CN113778483A - Cloud-edge-energized equipment decision and safety air upgrading device and method - Google Patents
Cloud-edge-energized equipment decision and safety air upgrading device and method Download PDFInfo
- Publication number
- CN113778483A CN113778483A CN202110961128.4A CN202110961128A CN113778483A CN 113778483 A CN113778483 A CN 113778483A CN 202110961128 A CN202110961128 A CN 202110961128A CN 113778483 A CN113778483 A CN 113778483A
- Authority
- CN
- China
- Prior art keywords
- decision
- equipment
- data
- cloud server
- real
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000013528 artificial neural network Methods 0.000 claims abstract description 28
- 238000013461 design Methods 0.000 claims abstract description 23
- 230000011218 segmentation Effects 0.000 claims abstract description 12
- 238000005516 engineering process Methods 0.000 claims abstract description 10
- 238000010276 construction Methods 0.000 claims abstract description 7
- 230000008569 process Effects 0.000 claims description 18
- 238000003062 neural network model Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 11
- 238000004519 manufacturing process Methods 0.000 claims description 9
- 238000004140 cleaning Methods 0.000 claims description 8
- 230000003993 interaction Effects 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 230000009467 reduction Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000003139 buffering effect Effects 0.000 claims description 2
- 238000010219 correlation analysis Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 description 5
- 230000001174 ascending effect Effects 0.000 description 3
- 244000141353 Prunus domestica Species 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/60—Software deployment
- G06F8/65—Updates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/70—Software maintenance or management
- G06F8/71—Version control; Configuration management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Computer Security & Cryptography (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Animal Behavior & Ethology (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a cloud side enabled equipment decision and safety air upgrading device and method, which take the realization of the method as an example and comprise the steps of collecting text type data generated in an equipment design stage and numerical type data generated in an operation stage; realizing the layered construction and the layered decision of a knowledge map and a deep neural network based on a cloud side architecture; and optimizing the decision model and performing quick version iteration based on the cloud edge architecture, the network technology and the segmentation technology. The method can scientifically and efficiently realize decision making of equipment, and has the outstanding characteristics of realizing the decision making of quantifiable decision and unquantizable decision; the real-time performance and the accuracy of the decision are considered; the method can quickly and safely realize the aerial upgrade of the decision-making model in the equipment, and the privacy can not be leaked due to the data leakage of a single edge node.
Description
Technical Field
The invention relates to a cloud-side energized equipment decision and safety air upgrading device and method, in particular to an equipment decision and safety air upgrading device and method based on a knowledge graph, deep learning, a network technology and a cloud-side cooperation technology.
Background
The external environment of the equipment may change drastically in the operation process, for example, the change of the driving road condition of the unmanned vehicle, the change of the weather of the aircraft in the flight process, and the like, and the decision algorithm deployed in the equipment needs to have adaptability to the change of the external environment.
For decision problems in the equipment operation process, the current methods are mainly divided into two types, namely model-based and rule-based, and have the following limitations: (1) the model has limited fidelity: the external environment in the equipment operation process can be extremely complex, a certain trade-off needs to be made between the fidelity and the complexity of the model in the modeling process, and meanwhile, under the current technical condition, some factors which influence the mechanism to be uncertain and are difficult to model exist. Therefore, it is difficult to completely describe the actual problem by the mathematical model. (2) Is susceptible to human factors: the rule base is usually established by a designer in an off-line mode, and the establishment process depends on a mathematical model and the experience of the designer and is easily influenced by personal factors of the designer. (3) It is difficult to exploit the knowledge contained in big data: the full lifecycle of equipment manufacturing produces a huge amount of data, but the knowledge in it is difficult to use efficiently due to data heterogeneity and the like. (4) Flexibility and adaptability are low: after factors influencing decision making are changed, the existing decision making system is possibly difficult to accurately judge, and the algorithm needs to be upgraded. Generally, decision-making algorithms are designed in advance by designers and are solidified in equipment, so that the equipment is difficult to change the versions of programs and algorithms after being manufactured; (5) the safety problem is as follows: fast algorithm iteration can be realized through an air upgrading technology, but the traditional air upgrading method has the problem of privacy disclosure.
Besides real-time data acquired in the operation process of the equipment, knowledge contained in mass data generated in the design and manufacturing stages of the equipment can also provide help for decision making. With the rapid development of new-generation information technologies such as mobile internet, big data, cloud computing, internet of things, artificial intelligence and the like, a large amount of data in the life cycle of a product can be deeply mined and utilized, and a new solution idea is brought to decision making of equipment.
In view of the above problems, there is an urgent need for a cloud-edge-powered device and method for making decisions and upgrading in the air, so as to implement scientific and efficient decision making and safe and fast version iteration.
Disclosure of Invention
The invention aims to provide a cloud-edge-end-energized equipment decision and safety air upgrading device and method so as to realize scientific decision making and quick program version iteration.
In order to solve the technical problem, the invention provides a cloud-edge-enabled equipment decision and safety air upgrading device, which comprises:
human-computer interaction equipment: interactive equipment such as a computer/mobile terminal and the like is used as a carrier to realize man-machine interaction and finish the operations of algorithm design, decision model design, automatic upgrade strategy formulation and the like;
cloud server: the method is used for realizing the acquisition of data in the stages of equipment design, manufacture and the like, the summarization of real-time data in the operation stage, the construction of a knowledge map, the training of a neural network, the encryption distribution of an upgrade package and the like;
edge nodes: the edge nodes are distributed at different positions and used for bearing wireless communication services, realizing data buffering between the cloud server and the equipment and training a neural network based on local data;
the method comprises the following steps: comprising a plurality of equipment distributed at different locations;
the real-time data acquisition module: is located in the equipment. The system is used for acquiring relevant information sensed by a sensor in the equipment operation stage in real time, such as information of radar, images, sound, temperature, distance and the like;
the heterogeneous data acquisition module: located in the cloud server. The method is used for collecting external unquantized data generated in the equipment design and manufacturing stages, and comprises various design criteria, design manuals, exception handling manuals, handling plans, historical fault analysis reports and the like which are summarized by designers for a long time and stored in a text form. Meanwhile, summarizing the data collected by the real-time data collection module in each device;
a data preprocessing module: and the edge nodes are positioned in the cloud server and the edge nodes and are used for cleaning the acquired data and reducing data noise. If necessary, reducing the dimension of the data so as to reduce the redundancy degree of the data;
a model construction module: located in the cloud server. The cloud server builds a knowledge graph based on unquantized data generated in the stages of equipment design, manufacturing and the like, such as text data of design rules and experience, and by means of various natural language intelligent processing technologies including methods of knowledge extraction, knowledge fusion, knowledge processing, event extraction, knowledge updating and the like. Training a deep neural network according to real-time data generated in each equipment operation stage;
a model segmentation module: located in the cloud server. The system is used for segmenting the deep neural network and the knowledge graph in the cloud server;
a local model training module: located at the edge nodes. The deep neural network is trained according to the real-time data of part of equipment acquired by the edge nodes;
a hierarchical decision module: located in cloud servers, edge nodes, equipment. The method is used for completing decision making in a cloud server, an edge node and equipment according to a deep neural network and a knowledge graph. The equipment dynamically selects a model in the equipment/edge/cloud server to make a corresponding decision according to the requirements of real-time performance and accuracy;
an encryption module: located in the cloud server. For implementing distributed encryption;
a decryption module: is located in the equipment. The file recovery device is used for recovering the encrypted file;
a network module: located in cloud servers, edge nodes, equipment. The method is used for supporting wireless communication among equipment, edge nodes and cloud servers.
In order to solve the technical problem, the invention provides a cloud-edge-enabled equipment decision and safety air upgrading method, which comprises the following steps:
in equipment, dynamically collecting big data generated in the operation process of the equipment based on a real-time data acquisition module, and transmitting the big data to a cloud edge node through a wireless network;
in the edge node, primarily cleaning the data returned by the equipment based on a data cleaning module;
in a cloud server, acquiring heterogeneous data generated in the stages of equipment design, manufacture and the like based on a heterogeneous data acquisition module, and summarizing real-time operation data uploaded by each equipment;
in the cloud server, based on a data preprocessing module, big data generated in the equipment operation process are cleaned. If necessary, reducing the dimension of the cleaned data;
in a cloud server, a knowledge graph is constructed through a model construction module based on unquantizable data generated in equipment design and other stages, and a deep neural network is constructed based on real-time data generated in each equipment operation stage;
and deploying the small-scale knowledge map and the deep neural network in the equipment, and selecting a model deployed in the equipment/edge/cloud end to make a decision according to the requirements of instantaneity and accuracy in the operation stage of the equipment. Based on the knowledge graph, realizing unquantized decision; based on the deep neural network, realizing quantifiable decision;
in a cloud server, segmenting the trained model based on a model segmentation module to generate a medium-scale knowledge map, a medium-scale deep neural network, a small-scale knowledge map and a small-scale deep neural network;
and deploying the complete neural network to a cloud server, and deploying the medium-scale knowledge graph and the medium-scale deep neural network to edge nodes. Integrating the small-scale knowledge map and the small-scale deep neural network into an update package in a cloud server;
the device continuously transmits data back to the edge node, the edge node preliminarily cleans the data based on the data preprocessing module and transmits the data to the cloud server, and the cloud server periodically updates the knowledge graph and the neural network model according to the new data transmitted back.
The cloud server calculates an integrity check code of the upgrade package based on the encryption module, encrypts the upgrade package, and divides the encrypted upgrade package into n sub-volumes (n is a positive integer) based on a division algorithm according to a division sequence O;
comprehensively considering factors such as network delay, signal strength, edge node load and the like, selecting n edge nodes, respectively sending n partial volumes of the upgrade package to the n edge nodes, randomly selecting two nodes from the n nodes, and sending a segmentation sequence O and an integrity check code;
the edge node is communicated with the equipment based on the network module, and transmits the integrity check code, the segmentation sequence O and each upgrade package to the equipment in a bundling manner; and restoring the upgrade package by the equipment according to the segmentation sequence O, generating an integrity check code after decryption, and checking the integrity of the data by comparing the integrity check code with the integrity check code received from the edge node. If the verification is correct, the upgrading operation is executed, otherwise, the error is reported, and retransmission is requested.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic view of an apparatus according to an embodiment of the present invention
FIG. 2 is a schematic flow chart of a method according to an embodiment of the present invention
FIG. 3 is a flow chart of a method according to an embodiment of the present invention
FIG. 4 is a flow chart of a method according to an embodiment of the present invention
FIG. 5 is a flow chart of a method according to an embodiment of the present invention
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings.
The device for realizing comprehensive decision-making and air upgrading of equipment in the embodiment of the invention is shown in figure 1 and comprises: human-computer interaction equipment, a cloud server, an edge node, and a system/device.
S1-1: and (3) a human-computer interaction device. And interactive equipment such as a computer/mobile terminal and the like is used as a carrier to realize man-machine interaction and finish the operations of algorithm design, model design, automatic upgrade strategy formulation and the like.
S1-2: and (4) a cloud server. And the operations of data acquisition, preprocessing, model construction, generation, encryption, distribution and the like of the upgrade package are realized.
S1-3: and (4) edge nodes. The system is composed of a plurality of edge servers, intelligent gateways and other devices distributed at different geographic positions. The edge nodes dynamically collect data collected by the system/equipment, training of the neural network model is achieved, and the edge nodes cooperate with the cloud server to achieve air upgrading of the equipment.
S1-4: and (5) assembling. The system comprises a plurality of pieces of equipment in different geographic positions, and the following modules are added on the basis of traditional equipment:
a network module: the equipment communicates with the edge node through a wireless communication technology, and then real-time communication with the cloud server is achieved.
A sensor module: the device is used for sensing and collecting relevant information inside and outside the device in real time, wherein the relevant information comprises radar, images, sound, temperature, distance and the like. Meanwhile, under the coordination of the edge nodes, a plurality of devices can share part of sensor data, so that a sensor network is formed;
as shown in fig. 2, a cloud-edge enabled equipment decision and security air upgrade method specifically includes the following steps:
step S2-1: and (5) collecting numerical data. Big data generated in the operation process of the equipment are dynamically collected and uploaded to an edge node, and the edge node uploads the big data to a cloud server after preliminarily cleaning the data;
step S2-2: and (4) preprocessing data. And preprocessing the data acquired in the last step through a data cleaning algorithm to reduce noise in the original data. Meanwhile, in the cloud server, based on multiple correlation analysis, influence weights of different dimensions in data collected in the operation process of the equipment are analyzed, and dimension reduction is carried out according to the weights, and the specific process is as follows:
suppose a sample can be denoted as (X)i,Yi) Then the Pearson correlation coefficient of the sample is
Wherein
And sXRespectively, normalized variable, sample mean and sample standard deviation.
Let the dependent variable y and the independent variable x1,x2…xnConstructing a linear model as follows:
note ry,xIs y and x1,x2…xnMultiple correlation coefficient of
Is y and
simple correlation coefficient of (2).
According to the complex correlation coefficient ry,xThe influence weights of different dimensionalities on the decision making after dimensionality reduction can be calculated, and data of dimensionality with lower influence weights can be deleted, so that dimensionality reduction of the data is realized.
Step S2-3: and (5) building a neural network model. And in the cloud server, training a deep neural network model for decision making based on the data subjected to dimensionality reduction.
Step S2-4: and collecting text type data. In the cloud server, external cross-boundary data of equipment design, manufacturing and other stages are collected, and the data comprise various design criteria, design manuals, exception handling manuals, handling plans, historical fault analysis reports and the like which are summarized by designers for a long time and stored in text form.
Step S2-5: and (4) preprocessing data. And preprocessing the data acquired in the last step through a data cleaning algorithm to reduce noise in the original data.
Step S2-6: and (5) constructing a knowledge graph. In the cloud server, a knowledge graph is constructed from big data generated in equipment design and manufacturing stages through natural language intelligent processing technologies such as knowledge extraction, knowledge fusion, knowledge processing, event extraction and knowledge updating.
Step S2-7: and updating the multi-level model. The specific process is shown in fig. 3.
Step S2-7-1: and the cloud server constructs a complete deep neural network model and a knowledge map model according to the global data (real-time data uploaded by all equipment).
Step S2-7-2: and the cloud server prunes the knowledge graph to obtain a medium-scale knowledge graph and a small-scale knowledge graph. Compared with the complete knowledge graph deployed in the cloud server, the trimmed knowledge graph occupies a smaller storage space, but the achievable decision type and accuracy are reduced. And the cloud server prunes the complete neural network to obtain a medium-scale deep neural network model and a small-scale deep neural network model. Compared with a complete deep neural network deployed in a cloud server, the medium-scale and small-scale deep neural networks occupy smaller storage space, and the reasoning process is smaller in calculation amount but limited in precision.
Step S2-7-3: the cloud server deploys the medium-scale knowledge-graph and the medium-scale deep neural network to each edge node.
Step S2-7-4: and the cloud server integrates the small-scale deep neural network and the small-scale knowledge map to generate an upgrade package.
Step S2-8: and (5) updating the equipment program. The specific process is shown in fig. 4.
Step S2-8-1: and the cloud server generates a corresponding upgrading packet according to the latest model and calculates the integrity check code of the upgrading packet.
Step S2-8-2: the cloud server encrypts the upgrade package, and divides the encrypted upgrade package into n parts with approximately the same size through the following algorithm. For convenience of description, n ═ 2 is described below as an example.
Assuming that the size of the encrypted upgrade package is M bytes, consecutive k bytes are extracted each time.
(1) Randomly generating a non-repeated random sequence between 0 and M, wherein the sequence length is M/2 k;
(2) arranging the random sequence in an ascending order;
(3) if the difference between two adjacent random numbers is larger than k, deleting the second random number to obtain a random sequence O;
(4) according to the sequence O, sequentially extracting k bytes from the positions corresponding to the encrypted upgrade packs respectively, and arranging the k bytes in an ascending order to serve as a bundling 1; the rest part is arranged in ascending order as a bundling 2;
step S2-8-3: considering network delay, signal strength, edge node load, and other factors, n edge nodes are selected (for convenience of description, n is 2, that is, two edge nodes A, B are selected respectively).
Step S2-8-4: and sending the integrity check code and the bundling 1 to an edge node A, and sending the sequence O and the bundling 2 to an edge node B.
Step S2-8-5: the edge node A, B communicates with the equipment and sends the integrity check code, sequence O, split 1, split 2 to the equipment.
Step S2-8-6: and restoring the upgrade package by the equipment according to the sequence O, the bundling 1 and the bundling 2, generating an integrity check code after decryption, and checking the data integrity by comparing the integrity check code with the integrity check code received from the edge node.
Step S2-8-7: if the verification is correct, the upgrading operation is executed, otherwise, the error is reported, and retransmission is requested.
Step S2-9: and (5) making a decision. The method is divided into two parts, namely quantifiable decision and non-quantifiable decision, as shown in fig. 5.
Step S2-9-1: and in the equipment operation process, quantifiable data collected by the sensor is dynamically collected.
Step S2-9-2: and selecting the deep neural network with the corresponding scale according to the requirements of real-time performance and accuracy. The real-time requirement of the decision is divided into three levels of high, medium and low. For high-real-time decision, the real-time is preferentially ensured, so that a small-scale deep neural network model deployed in the local equipment is selected for decision making. And for medium real-time decision, considering both real-time performance and accuracy, and selecting a medium-scale deep neural network model deployed at the edge node for decision making. For low-real-time decision making, accuracy is guaranteed to be limited, and a large-scale deep neural network model deployed in a cloud server is selected for decision making.
Step S2-9-3: and preprocessing the data to be used as the input of the deep neural network, and making a corresponding quantifiable decision according to the output of the deep neural network.
Step S2-9-4: and in the equipment operation process, dynamically collecting unquantizable data related to the knowledge graph.
Step S2-9-5: and selecting the knowledge graph with the corresponding scale according to the requirements of feasibility, real-time performance and accuracy. The real-time requirement of the decision is divided into three levels of high, medium and low. For high real-time decision, selecting a small-scale knowledge graph deployed in the local equipment for decision making, and uploading a decision task to an edge node if the small-scale knowledge graph cannot make a corresponding decision; for a medium real-time decision or a decision task (a decision task uploaded to an edge node by equipment) which cannot be completed by an equipment end, selecting a medium-scale knowledge graph deployed at the edge node for decision, and if the edge node cannot make a corresponding decision, uploading the task to a cloud server; and selecting a large-scale knowledge graph deployed in the cloud server for decision making for low real-time decisions or decisions which cannot be completed by the edge nodes.
Step S2-9-6: based on the data, a decision is made based on a knowledge graph and a corresponding decision algorithm. And simultaneously, carrying out rationality check on the decision according to experience knowledge and a pre-designed rule, and if the check is correct, executing a corresponding decision.
Claims (6)
1. The utility model provides a cloud limit end is equipped with decision-making and safe aerial upgrading device of enabling which characterized in that includes:
the human-computer interaction equipment is used for realizing human-computer interaction;
the cloud server is used for bearing tasks such as storage, calculation, decision and the like; the cloud server is provided with a network module, a heterogeneous data acquisition module, a data preprocessing module, a model construction module, a model segmentation module, a layered decision module and an encryption module;
the edge node is used for bearing tasks such as communication, calculation, data buffering, decision and the like; the edge node is provided with a network module, a data preprocessing module, a local model training module and a layering decision module;
the equipment is electromechanical integrated equipment for realizing certain functions, and the working environment is relatively complex; the device is provided with a network module, a real-time data acquisition module, a layered decision module and a decryption module.
2. A cloud edge enabled equipment decision and safety air upgrading method is characterized in that: the method comprises the following steps:
the cloud server collects text data generated in equipment design and other stages and integrates numerical data collected in equipment operation stages;
the cloud server preprocesses the data, including data cleaning, and optionally, dimensionality reduction is performed on numerical data based on complex correlation analysis;
the cloud server constructs a deep neural network and a knowledge graph based on the preprocessed data;
the cloud server cuts out a medium-scale knowledge graph, a medium-scale deep neural network, a small-scale knowledge graph and a small-scale deep neural network through the model;
deploying the medium-scale knowledge graph and the neural network at edge nodes, and deploying the small-scale knowledge graph and the neural network model at equipment;
the equipment collects real-time data and knowledge map related data, and makes a decision in a layered manner according to the requirements of real-time performance and accuracy;
based on the cloud side architecture and the wireless communication technology, the cloud server issues the optimized new model to the equipment periodically, and quick version iteration of the program and the model in the equipment is realized.
3. The method of claim 2, further comprising: the text type data generated in the stages of designing the acquisition equipment and the like and the numerical type data generated in the operation stage comprise:
the cloud server collects external unstructured data generated in the stages of equipment design, manufacturing and the like, wherein the external unstructured data comprises various design criteria, design manuals, exception handling manuals, handling plans, historical fault analysis reports and the like which are summarized by designers for a long time and stored in a text form;
the equipment collects relevant numerical data including radar, images, sound, temperature, distance and the like in the operation process and in the environment in real time, and the cloud server integrates the numerical data collected by the equipment.
4. The method of claim 2, further comprising: the knowledge graph construction comprises the following steps:
preliminarily cleaning the collected text data generated in the stages of product design and the like;
constructing a knowledge graph by means of methods such as knowledge extraction, knowledge fusion, knowledge processing, event extraction and knowledge updating;
and according to the constructed knowledge graph, decision making of the equipment operation stage is realized.
5. The method of claim 2, further comprising: the fast version iteration of the decision model in the equipment comprises the following steps:
continuously returning data to a cloud server by equipment, periodically updating the knowledge graph and the neural network model by the cloud server according to the returned new data, and generating a corresponding upgrade package and an integrity check code;
the cloud server encrypts the upgrade package, and based on a segmentation algorithm, the encrypted upgrade package is segmented into n sub-volumes (n is a positive integer) according to a segmentation sequence O;
comprehensively considering factors such as network delay, signal strength, edge node load and the like, selecting n edge nodes, respectively sending n partial volumes of the upgrade package to the n edge nodes, randomly selecting n nodes from the n nodes, and sending a segmentation sequence O and an integrity check code;
the edge node communicates with the equipment and transmits the integrity check code, the segmentation sequence O and each upgrade package to the equipment in a bundling manner; restoring an upgrade package by the equipment according to the segmentation sequence O, generating an integrity check code after decryption, and checking the integrity of the data by comparing the integrity check code with the integrity check code received from the edge node;
if the verification is correct, the upgrading operation is executed, otherwise, the error is reported, and retransmission is requested.
6. The method of claim 2, further comprising: the hierarchical decision comprises the following steps:
according to the requirements of real-time performance and accuracy, selecting a deep neural network with a corresponding scale to carry out quantifiable decision;
for high real-time decision, the real-time is preferentially ensured, so that a small-scale deep neural network model deployed in the local equipment is selected for decision making; for medium real-time decision, considering both real-time performance and accuracy, selecting a medium-scale deep neural network model deployed at an edge node for decision making; for low real-time decision making, the accuracy is guaranteed in a limited way, and a large-scale deep neural network model deployed in a cloud server is selected for decision making;
selecting a knowledge graph of a corresponding scale to carry out unquantized decision making according to the requirements of feasibility, real-time performance and accuracy;
dividing the real-time requirement of the decision into a high level, a middle level and a low level;
for high real-time decision, selecting a small-scale knowledge graph deployed in the local equipment for decision making, and uploading a decision task to an edge node if the small-scale knowledge graph cannot make a corresponding decision; for a decision task which is decided in a medium real-time manner or cannot be completed by an equipment end, selecting a medium-scale knowledge graph deployed at an edge node for decision, and if the edge node cannot make a corresponding decision, uploading the task to a cloud server; and selecting a large-scale knowledge graph deployed in the cloud server for decision making for low real-time decisions or decisions which cannot be completed by the edge nodes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110961128.4A CN113778483A (en) | 2021-08-20 | 2021-08-20 | Cloud-edge-energized equipment decision and safety air upgrading device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110961128.4A CN113778483A (en) | 2021-08-20 | 2021-08-20 | Cloud-edge-energized equipment decision and safety air upgrading device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113778483A true CN113778483A (en) | 2021-12-10 |
Family
ID=78838572
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110961128.4A Pending CN113778483A (en) | 2021-08-20 | 2021-08-20 | Cloud-edge-energized equipment decision and safety air upgrading device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113778483A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115277414A (en) * | 2022-07-26 | 2022-11-01 | 白犀牛智达(北京)科技有限公司 | Automatic vehicle upgrading system |
| CN117539520A (en) * | 2024-01-10 | 2024-02-09 | 深圳市东莱尔智能科技有限公司 | Firmware self-adaptive upgrading method, system and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107766889A (en) * | 2017-10-26 | 2018-03-06 | 济南浪潮高新科技投资发展有限公司 | A kind of the deep learning computing system and method for the fusion of high in the clouds edge calculations |
| CN112286691A (en) * | 2020-11-12 | 2021-01-29 | 济南浪潮高新科技投资发展有限公司 | Cloud edge-side cooperation method based on heterogeneous decision model generation technology |
| CN112600876A (en) * | 2020-11-25 | 2021-04-02 | 宝能(广州)汽车研究院有限公司 | OTA upgrade package downloading method, OTA server, electronic device and storage medium |
| US11017050B1 (en) * | 2020-04-01 | 2021-05-25 | Vmware Inc. | Hybrid quantized decision model framework |
-
2021
- 2021-08-20 CN CN202110961128.4A patent/CN113778483A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107766889A (en) * | 2017-10-26 | 2018-03-06 | 济南浪潮高新科技投资发展有限公司 | A kind of the deep learning computing system and method for the fusion of high in the clouds edge calculations |
| US11017050B1 (en) * | 2020-04-01 | 2021-05-25 | Vmware Inc. | Hybrid quantized decision model framework |
| CN112286691A (en) * | 2020-11-12 | 2021-01-29 | 济南浪潮高新科技投资发展有限公司 | Cloud edge-side cooperation method based on heterogeneous decision model generation technology |
| CN112600876A (en) * | 2020-11-25 | 2021-04-02 | 宝能(广州)汽车研究院有限公司 | OTA upgrade package downloading method, OTA server, electronic device and storage medium |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115277414A (en) * | 2022-07-26 | 2022-11-01 | 白犀牛智达(北京)科技有限公司 | Automatic vehicle upgrading system |
| CN115277414B (en) * | 2022-07-26 | 2023-10-27 | 白犀牛智达(北京)科技有限公司 | Automatic upgrading system for vehicle |
| CN117539520A (en) * | 2024-01-10 | 2024-02-09 | 深圳市东莱尔智能科技有限公司 | Firmware self-adaptive upgrading method, system and equipment |
| CN117539520B (en) * | 2024-01-10 | 2024-03-19 | 深圳市东莱尔智能科技有限公司 | Firmware self-adaptive upgrading method, system and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qiu et al. | Edge computing in industrial internet of things: Architecture, advances and challenges | |
| Guo et al. | Deep information fusion-driven POI scheduling for mobile social networks | |
| US20240046001A1 (en) | Automated standardized location digital twin and location digital twin method factoring in dynamic data at different construction levels, and system thereof | |
| Festa et al. | An annotated bibliography of GRASP–Part II: Applications | |
| CN103336790B (en) | Hadoop-based fast neighborhood rough set attribute reduction method | |
| CN103336791B (en) | Hadoop-based fast rough set attribute reduction method | |
| CN113778483A (en) | Cloud-edge-energized equipment decision and safety air upgrading device and method | |
| CN117808180B (en) | Path planning method, application and device based on knowledge and data combination | |
| KR20200052806A (en) | Operating method of deep learning based climate change prediction system | |
| CN104573915A (en) | Method and device for constructing high-speed train product structure tree | |
| CN107590749A (en) | A kind of processing method and system with electricity consumption data | |
| CN113032905B (en) | Comprehensive state evaluation method and device for deep space spacecraft | |
| Barolli et al. | Complex, intelligent, and software intensive systems | |
| CN101996257B (en) | Method for searching reconfigurable route exchange platform component | |
| Ryabko et al. | Graph model of Fog Computing system. | |
| Bâra et al. | Improving decision support systems with data mining techniques | |
| CN115827797A (en) | Environmental data analysis and integration method and system based on big data | |
| US20190005169A1 (en) | Dynamic Design of Complex System-of-Systems for Planning and Adaptation to Unplanned Scenarios | |
| CN114780864A (en) | Human-object-space interaction model construction method based on knowledge graph | |
| Si et al. | A hybrid framework of reinforcement learning and convex optimization for uav-based autonomous metaverse data collection | |
| CN118194487A (en) | Automatic arrangement method, medium and system for circuit and electric equipment | |
| CN117852324A (en) | Scene construction method based on data twinning | |
| CN117951636A (en) | Soil pollution detection method, device and medium based on data fusion of Internet of things | |
| CN117459374A (en) | Fault diagnosis method, device, equipment and computer storage medium | |
| WO2022102054A1 (en) | Neural architecture search system and search method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20211210 |
|
| WD01 | Invention patent application deemed withdrawn after publication |