CN116629252B

CN116629252B - Remote parameter adaptive configuration method and system based on Internet of Things

Info

Publication number: CN116629252B
Application number: CN202310896195.1A
Authority: CN
Inventors: 李广辉; 钱卿
Original assignee: Wuxi Xiaojing Sharing Network Technology Co ltd
Current assignee: Wuxi Xiaojing Sharing Network Technology Co ltd
Priority date: 2023-07-21
Filing date: 2023-07-21
Publication date: 2023-09-26
Anticipated expiration: 2043-07-21
Also published as: CN116629252A

Abstract

The invention provides a remote parameter adaptive configuration method and system based on the Internet of things, and relates to the technical field of artificial intelligence. In the invention, the operation control text to be analyzed is subjected to key information mining operation of a first dimension, and a first dimension description vector is output; when the second dimension description vector is excavated, according to the second dimension description vector and the first dimension description vector, analyzing the second dimension description vector of the second dimension excavation unit; if the second dimension description vector of the last second dimension mining unit is determined, marking the second dimension description vector as a designated second dimension description vector; performing operation reliability analysis on the appointed second dimension description vector, and outputting operation reliability analysis data; and regenerating new equipment operation data to be analyzed under the condition that the reliability analysis data does not exceed the reference reliability operation data duty ratio. Based on the above, the reliability of the adaptive configuration of the remote parameter can be improved.

Description

Remote parameter adaptive configuration method and system based on Internet of things

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a remote parameter adaptive configuration method and system based on the Internet of things.

Background

The internet of things is a network which is used for connecting any article with the internet through information sensing equipment according to a stipulated protocol and carrying out information exchange and communication so as to realize intelligent identification, positioning, tracking, monitoring and management. In popular terms, the internet of things is the internet of things, and comprises two layers of meanings: firstly, the Internet of things is an extension and expansion of the Internet, and the core and the foundation of the Internet are still the Internet; secondly, the user side of the Internet of things not only comprises people, but also comprises articles, and the Internet of things realizes the exchange and communication of information among people, articles and articles.

In many applications of the internet of things, including the background server performing remote parameter adaptive configuration operation on the front-end terminal device of the internet of things, however, in the prior art, there is a problem that the reliability of the remote parameter adaptive configuration is not high.

Disclosure of Invention

Accordingly, the present invention is directed to a method and a system for adaptively configuring remote parameters based on the internet of things, so as to improve the reliability of the adaptive configuration of the remote parameters to a certain extent.

In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme:

A remote parameter adaptive configuration method based on the Internet of things comprises the following steps:

extracting operation data of equipment to be analyzed, and determining an optimized operation data analysis network for analyzing the operation reliability of the operation data of the equipment to be analyzed, wherein the optimized operation data analysis network performs network optimization operation on candidate operation data analysis networks according to typical equipment operation data and operation feasibility actual parameters corresponding to the typical equipment operation data to form the operation feasibility actual parameters, the operation data of the typical equipment are formed by performing configuration operation on the typical equipment operation data, the operation data of the equipment to be analyzed are operation control texts to be analyzed generated aiming at operation control requirements of target internet of things equipment, the operation control texts to be analyzed comprise a plurality of local operation control texts generated through an operation data generation network, the optimized operation data analysis network comprises a first dimension mining model, a second dimension mining model and an operation data analysis model, the first dimension mining model comprises a first number of first dimension mining units, the second dimension mining model comprises a second number of second dimension mining units, a difference value between the second number of the second dimension mining units and the first number of the second dimension mining units is equal to 1, and each second dimension mining model comprises two dimension mining units different from each other in the first dimension mining model;

Loading the operation control text to be analyzed to an a-th first dimension mining unit in the first number of first dimension mining units so as to carry out first dimension key information mining operation on the operation control text to be analyzed, outputting a first dimension description vector of the a-th first dimension mining unit, wherein the a-th first dimension mining unit is provided with a first dimension mining subunit and a first dimension mapping subunit, the first dimension mining subunit is used for mining depth key information of the operation control text to be analyzed, and the first dimension description vector carries out vector mapping operation on the depth key information based on the first dimension mapping subunit so as to form the operation control text;

loading the operation control text to be analyzed, loading the operation control text into the second dimension mining model, and analyzing a second dimension description vector of a b second dimension mining unit according to the second dimension description vector of the a second dimension mining unit and a first dimension description vector of the a first dimension mining unit under the condition that a second dimension description vector of an a second dimension mining unit in the second plurality of second dimension mining units is mined, wherein the a second dimension mining unit belongs to a previous second dimension mining unit of the b second dimension mining unit;

If the second dimension description vector of the b second dimension mining unit is determined to belong to the second dimension description vector of the last second dimension mining unit in the second plurality of second dimension mining units, marking the second dimension description vector of the last second dimension mining unit to be marked as a designated second dimension description vector;

performing operation reliability analysis operation on the appointed second dimension description vector by using the operation data analysis model so as to output operation reliability analysis data of the operation control text to be analyzed, wherein the operation reliability analysis data of the operation control text to be analyzed is used for reflecting the reliable operation data duty ratio of the plurality of local operation control texts included in the operation control text to be analyzed;

and under the condition that the reliable operation data duty ratio of the plurality of local operation control texts does not exceed the reference reliable operation data duty ratio, regenerating new operation data of the equipment to be analyzed through the operation data generating network, wherein the new operation data of the equipment to be analyzed are used for being issued to the target internet of things equipment to serve as operation configuration parameters of the target internet of things equipment.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the b second dimension mining unit includes a second dimension mining subunit, a focusing feature analysis subunit and a parameter mapping subunit; and loading the operation control text to be analyzed to the second dimension mining model, and under the condition of mining the second dimension description vector of the a second dimension mining unit in the second plurality of second dimension mining units, analyzing the second dimension description vector of the b second dimension mining unit according to the second dimension description vector of the a second dimension mining unit and the first dimension description vector of the a first dimension mining unit, wherein the step comprises the following steps:

loading the operation control text to be analyzed to be loaded into the second dimension mining model, and performing cascading combination operation on the second dimension description vector of the a second dimension mining unit and the first dimension description vector of the a first dimension mining unit under the condition that the second dimension description vector of the a second dimension mining unit in the second plurality of second dimension mining units is mined to form a corresponding cascading combination description vector;

Loading the cascade combination description vector to the second dimension mining subunit included in the b second dimension mining unit, and performing key information mining operation on the cascade combination description vector to output a corresponding initial smoothness description vector;

loading the initial smoothness description vector to the focusing characteristic analysis subunit, analyzing focusing parameter distribution corresponding to the initial smoothness description vector, and analyzing an optimized smoothness description vector according to the focusing parameter distribution and the initial smoothness description vector;

and loading the optimized smoothness description vector to the parameter mapping subunit, and performing vector parameter mapping operation on the optimized smoothness description vector to form a second dimension description vector of the b second dimension mining unit.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the step of loading the to-be-analyzed running control text to load into the second dimension mining model, and performing a cascading combination operation on the second dimension description vector of the a-th second dimension mining unit and the first dimension description vector of the a-th first dimension mining unit to form a corresponding cascading combination description vector under the condition that the second dimension description vector of the a-th second dimension mining unit in the second plurality of second dimension mining units is mined includes:

Loading the operation control text to be analyzed to be loaded into the second dimension mining model, and under the condition that second dimension description vectors of an a-th second dimension mining unit in the second plurality of second dimension mining units are mined, performing information decimation operation on the second dimension description vectors of the a-th second dimension mining unit to output decimated second dimension description vectors corresponding to the a-th second dimension mining unit, wherein the vector size of the decimated second dimension description vectors is the same as the vector size of the first dimension description vectors of the a-th first dimension mining unit;

and performing cascading combination operation on the decimated second dimension description vector and the first dimension description vector of the a first dimension mining unit to form a corresponding cascading combination description vector.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the second dimension mining subunit includes a first-stage filtering structure, a first-stage mapping structure, and a second-stage filtering structure;

the step of loading the cascade combination description vector to load to the second dimension mining subunit included in the b second dimension mining unit, and performing key information mining operation on the cascade combination description vector to output a corresponding initial smoothness description vector includes:

Loading the cascade combination description vector to the first-stage filtering structure, and performing filtering operation on the cascade combination description vector to output a corresponding first-stage filtering description vector;

loading the first-stage filtering description vector to the first-stage mapping structure, and performing mapping operation on the first-stage filtering description vector to form a corresponding second-stage filtering description vector;

and loading the second-stage filtering description vector to load the second-stage filtering description vector into the second-stage filtering structure, and performing filtering operation on the second-stage filtering description vector to output a corresponding initial smoothness description vector.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the focusing characteristic analysis subunit includes a third level filtering structure, a second level mapping structure, a fourth level filtering structure, and a third level mapping structure; the third-stage filtering structure is the same as the network architecture of the fourth-stage filtering structure;

the step of loading the initial smoothness description vector to the focus feature analysis subunit, analyzing a focus parameter distribution corresponding to the initial smoothness description vector, and analyzing an optimized smoothness description vector according to the focus parameter distribution and the initial smoothness description vector, includes:

Loading the initial through description vectors to load into the third stage filtering structure, grouping the initial through description vectors to form a third number of local through description vectors, and performing a filtering operation on each of the third number of local through description vectors to form a third number of third stage filtering description vectors;

loading the third number of third-stage filter description vectors to the second-stage mapping structure, and mapping each third-stage filter description vector of the third number of third-stage filter description vectors to form a corresponding third number of fourth-stage filter description vectors;

loading the third number of fourth-stage filter description vectors to the fourth-stage filter structure, and performing vector combination operation on the third number of fourth-stage filter description vectors to form corresponding combined filter description vectors;

loading the combined filter description vector to the third-level mapping structure, and performing mapping operation on the combined filter description vector to form a focusing parameter distribution corresponding to the initial smoothness description vector, wherein each focusing parameter in the focusing parameter distribution is greater than or equal to 0;

And carrying out fusion operation on the focusing parameter distribution and the initial smoothness description vector to form a corresponding optimized smoothness description vector.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the step of loading the optimization smoothness description vector to the parameter mapping subunit, and performing a mapping operation of vector parameters on the optimization smoothness description vector to form a second dimension description vector of the b second dimension mining unit includes:

loading the optimized general description vector to the parameter mapping subunit, and performing filtering operation on the optimized general description vector to form a corresponding candidate filtering description vector;

and carrying out aggregation operation on the candidate filtering description vector and the optimization smoothness description vector, and carrying out parameter mapping operation on the description vector formed by the aggregation operation to form a second dimension description vector of the b second dimension mining unit.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the optimized operation data analysis network further includes a front-end key information mining model corresponding to the second dimension mining model;

The front-end key information mining model is used for mining key information description vectors of the operation control text to be analyzed; and under the condition that the a-th second dimension mining unit is a first second dimension mining unit in the second plurality of second dimension mining units, the first second dimension mining unit is used for carrying out second-dimension key information mining operation on the key information description vector when the key information description vector is loaded into the key information description vector mined by the front-end key information mining model so as to form the second dimension description vector of the first second dimension mining unit.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the operation data analysis model includes an information extraction unit and an operation data analysis unit;

the step of using the operation data analysis model to perform operation reliability analysis operation on the specified second dimension description vector to output operation reliability analysis data of the operation control text to be analyzed includes:

loading the appointed second-dimension description vector to the information extraction unit, and carrying out information extraction operation on the appointed second-dimension description vector to form an information extraction description vector;

And loading the information extraction description vector to the operation data analysis unit, and performing operation reliability analysis operation on the information extraction description vector to output operation reliability analysis data of the operation control text to be analyzed.

In some preferred embodiments, in the remote parameter adaptive configuration method based on the internet of things, the remote parameter adaptive configuration method further includes:

extracting typical equipment operation data for optimizing a candidate operation data analysis network and actual operation feasibility parameters of the typical equipment operation data;

loading the typical equipment operation data to be loaded into an a-th candidate first dimension mining unit in a first number of candidate first dimension mining units, performing first dimension key information mining operation on the typical equipment operation data, and outputting a first dimension description vector of the a-th candidate first dimension mining unit;

loading the typical equipment operation data to load into a candidate second dimension mining model, and analyzing a second dimension description vector of a b candidate second dimension mining unit according to the second dimension description vector of the a candidate second dimension mining unit and the first dimension description vector of the a candidate first dimension mining unit under the condition of mining the second dimension description vector of the a candidate second dimension mining unit in the second plurality of candidate second dimension mining units;

If the second dimension description vector of the b candidate second dimension mining unit is determined to belong to the second dimension description vector of the last candidate second dimension mining unit in the second number of candidate second dimension mining units, marking the second dimension description vector of the last candidate second dimension mining unit to be a typical second dimension description vector;

performing operation reliability analysis operation on the typical second dimension description vector by using a candidate operation data analysis model so as to output operation reliability typical data of the typical equipment operation data;

and carrying out network optimization operation on the candidate operation data analysis network according to the operation reliability typical data and the operation feasibility actual parameters so as to form an optimized operation data analysis network.

The embodiment of the invention also provides a remote parameter adaptive configuration system based on the Internet of things, which comprises a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program so as to realize the remote parameter adaptive configuration method based on the Internet of things.

According to the remote parameter adaptive configuration method and system based on the Internet of things, which are provided by the embodiment of the invention, the operation control text to be analyzed can be subjected to key information mining operation of a first dimension, and a first dimension description vector is output; when the second dimension description vector is excavated, according to the second dimension description vector and the first dimension description vector, analyzing the second dimension description vector of the second dimension excavation unit; if the second dimension description vector of the last second dimension mining unit is determined, marking the second dimension description vector as a designated second dimension description vector; performing operation reliability analysis on the appointed second dimension description vector, and outputting operation reliability analysis data; and regenerating new equipment operation data to be analyzed under the condition that the reliability analysis data does not exceed the reference reliability operation data duty ratio. Based on the foregoing, the optimizing operation data analysis network is utilized to perform two-dimensional key information mining on the operation data of the device to be analyzed, and the mining results of the two dimensions are fused, so that the obtained appointed second dimension description vector for performing the operation reliability analysis operation has better characterization capability, the reliability of the operation reliability analysis operation is ensured, the reliability of remote parameter adaptive configuration can be improved to a certain extent, and the defects of the prior art are overcome.

In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

Fig. 1 is a block diagram of a remote parameter adaptive configuration system based on the internet of things according to an embodiment of the present invention.

Fig. 2 is a schematic flow chart of steps included in the remote parameter adaptive configuration method based on the internet of things according to the embodiment of the present invention.

Fig. 3 is a schematic diagram of each module included in the remote parameter adaptive configuration device based on the internet of things according to the embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

As shown in fig. 1, the embodiment of the invention provides a remote parameter adaptive configuration system based on the internet of things. The remote parameter adaptive configuration system based on the Internet of things can comprise a memory and a processor.

In detail, the memory and the processor are electrically connected directly or indirectly to realize transmission or interaction of data. For example, electrical connection may be made to each other via one or more communication buses or signal lines. The memory may store at least one software functional module (computer program) that may exist in the form of software or firmware. The processor may be configured to execute the executable computer program stored in the memory, so as to implement the remote parameter adaptive configuration method based on the internet of things provided by the embodiment of the present invention.

It should be appreciated that in some possible embodiments, the Memory may be, but is not limited to, random access Memory (Random Access Memory, RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-Only Memory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), and the like.

It should be appreciated that in some possible embodiments, the processor may be a general purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), a System on Chip (SoC), etc.; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

It should be appreciated that in some possible embodiments, the remote parameter adaptive configuration system based on the internet of things may be a server with data processing capability.

With reference to fig. 2, the embodiment of the invention further provides a remote parameter adaptive configuration method based on the internet of things, which can be applied to the remote parameter adaptive configuration system based on the internet of things. The method steps defined by the flow related to the remote parameter adaptive configuration method based on the internet of things can be realized by the remote parameter adaptive configuration system based on the internet of things.

The specific flow shown in fig. 2 will be described in detail.

Step S110, extracting operation data of equipment to be analyzed, and determining an optimized operation data analysis network for analyzing the operation reliability of the operation data of the equipment to be analyzed.

In the embodiment of the invention, the remote parameter adaptive configuration system based on the Internet of things can extract the operation data of the equipment to be analyzed and determine the optimized operation data analysis network for analyzing the operation reliability of the operation data of the equipment to be analyzed. The optimized operation data analysis network performs network optimization operation on candidate operation data analysis networks according to typical equipment operation data and operation feasibility actual parameters corresponding to the typical equipment operation data to form (the network optimization operation can refer to related description later), the operation feasibility actual parameters are formed by performing configuration operation on the typical equipment operation data (such as manual configuration, etc.), the equipment operation data to be analyzed is an operation control text to be analyzed generated according to operation control requirements of target internet of things equipment, the operation control text to be analyzed comprises a plurality of local operation control texts generated through an operation data generation network (for example, each local operation control text is used for controlling one function of the target internet of things equipment, such as washing time length, washing mode, etc.), the optimized operation data analysis network comprises a first dimension mining model, a second dimension mining model and an operation data analysis model, the first dimension mining model comprises a first number of first dimension mining units, the second dimension mining model comprises a second number of second dimension mining units, the second dimension mining units is different from the first dimension mining units, the second dimension mining units are different from the first dimension mining units, the second dimension mining models are different from the first dimension mining units, the first dimension mining models are different from the first dimension mining units, the second dimension mining models are different from each other, except that the parameters are different.

Step S120, loading the operation control text to be analyzed, so as to load the operation control text to be analyzed into an a-th first dimension mining unit in the first number of first dimension mining units, so as to perform key information mining operation of the first dimension on the operation control text to be analyzed, and outputting a first dimension description vector of the a-th first dimension mining unit.

In the embodiment of the invention, the remote parameter adaptive configuration system based on the internet of things can load the operation control text to be analyzed to be loaded into an a-th first dimension mining unit in the first number of first dimension mining units so as to carry out the key information mining operation of the first dimension on the operation control text to be analyzed and output the first dimension description vector of the a-th first dimension mining unit. The a-th first dimension mining unit is provided with a first dimension mining subunit and a first dimension mapping subunit, the first dimension mining subunit (which may include a convolution kernel or a filtering operator) is used for mining depth key information of the running control text to be analyzed, and the first dimension description vector is formed by performing vector mapping operation on the depth key information based on the first dimension mapping subunit (which may include an excitation mapping function). The a-th first dimension mining unit may be any one of the first number of first dimension mining units, so that a first dimension description vector of each first dimension mining unit may be obtained.

Step S130, loading the operation control text to be analyzed to load the operation control text into the second dimension mining model, and under the condition that the second dimension description vector of the a second dimension mining unit in the second plurality of second dimension mining units is mined, analyzing the second dimension description vector of the b second dimension mining unit according to the second dimension description vector of the a second dimension mining unit and the first dimension description vector of the a first dimension mining unit.

In the embodiment of the invention, the remote parameter adaptive configuration system based on the internet of things can load the operation control text to be analyzed to load the operation control text into the second dimension mining model, and analyze the second dimension description vector of the b second dimension mining unit according to the second dimension description vector of the a second dimension mining unit and the first dimension description vector of the a first dimension mining unit under the condition that the second dimension description vector of the a second dimension mining unit in the second plurality of second dimension mining units is mined. The a-th second-dimension mining unit belongs to a second-dimension mining unit before the b-th second-dimension mining unit.

Step S140, if it is determined that the second dimension description vector of the b-th second dimension mining unit belongs to the second dimension description vector of the last second dimension mining unit in the second number of second dimension mining units, marking the second dimension description vector of the last second dimension mining unit to be a specified second dimension description vector.

In the embodiment of the present invention, the remote parameter adaptive configuration system based on the internet of things may mark the second dimension description vector of the last second dimension mining unit when it is determined that the second dimension description vector of the b-th second dimension mining unit belongs to the second dimension description vector of the last second dimension mining unit in the second plurality of second dimension mining units, so as to mark the second dimension description vector as the designated second dimension description vector.

And step S150, utilizing the operation data analysis model to conduct operation reliability analysis operation on the specified second dimension description vector so as to output operation reliability analysis data of the operation control text to be analyzed.

In the embodiment of the invention, the remote parameter adaptive configuration system based on the internet of things can utilize the operation data analysis model to perform operation reliability analysis operation on the appointed second dimension description vector so as to output operation reliability analysis data of the operation control text to be analyzed. The operation reliability analysis data of the operation control text to be analyzed is used for reflecting the reliable operation data duty ratio of the plurality of local operation control texts included in the operation control text to be analyzed. The reliable operation data duty ratio refers to a duty ratio of the number of the partial operation control texts capable of being reliably operated in the plurality of partial operation control texts.

And step S160, regenerating new equipment operation data to be analyzed through the operation data generating network under the condition that the reliable operation data duty ratio of the plurality of local operation control texts does not exceed the reference reliable operation data duty ratio.

In the embodiment of the invention, the remote parameter adaptive configuration system based on the internet of things can regenerate new equipment operation data to be analyzed through the operation data generation network under the condition that the reliable operation data duty ratio of the plurality of local operation control texts does not exceed the reference reliable operation data duty ratio. And the new equipment operation data to be analyzed is used for being issued to the target internet of things equipment to serve as operation configuration parameters of the target internet of things equipment. The specific values of the reference reliable operation data duty cycle are not limited, such as values of 0.8, 0.85, etc. In addition, the operation data generating network may be a neural network formed through a network optimization operation, based on which the step of regenerating new operation data of the device to be analyzed through the operation data generating network may include: and carrying out network optimization operation again on the operation data generating network based on the typical equipment operation data and the typical operation control demand text which are not used in the previous network optimization process by the operation data generating network, so that the mapping relationship between the equipment operation data and the operation control demand text which are learned by the operation data generating network is more reliable, the operation control demand (text) of the target Internet of things equipment is analyzed and predicted based on the more reliable operation data generating network, and the new equipment operation data to be analyzed is obtained.

Based on the foregoing, the optimizing operation data analysis network is utilized to perform two-dimensional key information mining on the operation data of the device to be analyzed, and the mining results of the two dimensions are fused, so that the obtained appointed second dimension description vector for performing the operation reliability analysis operation has better characterization capability, the reliability of the operation reliability analysis operation is ensured, the reliability of remote parameter adaptive configuration can be improved to a certain extent, and the defects of the prior art are overcome.

It should be understood that, in some possible embodiments, the b-th second dimension mining unit includes a second dimension mining subunit, a focusing feature analysis subunit, and a parameter mapping subunit, based on which, in the above step S130, the following specific embodiments may be further included:

loading the operation control text to be analyzed to be loaded into the second dimension mining model, and under the condition that second dimension description vectors of a second dimension mining unit in the second plurality of second dimension mining units are mined, performing cascading combination operation on the second dimension description vectors of the a second dimension mining unit and the first dimension description vectors of the a first dimension mining unit to form corresponding cascading combination description vectors, namely, performing key information mining operation on the operation control text to be analyzed by using the second plurality of second dimension mining units to form second plurality of second dimension description vectors, and then performing cascading combination operation on the second dimension description vectors and the first dimension description vectors of the corresponding first dimension mining units to obtain corresponding cascading combination description vectors, such as { first dimension description vectors, first dimension description vectors }, for example, the second dimension description vectors of the 1 st second dimension mining unit and the first dimension description vectors of the 1 st dimension mining unit can be subjected to cascading combination operation to form cascading combination description vectors;

Loading the cascade combination description vector to the second dimension mining subunit included in the b-th second dimension mining unit (a second dimension mining unit after the a-th second dimension mining unit), and performing key information mining operation on the cascade combination description vector to output a corresponding initial smoothness description vector, wherein the initial smoothness description vector has information representing the semantic smoothness of the operation control text to be analyzed;

loading the initial smoothness description vector to the focusing characteristic analysis subunit (belonging to the b second dimension mining unit), analyzing focusing parameter distribution corresponding to the initial smoothness description vector, and analyzing an optimized smoothness description vector according to the focusing parameter distribution and the initial smoothness description vector;

and loading the optimized smoothness description vector to the parameter mapping subunit, performing vector parameter mapping operation on the optimized smoothness description vector, and mapping the vector parameter to a corresponding interval to form a second dimension description vector of the b second dimension mining unit.

It should be understood that, in some possible embodiments, the step of loading the running control text to be analyzed to load into the second dimension mining model and performing a cascading operation on the second dimension description vector of the a-th second dimension mining unit and the first dimension description vector of the a-th first dimension mining unit to form a corresponding cascading combination description vector in the case of mining the second dimension description vector of the a-th second dimension mining unit in the second number of second dimension mining units further includes the following specific embodiments:

loading the operation control text to be analyzed, loading the operation control text into the second dimension mining model, under the condition that a second dimension description vector of an a second dimension mining unit in the second number of second dimension mining units is mined, performing information extraction operation on the second dimension description vector of the a second dimension mining unit, so as to output an extracted second dimension description vector corresponding to the a second dimension mining unit, wherein the vector size of the extracted second dimension description vector is the same as the vector size of a first dimension description vector of the a first dimension mining unit, for example, when performing information extraction operation, a window with a specified size can be determined, then, for each vector parameter in the second dimension description vector, taking the vector parameter as a center, determining a region by taking the window as a center, taking the average value or the maximum value of each parameter in the region as a representative parameter of the vector parameter, so as to form a corresponding intermediate vector, and forming a corresponding dimension, and performing sampling operation on the first dimension description vector, such as a first dimension description vector, a corresponding to the first dimension description vector, or a second dimension description vector, or a corresponding to the first dimension description vector, and performing sampling operation on the first dimension description vector, and the like;

And performing cascade combination operation on the decimated second dimension description vector and the first dimension description vector of the a first dimension mining unit to form a corresponding cascade combination description vector, wherein the cascade combination description vector is { decimated second dimension description vector, first dimension description vector }.

It should be understood that, in some possible embodiments, the second dimension mining subunit includes a first level filtering structure, a first level mapping structure, and a second level filtering structure, based on which the step of loading the cascaded combination description vector to load to the second dimension mining subunit included in the b-th second dimension mining unit, and performing a key information mining operation on the cascaded combination description vector to output a corresponding initial generic description vector may further include the following specific embodiments:

loading the cascade combination description vector to the first-stage filtering structure, and performing filtering operation on the cascade combination description vector to output a corresponding first-stage filtering description vector, wherein parameters of a filtering operator in the first-stage filtering structure can be formed in a network optimization process;

Loading the first-stage filter description vector to the first-stage mapping structure, and performing mapping operation on the first-stage filter description vector to form a corresponding second-stage filter description vector, for example, the first-stage mapping structure may include an excitation mapping function (may also be referred to as an excitation function) to process the first-stage filter description vector;

and loading the second-stage filtering description vector to the second-stage filtering structure, performing filtering operation on the second-stage filtering description vector, and outputting a corresponding initial smoothness description vector, wherein parameters of a filtering operator in the second-stage filtering structure can be formed in a network optimization process, namely, filtering operation is performed based on the filtering operator.

It should be appreciated that in some possible implementations, the focus feature analysis subunit includes a third level filtering structure, a second level mapping structure, a fourth level filtering structure, and a third level mapping structure; the third stage filtering structure and the fourth stage filtering structure have the same network architecture, based on which the initial smoothness description vector is loaded to be loaded into the focusing feature analysis subunit, to analyze a focusing parameter distribution corresponding to the initial smoothness description vector, and the step of analyzing an optimized smoothness description vector according to the focusing parameter distribution and the initial smoothness description vector may further include the following specific embodiments:

Loading the initial through description vectors for loading into the third stage filtering structure, grouping the initial through description vectors to form a third number of local through description vectors (the third number of local through description vectors may be combined to form the initial through description vector, specific grouping parameters may be formed during network optimization), and performing a filtering operation on each of the third number of local through description vectors to form a third number of third stage filtering description vectors, that is, one local through description vector corresponds to one third stage filtering description vector;

loading the third number of third-stage filter description vectors for loading into the second-stage mapping structure, mapping each of the third number of third-stage filter description vectors, e.g., by including an excitation mapping function, to form a corresponding third number of fourth-stage filter description vectors, that is, one fourth-stage filter description vector corresponding to each third-stage filter description vector;

Loading the third number of fourth-stage filter description vectors to the fourth-stage filter structure, and performing vector combination operation on the third number of fourth-stage filter description vectors to form corresponding combined filter description vectors, wherein the vector combination operation can be reciprocal to a mode of grouping the initial smoothness description vectors;

loading the combined filter description vector to the third-level mapping structure, and performing mapping operation on the combined filter description vector, for example, by using an included excitation mapping function, so as to form a focusing parameter distribution corresponding to the initial smoothness description vector, wherein each focusing parameter in the focusing parameter distribution is greater than or equal to 0;

and performing a fusion operation on the focusing parameter distribution and the initial smoothness description vector to form a corresponding optimized smoothness description vector, for example, performing a multiplication operation on the focusing parameter distribution and the initial smoothness description vector to form an optimized smoothness description vector.

It should be appreciated that, in some possible embodiments, the step of loading the optimization smoothness description vector to load into the parameter mapping subunit, and performing the mapping operation of vector parameters on the optimization smoothness description vector to form the second dimension description vector of the b second dimension mining unit may further include the following specific embodiments:

Loading the optimized smoothness description vector to the parameter mapping subunit, and performing filtering operation on the optimized smoothness description vector to form a corresponding candidate filtering description vector, that is, the parameter mapping subunit comprises a filtering operator to perform filtering operation on the optimized smoothness description vector;

the candidate filter description vector and the optimization smoothness description vector are subjected to aggregation operation, and the description vector formed by the aggregation operation is subjected to parameter mapping operation to form a second dimension description vector of the b second dimension mining unit, for example, the candidate filter description vector and the optimization smoothness description vector may be subjected to superposition operation, and then, a result of the superposition operation may be subjected to parameter mapping operation to form a second dimension description vector of the b second dimension mining unit, and in particular, parameters may be mapped to intervals 0 to 1.

It should be appreciated that in some possible embodiments, the optimized operation data analysis network may further include a front-end key information mining model corresponding to the second dimension mining model. Based on the above, the front-end key information mining model is used for mining the key information description vector of the to-be-analyzed operation control text, namely mining the key information of the to-be-analyzed operation control text, for example, the front-end key information mining model can be a coding network.

And under the condition that the a-th second dimension mining unit is a first second dimension mining unit in the second plurality of second dimension mining units, the first second dimension mining unit is used for carrying out second-dimension key information mining operation on the key information description vector when the key information description vector is loaded into the key information description vector mined by the front-end key information mining model so as to form the second dimension description vector of the first second dimension mining unit.

It should be understood, that in some possible embodiments, the step of mining the key information description vector of the to-be-analyzed running control text by using the front-end key information mining model may further include the following specific embodiments:

determining each text word in the operation control text to be analyzed, and then performing embedding operation on each text word to obtain an embedded word vector corresponding to each text word;

sequencing the embedded word vectors corresponding to each text word according to the sequence from front to back in the operation control text to be analyzed so as to form a first embedded word vector sequence;

Sequencing the embedded word vectors corresponding to each text word according to the sequence from back to front in the operation control text to be analyzed so as to form a second embedded word vector sequence;

for a first embedded word vector in the first embedded word vector sequence, directly taking the embedded word vector as a first target word vector corresponding to the embedded word vector, and for each other embedded word vector except for the first embedded word vector in the first embedded word vector sequence, performing association analysis operation on the other embedded word vectors based on the first target word vector corresponding to the previous embedded word vector of the other embedded word vectors so as to output the first target word vector corresponding to the other embedded word vectors;

for a first embedded word vector in the second embedded word vector sequence, directly taking the embedded word vector as a second target word vector corresponding to the embedded word vector, and for each other embedded word vector except the first embedded word vector in the second embedded word vector sequence, performing association analysis operation on the other embedded word vectors based on the second target word vector corresponding to the previous embedded word vector of the other embedded word vectors so as to output the second target word vector corresponding to the other embedded word vectors;

For each embedded word vector, performing fusion operation on the embedded word vector, a first target word vector corresponding to the embedded word vector and a second target word vector corresponding to the embedded word vector, for example, performing weighted summation calculation to form a corresponding fused target word vector, where, illustratively, the weighting coefficient of the embedded word vector is greater than the weighting coefficient of the first target word vector corresponding to the embedded word vector, and the weighting coefficient of the first target word vector corresponding to the embedded word vector is greater than the weighting coefficient of the second target word vector corresponding to the embedded word vector;

and carrying out cascading combination operation on each embedded word vector to form a key information description vector of the operation control text to be analyzed.

It should be understood, that in some possible embodiments, the step of performing, based on the first target word vector corresponding to the previous embedded word vector of the other embedded word vector, performing a correlation analysis operation on the other embedded word vector to output the first target word vector corresponding to the other embedded word vector may further include the following specific embodiments:

Weighting (e.g., multiplying) the other embedded term vectors based on a first parameter distribution (the first parameter distribution may be a matrix of a plurality of parameters, which may be formed during network optimization of the respective neural network) to form corresponding first weighting vectors;

extracting a first target word vector corresponding to a previous embedded word vector of the other embedded word vectors, and performing a weighting operation on the first target word vector based on a second parameter distribution (the second parameter distribution may be a matrix formed by a plurality of parameters and may be formed in a network optimization process of a corresponding neural network) to form a corresponding second weighting vector, and performing a weighting operation on the first target word vector based on a third parameter distribution (the third parameter distribution may be a matrix formed by a plurality of parameters and may be formed in a network optimization process of a corresponding neural network) to form a corresponding third weighting vector;

performing a number product calculation operation on the transpose vectors of the first and second weighting vectors to output a corresponding target number product, and performing an adjustment operation on the number product based on the vector dimension of the first weighting vector to form a corresponding adjusted number product, e.g., a ratio of the target number product to the number of vector dimensions may be calculated;

And multiplying the adjustment quantity product and the third weighted vector to output a first target word vector corresponding to the other embedded word vectors.

It should be appreciated that in some possible embodiments, the operation data analysis model may include an information extraction unit and an operation data analysis unit, based on which the above step S150 may further include the following specific embodiments:

loading the appointed second-dimension description vector to the information extraction unit, and performing information extraction operation on the appointed second-dimension description vector to form an information extraction description vector, wherein the information extraction operation can refer to the related description;

Wherein, it should be understood that, in some possible embodiments, the operation data analysis unit may include a linear integration subunit and an operation data analysis subunit, based on which the steps of loading the information extraction description vector to load into the operation data analysis unit, and performing an operation reliability analysis operation on the information extraction description vector to output operation reliability analysis data of the operation control text to be analyzed may further include the following specific embodiments:

Performing linear integration operation on the information extraction description vector by using the linear integration subunit (such as a multi-layer Perceptron, MLP) to form a corresponding linear integration description vector;

the linear integration description vector is loaded to be loaded into the operation data analysis subunit, an operation data analysis result corresponding to the linear integration description vector is analyzed, the operation data analysis result corresponding to the linear integration description vector is marked to be the operation reliability analysis data of the operation control text to be analyzed, and the operation data analysis subunit can comprise an operation data analysis function, such as a softmax function and the like, so as to calculate the operation data analysis result.

It should be understood that, in some possible embodiments, the remote parameter adaptive configuration method based on the internet of things may further include the following specific embodiments:

loading the typical equipment operation data to load into an a-th candidate first dimension mining unit in a first number of candidate first dimension mining units, performing first dimension key information mining operation on the typical equipment operation data, and outputting a first dimension description vector of the a-th candidate first dimension mining unit, as described in the previous related description, namely, the step S120;

Loading the typical equipment operation data to load into a candidate second dimension mining model, and under the condition of mining the second dimension description vector of an a-th candidate second dimension mining unit in the second number of candidate second dimension mining units, analyzing the second dimension description vector of an b-th candidate second dimension mining unit according to the second dimension description vector of the a-th candidate second dimension mining unit and the first dimension description vector of the a-th candidate first dimension mining unit, wherein the previous related description is the step S130;

if it is determined that the second dimension description vector of the b-th candidate second dimension mining unit belongs to the second dimension description vector of the last candidate second dimension mining unit in the second number of candidate second dimension mining units, marking the second dimension description vector of the last candidate second dimension mining unit to be a typical second dimension description vector, as described in the foregoing description, i.e. step S140;

performing operation reliability analysis operation on the representative second dimension description vector by using a candidate operation data analysis model to output operation reliability representative data of the representative equipment operation data, as described in the foregoing related description, i.e. the foregoing step S150;

And performing network optimization operation on the candidate operation data analysis network according to the operation reliability typical data and the operation feasibility actual parameter to form an optimized operation data analysis network, for example, determining a corresponding network optimization index based on distinguishing information between the operation reliability typical data and the operation feasibility actual parameter, and performing parameter optimization adjustment operation on the candidate operation data analysis network along the direction of reducing the network optimization index, wherein the network optimization index and the distinguishing information can have a positive correlation corresponding relation.

In an application example, the target internet of things device may be a laundry detergent dispenser, and the laundry detergent dispenser may also communicate with an internet of things cloud server (e.g. the remote parameter adaptive configuration system based on internet of things), feedback payment conditions and control laundry detergent output, for example, the laundry detergent dispenser is connected with a washing machine for use, and is mainly used for storing laundry detergent, and outputting a certain amount of laundry detergent (for a laundry payment scene such as school) according to the payment conditions.

With reference to fig. 3, the embodiment of the invention further provides a remote parameter adaptive configuration device based on the internet of things, which can be applied to the remote parameter adaptive configuration system based on the internet of things. The remote parameter adaptive configuration device based on the internet of things may include:

The system comprises a data determining module, a data analyzing module and a data analyzing module, wherein the data determining module is used for extracting operation data of equipment to be analyzed and determining an optimized operation data analysis network for analyzing the operation reliability of the operation data of the equipment to be analyzed, the optimized operation data analysis network is formed by carrying out network optimization operation on candidate operation data analysis networks according to typical equipment operation data and operation feasibility actual parameters corresponding to the typical equipment operation data, the operation feasibility actual parameters are formed by carrying out configuration operation on the typical equipment operation data, the operation data of the equipment to be analyzed are operation control texts to be analyzed, which are generated according to operation control requirements of target internet of things equipment, the operation control texts to be analyzed comprise a plurality of local operation control texts generated through operation data generating networks, the optimized operation data analysis network comprises a first dimension mining model, a second dimension mining model and an operation data analysis model, the first dimension mining model comprises a first number of first dimension mining units, the second dimension mining model comprises a second dimension mining units, a difference value between the second number of the second dimension mining units and the first dimension mining units is equal to 1, and the difference between the second dimension mining models in the first dimension mining model and the second dimension mining model comprises two dimension mining units different from each other;

The first dimension mining module is used for loading the operation control text to be analyzed, loading the operation control text to be analyzed into an a-th first dimension mining unit in the first number of first dimension mining units, carrying out first dimension key information mining operation on the operation control text to be analyzed, outputting a first dimension description vector of the a-th first dimension mining unit, wherein the a-th first dimension mining unit is provided with a first dimension mining subunit and a first dimension mapping subunit, the first dimension mining subunit is used for mining depth key information of the operation control text to be analyzed, and carrying out vector mapping operation on the depth key information based on the first dimension mapping subunit to form the depth key information;

the second dimension mining module is used for loading the operation control text to be analyzed, loading the operation control text into the second dimension mining model, and analyzing a second dimension description vector of a b second dimension mining unit according to the second dimension description vector of an a second dimension mining unit and a first dimension description vector of the a first dimension mining unit under the condition that a second dimension description vector of an a second dimension mining unit in the second plurality of second dimension mining units is mined, wherein the a second dimension mining unit belongs to a previous second dimension mining unit of the b second dimension mining unit;

The appointed vector determining module is used for marking the second dimension description vector of the last second dimension mining unit in the second number of second dimension mining units to be marked as an appointed second dimension description vector if the second dimension description vector of the b second dimension mining unit is determined to belong to the second dimension description vector of the last second dimension mining unit in the second number of second dimension mining units;

the operation reliability analysis module is used for carrying out operation reliability analysis operation on the appointed second dimension description vector by utilizing the operation data analysis model so as to output operation reliability analysis data of the operation control text to be analyzed, wherein the operation reliability analysis data of the operation control text to be analyzed is used for reflecting the reliable operation data duty ratio of the plurality of local operation control texts included in the operation control text to be analyzed;

the operation data generation module is used for regenerating new operation data of the equipment to be analyzed through the operation data generation network under the condition that the reliable operation data duty ratio of the plurality of local operation control texts does not exceed the reference reliable operation data duty ratio, and the new operation data of the equipment to be analyzed is used for being issued to the target internet of things equipment to serve as operation configuration parameters of the target internet of things equipment.

In summary, the remote parameter adaptive configuration method and system based on the internet of things provided by the invention can perform the key information mining operation of the first dimension on the operation control text to be analyzed, and output the first dimension description vector; when the second dimension description vector is excavated, according to the second dimension description vector and the first dimension description vector, analyzing the second dimension description vector of the second dimension excavation unit; if the second dimension description vector of the last second dimension mining unit is determined, marking the second dimension description vector as a designated second dimension description vector; performing operation reliability analysis on the appointed second dimension description vector, and outputting operation reliability analysis data; and regenerating new equipment operation data to be analyzed under the condition that the reliability analysis data does not exceed the reference reliability operation data duty ratio. Based on the foregoing, the optimizing operation data analysis network is utilized to perform two-dimensional key information mining on the operation data of the device to be analyzed, and the mining results of the two dimensions are fused, so that the obtained appointed second dimension description vector for performing the operation reliability analysis operation has better characterization capability, the reliability of the operation reliability analysis operation is ensured, the reliability of remote parameter adaptive configuration can be improved to a certain extent, and the defects of the prior art are overcome.

The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. The remote parameter adaptive configuration method based on the Internet of things is characterized by comprising the following steps of:

extracting operation data of equipment to be analyzed, and determining an optimized operation data analysis network for analyzing the operation reliability of the operation data of the equipment to be analyzed, wherein the optimized operation data analysis network performs network optimization operation on candidate operation data analysis networks according to typical equipment operation data and operation feasibility actual parameters corresponding to the typical equipment operation data to form the operation feasibility actual parameters, the operation feasibility actual parameters are formed by performing configuration operation on the typical equipment operation data, the operation data of the equipment to be analyzed are operation control texts to be analyzed generated aiming at operation control requirements of target internet of things equipment, the operation control texts to be analyzed comprise a plurality of local operation control texts generated through an operation data generation network, the optimized operation data analysis network comprises a first dimension mining model, a second dimension mining model and an operation data analysis model, the first dimension mining model comprises a first number of first dimension mining units, and the second dimension mining model comprises a second number of second dimension mining units;

loading the operation control text to be analyzed to load the operation control text into the second dimension mining model, and analyzing a second dimension description vector of a b second dimension mining unit according to the second dimension description vector of the a second dimension mining unit and a first dimension description vector of the a first dimension mining unit under the condition that a second dimension description vector of an a second dimension mining unit in the second plurality of second dimension mining units is mined, wherein the a second dimension mining unit belongs to a previous second dimension mining unit of the b second dimension mining unit;

2. The method for adaptively configuring remote parameters based on the internet of things according to claim 1, wherein the b second dimension mining unit comprises a second dimension mining subunit, a focusing characteristic analysis subunit and a parameter mapping subunit; and loading the operation control text to be analyzed to the second dimension mining model, and under the condition of mining the second dimension description vector of the a second dimension mining unit in the second plurality of second dimension mining units, analyzing the second dimension description vector of the b second dimension mining unit according to the second dimension description vector of the a second dimension mining unit and the first dimension description vector of the a first dimension mining unit, wherein the step comprises the following steps:

3. The method for adaptively configuring remote parameters based on the internet of things according to claim 2, wherein the step of loading the to-be-analyzed running control text to load into the second dimension mining model and performing a cascading combination operation on the second dimension description vector of the a-th second dimension mining unit and the first dimension description vector of the a-th first dimension mining unit to form a corresponding cascading combination description vector in the case of mining the second dimension description vector of the a-th second dimension mining unit in the second plurality of second dimension mining units comprises:

4. The internet of things-based remote parameter adaptive configuration method of claim 2, wherein the second dimension mining subunit comprises a first-stage filtering structure, a first-stage mapping structure and a second-stage filtering structure;

5. The method for adaptively configuring remote parameters based on the internet of things according to claim 2, wherein the focusing characteristic analysis subunit comprises a third-level filtering structure, a second-level mapping structure, a fourth-level filtering structure and a third-level mapping structure; the third-stage filtering structure is the same as the network architecture of the fourth-stage filtering structure;

6. The method for adaptively configuring remote parameters based on the internet of things according to claim 2, wherein the step of loading the optimized generic description vector to be loaded into the parameter mapping subunit and performing a mapping operation of vector parameters on the optimized generic description vector to form a second dimension description vector of the b-th second dimension mining unit comprises:

7. The method for adaptively configuring remote parameters based on the internet of things according to claim 1, wherein the optimized operation data analysis network further comprises a front-end key information mining model corresponding to the second dimension mining model;

8. The remote parameter adaptive configuration method based on the internet of things according to claim 1, wherein the operation data analysis model comprises an information extraction unit and an operation data analysis unit;

9. The method for adaptively configuring remote parameters based on the internet of things according to any one of claims 1-8, wherein the method for adaptively configuring remote parameters further comprises:

10. A remote parameter adaptive configuration system based on the internet of things, comprising a processor and a memory, the memory for storing a computer program, the processor for executing the computer program to implement the method of any one of claims 1-9.