the data value density representing the node data value level as the coefficient of the decision variable in the objective function, so as to quantify the contribution rate of different frequency configurations to the total value. The constraint target is set to maximize the weighted sum of all node data value densities, that is, to maximize the overall data value output of the system. At the same time, the node constraint model converts the network resource constraint into a set of linear inequalities, accurately representing that the bandwidth resource occupied by each node data flow should not exceed the upper limit specified by the delay constraint, and the total consumption of all node resources should not exceed the total bandwidth capacity of the cloud center. By using optimization algorithms such as the simplex method to solve the node constraint model, a set of optimal collection frequencies can be obtained, which can maximize the total data value of the system under the premise of meeting all network constraints, thereby realizing individualized and accurate regulation based on node value difference and resource status.

[0039] In addition, another aspect of the present application, in some embodiments, the present application provides an industrial internet edge multi-source heterogeneous data collaborative collection device, referring to Figure 3 The figure is a structural schematic diagram of industrial internet edge multi-source heterogeneous data collaborative collection according to some embodiments of the present application. The industrial internet edge multi-source heterogeneous data collaborative collection includes an acquisition module 201, a processing module 202 and an execution module 203, which are described as follows: The acquisition module 201 is mainly used to acquire multi-source data transmission information between the cloud center and each edge node in the industrial internet, and monitor the device state information of the cloud center. The processing module 202 is used to calculate the time sequence change gradient of each data source information transmission amount in the edge node through the multi-source data transmission information of the edge node for each edge node in the industrial internet, and fuse all the time sequence change gradients into the data value density of the edge node using the cross-source data difference in the edge node and the business function weight of each data source, and then obtain the data value density of each edge node in the industrial internet. It should be noted that the processing module 202 is also used to extract the transmission delay feature between the cloud center and each edge node from the device state information, delay constraint is performed on the transmission resource between the cloud center and each edge node through each transmission delay feature and the fluctuation feature of the network bandwidth in the cloud center, and the network resource constraint of the cloud center is obtained. The execution module 203 is mainly used for maximizing the total data value transmitted to the cloud center according to all data value densities and taking the network resource constraint as the constraint condition of each edge node in the industrial internet, and performing individual optimization on the data acquisition frequency of each edge node in the industrial internet.

[0040] The above describes the examples of the industrial internet edge multi-source heterogeneous data collaborative collection method and device provided by the embodiments of the present application in detail. It can be understood that the corresponding device contains the hardware structure and / or software module corresponding to the execution of each function in order to realize the above functions. Those skilled in the art should easily realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is realized in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0041] In some embodiments, the present application also provides a computer device, which comprises a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the industrial internet edge multi-source heterogeneous data collaborative collection method described above.

[0042] In some embodiments, with reference to Figure 4 The dashed line in the figure represents that the unit or the module is optional, and the figure is a structural schematic diagram of a computer device for implementing the industrial internet edge multi-source heterogeneous data collaborative collection method according to the embodiments of the present application. The industrial internet edge multi-source heterogeneous data collaborative collection method described in the above embodiments can be realized by the computer device shown in the figure, which comprises at least one processor 301, a memory 302 and at least one communication unit 305. The computer device can be a terminal device or a server or a chip. Figure 4

[0043] ​The processor 301 can be a general processor or a special-purpose processor. For example, the processor 301 can be a central processing unit (CPU), which can be used to control a computer device, execute a software program, and process data of the software program. The computer device can further include a communication unit 305 to implement input (reception) and output (transmission) of signals.

[0044] For example, the computer device can be a chip, and the communication unit 305 can be an input and / or output circuit of the chip, or the communication unit 305 can be a communication interface of the chip, and the chip can be a component of a terminal device or a network device or other device.

[0045] For another example, the computer device can be a terminal device or a server, and the communication unit 305 can be a transceiver of the terminal device or the server, or the communication unit 305 can be a transceiver circuit of the terminal device or the server.

[0046] The computer device can include one or more memories 302, which store a program 304 that can be executed by the processor 301 to generate instructions 303, so that the processor 301 performs the method described in the above method embodiments according to the instructions 303. Optionally, the memory 302 can further store data (such as a target audit model). Optionally, the processor 301 can further read data stored in the memory 302, and the data can be stored in the same storage address as the program 304, or the data can be stored in a different storage address from the program 304.

[0047] The processor 301 and the memory 302 can be separately arranged or integrated together, for example, integrated on a system on chip (SOC) of a terminal device.

[0048] It should be understood that each step of the above method embodiments can be completed by a logic circuit in the form of hardware or instructions in the form of software in the processor 301. The processor 301 can be a CPU, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, for example, discrete gates or transistor logic devices, or discrete hardware components.

[0049] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk memory, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0050] For example, in some embodiments, the present application also provides a computer readable storage medium, having stored therein instructions or codes, which when executed on a computer, cause the computer to perform the above-mentioned method for collaborative collection of multi-source heterogeneous data of an industrial internet edge.

[0051] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those skilled in the art once they learn of the basic inventive concepts. Therefore, the appended claims are intended to encompass within their scope all such variations and modifications as are included within the scope of the application.

[0052] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A method for collaborative acquisition of multi-source heterogeneous data at the edge of an industrial internet, characterized in that, Includes the following steps: Acquire multi-source data transmission information between the cloud center and various edge nodes in the industrial internet, and monitor the device status information of the cloud center; For each edge node in the Industrial Internet, the temporal variation gradient of the information transmission volume of each data source in the edge node is calculated by using the multi-source data transmission information of the edge node. The cross-source data differences in the edge node and the business function weights of each data source are used to fuse all the temporal variation gradients into the data value density of the edge node, thereby obtaining the data value density of each edge node in the Industrial Internet. The transmission delay features between the cloud center and each edge node are extracted from the device status information. The transmission resources between the cloud center and each edge node are then constrained by the transmission delay features and the fluctuation features of the network bandwidth in the cloud center, so as to obtain the network resource constraints of the cloud center. Based on all data value densities and using the network resource constraints as constraints for each edge node in the Industrial Internet, the total data value transmitted to the cloud center is maximized, and the data acquisition frequency of each edge node in the Industrial Internet is optimized in a personalized manner.

2. The method as described in claim 1, characterized in that, Calculating the temporal variation gradient of information transmission volume from various data sources in edge nodes through multi-source data transmission information of edge nodes specifically includes: For each data source in the edge node, the information transmission volume of the data source at each sampling point is obtained from the multi-source data transmission information of the edge node. By determining the temporal variation gradient of the data source information transmission volume through all information transmission volumes, the temporal variation gradient of the information transmission volume of each data source in the edge node can be obtained.

3. The method as described in claim 1, characterized in that, Using cross-source data differences in edge nodes and the business function weights of each data source, all time-series variation gradients are fused into the data value density of the edge nodes, specifically including: Determine the cross-source data differences in edge nodes and obtain the business function weights of each data source; By merging the temporal variation gradients of each business function weight into an initial value density for the edge nodes; The initial value of the value density is adjusted based on the cross-source data differences to obtain the data value density of the edge nodes.

4. The method as described in claim 1, characterized in that, Extracting transmission delay features between the cloud center and each edge node from the device status information specifically includes: For each edge node, multiple echo request data packets between the cloud center and the edge node are obtained from the device status information; Extract the round-trip time value of each echo request data packet; The transmission delay characteristics between the cloud center and edge nodes are determined by all round-trip delay values, thereby obtaining the transmission delay characteristics between the cloud center and each edge node.

5. The method as described in claim 1, characterized in that, By applying various transmission delay characteristics and the fluctuation characteristics of network bandwidth in the cloud center, delay constraints are imposed on the transmission resources between the cloud center and each edge node. The specific network resource constraints at the cloud center include: Extract the fluctuation characteristics of network bandwidth in the cloud center; The delay constraint of transmission resources between the cloud center and each edge node is determined based on the fluctuation characteristics and various transmission delay characteristics. The network resource constraints at the cloud center are determined by all latency constraints.

6. The method as described in claim 1, characterized in that, Based on the total data value density and using the network resource constraints as constraints for each edge node in the Industrial Internet, the data acquisition frequency of each edge node in the Industrial Internet is optimized in a personalized manner to maximize the total data value transmitted to the cloud center. This includes: Initialize a node constraint model based on linear programming; Use all data value density as the value decision parameter of the node constraint model; Maximizing the total value of data transmitted to the cloud center is taken as the constraint objective of the node constraint model; The network resource constraints are used as resource constraints for each edge node in the node constraint model. The node constraint model described above is used to apply personalized constraints to the data acquisition frequency of each edge node in the Industrial Internet.

7. The method as described in claim 1, characterized in that, The edge node is an embedded gateway based on the ARM Cortex-A7 architecture.

8. A multi-source heterogeneous data collaborative acquisition device for the edge of an industrial internet, characterized in that, include: The acquisition module is used to acquire multi-source data transmission information between the cloud center and various edge nodes in the industrial internet, and to monitor the device status information of the cloud center. The processing module is used to calculate the temporal variation gradient of the information transmission volume of each data source in the edge node through the multi-source data transmission information of the edge node. It uses the cross-source data differences in the edge node and the business function weights of each data source to fuse all the temporal variation gradients into the data value density of the edge node, thereby obtaining the data value density of each edge node in the industrial internet. The processing module is also used to extract the transmission delay features between the cloud center and each edge node from the device status information, and to impose delay constraints on the transmission resources between the cloud center and each edge node by using the transmission delay features and the fluctuation features of the network bandwidth in the cloud center, so as to obtain the network resource constraints of the cloud center. The execution module is used to optimize the data acquisition frequency of each edge node in the industrial internet in a personalized manner, based on all data value densities and using the network resource constraints as constraints for each edge node in the industrial internet, to maximize the total data value transmitted to the cloud center.

9. A computer device, characterized in that, The computer device includes a memory and a processor. The memory is used to store computer programs, and the processor is used to call and run the computer programs from the memory, so that the computer device executes the multi-source heterogeneous data collaborative acquisition method for the industrial internet edge as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions or code that, when executed on a computer, cause the computer to implement the multi-source heterogeneous data collaborative acquisition method for the industrial internet edge as described in any one of claims 1 to 7.

Images