AI Internet of Things multi-domain data sharing method and system

Abstract

The invention discloses an AI Internet of Things multi-domain data sharing method and system, and particularly relates to the technical field of Internet of Things, and the method comprises the following steps: obtaining target domain data, calculating the correlation between the target domain data and different domain data based on correlation weight fusion, generating a correlation vector, further analyzing the correlation degree, and obtaining a correlation vector; judging whether to establish a data pairing relationship according to the result; after all the pairing relationships are summarized, generating a pairing relationship form, and setting an actual time threshold value of the data sharing channel; and then optimizing and distributing resources established by the channels by using a genetic algorithm to finally obtain an optimal resource distribution scheme, and applying the optimal resource distribution scheme to data sharing of a multi-field Internet of Things system to realize efficient cross-field data collaboration and sharing. According to the method, data fusion among a plurality of different fields is supported, which fields of data need to be shared and cooperated is automatically judged through AI, a cross-field and multi-dimensional intelligent data processing scheme is provided, and rapid establishment of a data sharing channel is ensured.

Description

Technical Field

[0001] The present invention relates to the field of Internet of Things technology, and more specifically, to an AI Internet of Things multi-domain data sharing method and system. Background Art

[0002] With the rapid development of the Internet of Things, more and more devices and systems are connected through the Internet, generating massive amounts of data. However, these data are often scattered in different fields and platforms, forming so-called data "islands". Data in various fields are not interconnected, which greatly limits the effective use of data.

[0003] In practical applications, IoT systems such as smart transportation, smart healthcare, and smart city management require comprehensive analysis of data from different fields. For example, the traffic management system needs to share data from the medical emergency system to quickly dispatch resources in the event of a car accident; similarly, the medical system may also need to access traffic data to optimize the dispatch and route planning of emergency vehicles.

[0004] However, most existing IoT systems can only interact with data within a single domain, and lack an intelligent cross-domain data sharing mechanism. In particular, when data between multiple IoT domains needs to be shared, there is a lack of effective association judgment and resource allocation methods, resulting in low efficiency of data sharing and failure to meet real-time data requirements in dynamic and complex scenarios. Therefore, an AI IoT multi-domain data sharing method and system is proposed. Summary of the invention

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] An AI Internet of Things multi-domain data sharing method comprises the following steps:

[0007] Obtain target domain data and use a weight fusion method based on relevance to calculate the correlation between data in different domains, and obtain a correlation vector after the target domain data is pre-paired with multiple different domain data;

[0008] Continue to perform correlation analysis based on the paired correlation vectors, and make correlation level judgment based on the correlation analysis results;

[0009] Determine whether to establish a pairing relationship between the target domain data and the different domain data according to the correlation level obtained by judgment, summarize all established pairing relationships, and obtain a pairing relationship form between the target domain data and multiple different domain data;

[0010] The actual time threshold for establishing the data sharing channel corresponding to the target pairing relationship is set based on the pairing relationship form, and then the genetic algorithm is used to allocate resources for channel establishment, and the final resource allocation plan is obtained and applied.

[0011] In a preferred embodiment, calculating the correlation between data in different fields in a weighted fusion method based on correlation means:

[0012] Get the correlation coefficient r between the pre-paired target domain data and the target different domain data x，vi ; i represents the number corresponding to the pre-pairing, yi represents the type of the target different field corresponding to the pre-pairing number i, and x represents the type of the target field;

[0013] Then assign weights based on the importance of the fields:

[0014] N represents the total number of pre-pairing pairs, and wi represents the allocation weight corresponding to the pre-pairing number i.

[0015] In a preferred embodiment, the correlation vector after pre-pairing refers to the correlation coefficient r corresponding to the pre-pairing number i. x，yi Summarize the importance distribution weights wi to obtain the relevance vector Gi = [r x，yi , wi].

[0016] Correlation coefficient r x，yi It is measured by Pearson's correlation coefficient.

[0017] In a preferred embodiment, continuing to perform correlation analysis based on the paired correlation vectors refers to:

[0018] Get the correlation vector Gi after pairing = [r x，yi , wi], extract the correlation coefficient r x，yi And do the following conversion:

[0019] R1i=|r x，yi |*C;

[0020] R2i=|r x，yi | α *C;

[0021] R3i=ln(1+|r x，yi |)*C;

[0022] α is greater than 1, R1i represents the correlation coefficient r x，yi The corresponding type I conversion value, R2i represents the correlation coefficient r x，yi The corresponding type II conversion value, R3i represents the correlation coefficient r x，yiThe corresponding three-type conversion value, and then all types of conversion values ​​​​are r x，yi After replacement, we obtain the vector Fi=[R1i, R2i, R3i, wi] under the correlation analysis.

[0023] In a preferred embodiment, judging the correlation level according to the correlation analysis result means:

[0024] The vector Fi=[R1i, R2i, R3i, wi] under the correlation analysis is taken as the input variable, and the correlation level of the pre-pairing number i is taken as the output variable. The input variable is fuzzified and the value of the input variable is converted into a fuzzy set. The output variable is fuzzified and the output variable is converted into a fuzzy set. Fuzzy rules are formulated to describe the correlation levels under different data type combinations. The fuzzified input variables are inferred through fuzzy rules to obtain the correlation level of the pre-pairing number i.

[0025] In a preferred embodiment, determining whether to establish a pairing relationship between the target domain data and the different domain data according to the determined correlation level refers to:

[0026] Obtain the relevant level of the pre-pairing number i, and compare it with the preset level threshold. If the relevant level of the pre-pairing number i is greater than or equal to the preset level threshold, the pairing relationship of the pre-pairing number i is retained; if the relevant level of the pre-pairing number i is less than the preset level threshold, the pairing relationship of the pre-pairing number i is deleted.

[0027] In a preferred embodiment, the logic for obtaining the actual time threshold for establishing the data sharing channel is:

[0028] Di represents the relevant level corresponding to the retained pairing relationship i, fi represents the preset urgency adjustment coefficient in the current scenario, T0 represents the preset basic duration, Zi represents the preset load index data of the data fusion module in the current scenario, ZB represents the standard value corresponding to the load index data of the data fusion module in the current scenario, and Si represents the actual time threshold for establishing the data sharing channel corresponding to the pairing relationship i.

[0029] In a preferred embodiment, resource allocation for channel establishment using a genetic algorithm refers to:

[0030] Encoding and initial population: Encode the channel resource allocation parameter group, i.e., the decision variable, into a chromosome form and randomly generate m chromosomes as the initial population;

[0031] Fitness evaluation: The constraint condition is defined as the establishment time of each data sharing channel cannot exceed its actual time threshold. At the same time, the total establishment time of all data sharing channels is calculated. The total establishment time is the fitness value of the chromosome. The goal is to minimize the total establishment time of all data sharing channels.

[0032] Selection operation: Use the roulette wheel selection method to select offspring as new parents;

[0033] Crossover operation: randomly exchange data in chromosomes of different parents;

[0034] Mutation operation: randomly select data of different offspring chromosomes for adjustment;

[0035] Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal resource allocation plan.

[0036] In a preferred embodiment, an AI IoT multi-domain data sharing system includes:

[0037] The pre-pairing module obtains the target domain data and calculates the correlation between the data in different domains by using a weight fusion method based on relevance, and obtains the correlation vector after the target domain data is pre-paired with the data in multiple different domains;

[0038] The secondary pairing module continues to perform correlation analysis based on the correlation vector after pairing, and performs correlation level judgment according to the correlation level analysis result, and determines whether to establish a pairing relationship between the target domain data and the different domain data according to the correlation level obtained by judgment, and summarizes all the established pairing relationships to obtain a pairing relationship table between the target domain data and multiple different domain data;

[0039] The data sharing channel establishment module sets the actual time threshold for establishing the data sharing channel corresponding to the target pairing relationship based on the pairing relationship form, and then uses the genetic algorithm to allocate resources for channel establishment, obtains the final resource allocation plan and applies it.

[0040] Technical effects and advantages of the present invention:

[0041] The present invention realizes intelligent data sharing among multiple fields of the Internet of Things through AI technology, breaking the data "island" problem in traditional systems. It can dynamically analyze the correlation of data in different application scenarios (such as smart transportation and smart medical care), ensure the efficient flow of valuable data between different fields, and improve data utilization.

[0042] By optimizing the allocation of resources in the data sharing process through genetic algorithms, the present invention can automatically adjust the allocation scheme of network bandwidth, CPU and memory according to different scenarios and real-time changes in system load. Compared with the traditional fixed resource allocation method, the present invention can reduce resource waste while ensuring the rapid establishment of data sharing channels, thereby improving the overall operating efficiency of the system.

[0043] The present invention is particularly suitable for emergency scenarios (such as car accidents, natural disasters, etc.). Through intelligent emergency adjustment mechanisms and fast data channel establishment, the delay of data transmission and processing can be greatly shortened. The system can quickly share high-priority data according to real-time needs, greatly improving the emergency response speed and providing effective support for on-site decision-making.

[0044] The present invention supports data fusion between multiple different fields. Through AI correlation analysis, the system can automatically determine which fields of data need to be shared and collaborated, providing a cross-domain, multi-dimensional intelligent data processing solution, which is particularly suitable for complex scenarios such as smart city management, medical emergency, and smart transportation.

[0045] The present invention ensures the stability of the system under high load by strictly controlling the establishment time and resource allocation of the data sharing channel. At the same time, through correlation analysis and optimized resource allocation strategy, unnecessary data transmission is reduced, system overload or delay is avoided, and the security and reliability of the data sharing process are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0047] Figure 1 This is a schematic diagram of an AI Internet of Things multi-domain data sharing method in the present invention.

[0048] Figure 2 This is a schematic diagram of an AI Internet of Things multi-domain data sharing system in the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Reference Figure 1 - Figure 2 The following examples are obtained:

[0051] Embodiment 1: An AI IoT multi-domain data sharing method comprises the following steps:

[0052] Multi-dimensional data is collected in different fields (such as transportation, medical care, industry, etc.) through IoT devices. The data includes structured and unstructured types. The collected data is cleaned, formatted and standardized to ensure that the data in different fields are comparable and consistent. Then the target field data is obtained and the correlation between data in different fields is calculated using a weight fusion method based on correlation to obtain the correlation vector after the target field data and multiple different field data are pre-paired; the purpose of this step is to identify the correlation between the target field data and other different field data. By calculating the correlation of data in each field, the potential connection between cross-domain data can be determined, providing a basis for subsequent data sharing. The weight fusion method further refines the correlation calculation of data in different fields to ensure accurate results.

[0053] A divergent domain is a data domain that is different from the target domain and has an independent data source. The data in a divergent domain is usually related to the data in the target domain, but comes from a different system or platform. A divergent domain can cover multiple IoT application scenarios, and the need to share data is determined by calculating the data correlation between these domains. The following are some typical examples of divergent domains:

[0054] In the field of intelligent transportation, the medical emergency system belongs to a different field. For example, in a traffic accident scenario, data such as the arrival time of the ambulance, the availability of emergency resources (such as the number of medical staff and the availability of emergency equipment) belong to a different field. These data are associated with the target field (such as traffic flow and accident location) to help make quick decisions and resource scheduling.

[0055] In smart city management, the fire protection system can be considered as a different domain. For example, in a building fire scenario, the arrival time of the fire truck and the dispatch data of fire protection resources belong to different domains. These data are associated with the data of the traffic management system and monitoring system of the smart city, and sharing these data can improve rescue efficiency.

[0056] In the agricultural IoT system, environmental monitoring data can be used as different fields. For example, data such as weather conditions and soil moisture of farmland come from the environmental monitoring system, which is different from the data of the agricultural production system (such as crop growth and irrigation data). These different fields of data can help the intelligent agricultural system optimize production decisions.

[0057] Based on the paired correlation vectors, the correlation analysis is continued, and the correlation level is judged according to the correlation analysis results; the correlation analysis is used to evaluate the specific correlation strength between the target domain data and the data in different domains. Through further analysis, the correlation level judgment provides a decision-making basis for the adaptability of cross-domain data and helps identify which fields of data are most suitable for sharing and collaboration.

[0058] According to the correlation level obtained, it is decided whether to establish a pairing relationship between the target domain data and the different domain data. All the established pairing relationships are summarized to obtain a pairing relationship form between the target domain data and multiple different domain data. When the correlation level of some different domain data meets the requirements, these data will establish a pairing relationship with the target domain data. By summarizing the established pairing relationship form, it can be clarified which data will be used for sharing. This step realizes the transition from theoretical analysis to actual data sharing and prepares for the implementation of data sharing.

[0059] The actual time threshold for establishing the data sharing channel corresponding to the target pairing relationship is set based on the pairing relationship form. The time threshold is set to optimize the establishment process of the shared channel and ensure the efficiency of data sharing. According to the needs of different data and the sharing time requirements, setting a reasonable time threshold can maximize resource utilization and reduce delays. Then use the genetic algorithm to allocate resources for channel establishment, obtain the final resource allocation plan and apply it. The genetic algorithm is used to optimize resource allocation to ensure that when establishing a cross-domain data sharing channel, each resource can be allocated in the best way. The application of the genetic algorithm can effectively improve sharing efficiency, reduce resource waste, and ensure the stability of the shared channel. The final resource allocation plan is applied to ensure the smooth progress of data sharing.

[0060] The weight fusion method based on correlation calculates the correlation between data in different fields, which means:

[0061] Get the correlation coefficient r between the pre-paired target domain data and the target different domain data x，yi; i represents the number corresponding to the pre-pairing, yi represents the type of target different field corresponding to the pre-pairing number i, and x represents the type of target field; the correlation coefficient is used to measure the strength of the association between the target field data and the different field data. This coefficient usually takes a value between -1 and 1. The closer the value is to 1, the stronger the positive correlation between the two, the closer it is to -1, the stronger the negative correlation, and close to 0 means that there is almost no correlation between the two. The weight is assigned according to the importance of each field, reflecting the priority of a certain field in the pairing process. Weight fusion will make the correlation calculation more accurate, taking into account the degree of influence of data in different fields. The correlation vectors after pairing are summarized. Through the correlation vector Gi, the degree of correlation between the data in each field can be comprehensively reflected. Combined with the weight, a more accurate correlation measurement result can be generated for subsequent correlation level judgment.

[0062] The weight indicates the weight assigned to each field during the pairing process. It is assigned based on the importance of the field. A larger weight indicates that the field has a higher priority in the pairing, and vice versa.

[0063] N represents the total number of pre-pairing pairs, and wi represents the weight assigned to the pre-pairing number i. The correlation vector after pre-pairing refers to the correlation coefficient r corresponding to the pre-pairing number i. x，yi Summarize the importance distribution weights wi to obtain the relevance vector Gi = [r x，yi , wi]. The correlation vector combines the results of the correlation coefficient and the weight, reflecting the strength and importance of the association between the target field and the data in different fields.

[0064] Correlation coefficient r x，yi It is measured by the Pearson correlation coefficient. For example, in a car accident scenario, the correlation coefficient between the traffic management field and the medical health field is determined as follows:

[0065] Define variables: First, it is necessary to identify the key data variables between the traffic management field and the medical and health field in order to determine the relationship between them. For example: traffic flow (such as the number of vehicles passing the accident site per unit time); vehicle speed at the accident site; road congestion (which can be measured by congestion index or speed distribution); traffic light status at the time of the accident (ratio of red lights to green lights); arrival time of the ambulance; number of injured and severity of injuries; availability of hospital emergency resources (such as number of beds, emergency equipment, number of medical staff); response time of ambulance dispatch.

[0066] By collecting historical data, key variables in the fields of traffic management and healthcare can be statistically analyzed. For example, the following information can be obtained from historical records of traffic accidents: traffic flow and congestion at the accident site, the time it takes for the ambulance to arrive at the scene after the alarm, the distribution of the arrival time of the ambulance under different congestion conditions or traffic flows, the use of medical resources in the hospital and the rescue response time. These data can be collected through the Internet of Things and other means to form a multidimensional data set.

[0067] It should be noted that the target domain data and multiple different domain data mentioned in the present invention are pre-paired, which means that in a certain application scenario such as a car accident scenario, the target domain data refers to the target data variable in the field such as traffic flow, and the different domain data refers to the arrival time of the ambulance and the number of medical staff. The correlation coefficient between the traffic flow and the arrival time of the ambulance is calculated, and the correlation coefficient between the traffic flow and the number of medical staff is calculated. Finally, the correlation strength of the relevant data can be judged in the current application scenario, providing a basis for whether data sharing is ultimately needed.

[0068] Continuing the correlation analysis based on the paired correlation vector means: obtaining the paired correlation vector Gi = [r x，yi , wi], extract the correlation coefficient r x，yi And do the following conversion:

[0069] R1i=|r x，yi |*C; R2i=|r x，yi | α *C; R3i = ln(1+|r x，yi |)*C;

[0070] α is greater than 1, C is a positive constant, and R1i represents the correlation coefficient r x，yi The corresponding type 1 conversion value is amplified by its absolute value. Here C is a constant factor and is a positive number, which is usually used to amplify or reduce the converted value. The role of the absolute value is to avoid the positive and negative signs affecting the conversion results and focus on the correlation strength itself. R2i represents the correlation coefficient r x，yi The corresponding type II transformation value is calculated by x，yi The value is further magnified or reduced by raising it to the αth power. The purpose of setting α greater than 1 is to enhance the sensitivity to high correlation coefficients, that is, the stronger the correlation, the greater the difference in the value obtained after the power amplification. For weak correlation, the difference will not be too obvious due to the effect of the αth power. R3i represents the correlation coefficient r x，yiThe corresponding three types of conversion values ​​are transformed by logarithmic conversion. The characteristic of logarithmic function is to expand smaller values ​​and compress larger values. Therefore, this conversion can amplify the coefficients with weak correlation, and at the same time, it can suppress the extremely large or extremely small correlation values ​​to prevent them from dominating the results. Then all types of conversion values ​​are transformed by r x，yi After replacement, we get the vector Fi = [R1i, R2i, R3i, wi] under the correlation analysis. This comprehensive vector reflects the correlation calculation results under multiple conversion methods, so that the high correlation data can be more significantly amplified, which is convenient for more detailed correlation analysis and judgment.

[0071] Judging the relevance level based on the relevance analysis results means:

[0072] The vector Fi = [R1i, R2i, R3i, wi] under the correlation analysis is used as the input variable, and the correlation level of the pre-pairing number i is used as the output variable. The input variable is fuzzified, and the value of the input variable is converted into a fuzzy set. The output variable is fuzzified, and the output variable is converted into a fuzzy set. Fuzzy rules are formulated to describe the correlation level under different data type combinations. The fuzzified input variables are inferred through fuzzy rules to obtain the correlation level of the pre-pairing number i. The correlation level is the final result judgment of the input variable, indicating that the correlation degree between the input data is divided into different levels. The goal of the level division is to grade the correlation under different combinations for subsequent use or further analysis. The first step in the fuzzy logic process is to fuzzify the input correlation vector. Convert the precise numerical data into linguistic fuzzy sets, such as "low correlation", "medium correlation" and "high correlation". This step is to facilitate processing in the subsequent fuzzy reasoning. According to historical data or prior knowledge, fuzzy rules are formulated to describe the corresponding correlation levels under different correlation combinations. For example, there may be the following rules: if the primary conversion value is high, the secondary conversion value is high, and the tertiary conversion value is medium, then the correlation level is high; if the primary conversion value is low, the secondary conversion value is low, and the tertiary conversion value is high, then the correlation level is low. Through these rules, different correlation vector combinations can be mapped to corresponding levels. Logical inference based on fuzzy rules and fuzzified input variables. After each input variable is converted into a fuzzy set, it is matched through the fuzzy rule set to infer a corresponding correlation level. The result of the inference is expressed in multiple input dimensions, giving a comprehensive judgment of different correlations. Finally, the result of the fuzzy inference is defuzzified to obtain a clear correlation level output. For example, through inference, it can be obtained that the correlation level of a certain data pair is "high" or "medium". This process is an output from the fuzzy space back to the actual numerical value. The whole process converts multiple different correlation conversion values ​​and weight inputs into linguistic fuzzy sets through the fuzzy logic system, and after reasoning with fuzzy rules, a specific correlation level is obtained. This level is used to make the final evaluation and judgment of the correlation between the target domain data and the data in different domains.

[0073] Determining whether to establish a pairing relationship between the target domain data and the different domain data based on the correlation level obtained by judgment means: obtaining the correlation level of the pre-pairing number i, comparing it with the preset level threshold, if the correlation level of the pre-pairing number i is greater than or equal to the preset level threshold, then the pairing relationship of the pre-pairing number i is retained, if the correlation level of the pre-pairing number i is less than the preset level threshold, then the pairing relationship of the pre-pairing number i is deleted. In the previous steps, the correlation level of each pre-pairing number was obtained through fuzzy logic reasoning. This level represents the strength of association between the target domain data and the different domain data. By assigning a correlation level to each pairing number, we can evaluate the quality of each data pairing.

[0074] In order to ensure the quality and effectiveness of data sharing, there is a preset level threshold in the system. This threshold represents the minimum standard for establishing a data pairing relationship. Only when the correlation level of two data is higher than or equal to the threshold, the system will consider the association between these data to be strong enough to keep the pairing relationship. When the correlation level of a pre-pairing number is higher than or equal to the preset level threshold, it means that the target domain data and the different domain data have a strong correlation. At this time, the system believes that these data are valuable for sharing, so their pairing relationship will be retained, allowing data sharing or resource collaboration to continue in subsequent processes. If the correlation level of a pre-pairing number is lower than the preset level threshold, it means that the target domain data and the different domain data have a weak correlation. At this time, the system believes that the sharing between these data does not have much practical significance or does not have enough correlation value. Therefore, the pairing relationship will be deleted and the data will no longer be shared. This process helps to exclude insignificant data pairs, reduce resource waste and the burden of data processing.

[0075] By comparing the correlation level with the threshold, the system can filter out high-quality, highly relevant data pairs for sharing, thereby improving the efficiency and reliability of data sharing. This avoids unnecessary exchange of low-relevance data and ensures that system resources are allocated reasonably.

[0076] The logic for obtaining the actual time threshold for establishing a data sharing channel is:

[0077] Di represents the correlation level corresponding to the retained pairing relationship i. The higher the correlation level, the stronger the correlation between the data pairs, and the greater the necessity of data sharing. Therefore, 1-Di*fi is used to reduce the impact on the actual time. The higher the correlation level, the shorter the time to establish the shared channel. fi represents the preset urgency adjustment coefficient in the current scenario, which is used to adjust the time requirement in a specific scenario. For example, in emergency scenarios (such as disaster relief), time is particularly important, and the adjustment coefficient will affect the establishment time of the data sharing channel according to the urgency. The higher the urgency, the shorter the time. This reflects the flexibility of the system, which can adjust the response speed according to the actual scenario. T0 represents the preset basic duration, which is the basic time for establishing the data sharing channel, indicating the standard time required for channel establishment when there is no other factor affecting it. The introduction of basic time can ensure that the system has a standardized time frame. Zi represents the preset load index data of the data fusion module in the current scenario, and ZB represents the standard value corresponding to the load index data of the data fusion module in the current scenario. These two indicators are used to measure the system load in the current scenario. The greater the system load, the longer the resource allocation time will be. On the contrary, the smaller the system load, the shorter the time to establish the channel. Standardizing and comparing load indicator data can dynamically adjust the time to establish a channel so that the system can operate reasonably under different load conditions. Load indicator data is one of the parameters of various system performance and resource usage in the AI ​​IoT multi-domain data sharing system, depending on the application scenario of the system. The following are some common load indicator data examples, one of which is CPU usage: CPU usage indicates the degree of processor load during the operation of the system. If the CPU usage is close to 100%, it means that the system is running at full load and the resources are close to the limit. By monitoring the CPU usage, the establishment time of the data sharing channel can be dynamically adjusted, and the time can be extended when the CPU load is high to avoid further system overload. Memory usage reflects the RAM (random access memory) resources consumed during the operation of the system. If the memory usage is high (for example, more than 80%), the performance of the system may decline, especially when large-scale data processing is involved. High memory load may affect the speed of data sharing, so it can also be used as a load indicator. Si represents the actual time threshold for the establishment of the data sharing channel corresponding to the pairing relationship i. The smaller the time threshold, the faster the data sharing channel should be established, reducing the waiting time. This indicates that under current conditions, the system can establish a data sharing channel for the pairing relationship at a faster speed to ensure timely data transmission and use. If the data sharing pairing involves an emergency scenario (adjusted by the urgency adjustment coefficient fi), the system will accelerate the establishment of the channel to ensure that data can be quickly transmitted and used in the emergency scenario.This is crucial for emergency response to unexpected events. For example, in the event of a traffic accident, urgent medical data and traffic data need to be shared as quickly as possible.

[0078] Resource allocation for channel establishment using genetic algorithms refers to:

[0079] Encoding and initial population: First, the resource allocation problem of the data sharing channel needs to be converted into a form that can be processed by the genetic algorithm. Here, the resource allocation parameters of each channel (such as bandwidth, CPU allocation, memory, etc.) are encoded as chromosomes, that is, each chromosome represents a possible resource allocation scheme. The initial population is formed by randomly generating multiple (for example, m) different chromosomes. These chromosomes represent the possible resource allocation combinations that the system may adopt. The main purpose of this step is to provide the genetic algorithm with multiple possible solutions to ensure that the initial exploration of the algorithm is wide-ranging.

[0080] Fitness evaluation: The fitness value is used to measure the quality of each chromosome (i.e., resource allocation scheme). Here, the fitness value is measured by the total establishment time, that is, the sum of the establishment time of each channel after the system allocates resources to all data sharing channels. The goal is to minimize this total time as much as possible to ensure that all channels are established as soon as possible. The total time cannot exceed the actual time threshold of each channel (this threshold is calculated according to the formula discussed earlier). The purpose of fitness evaluation is to evaluate each resource allocation scheme to see if it is effective and how efficient it is. A lower total establishment time means a higher fitness value, which means that the scheme can complete the establishment of all channels faster and the resource allocation is more reasonable.

[0081] Selection operation: The selection operation retains chromosomes with higher fitness by simulating the process of natural selection. Through the "roulette wheel selection method", each chromosome is given a corresponding probability to enter the next generation according to its fitness. Chromosomes with high fitness are more likely to be selected, thereby increasing the chances of high-quality solutions continuing to evolve in subsequent iterations. This process ensures that better resource allocation solutions are retained while eliminating poorer solutions, so that the overall resource allocation solution in the population is gradually optimized.

[0082] Crossover operation: The crossover operation simulates the process of gene recombination by exchanging data (resource allocation parameters) in different parent chromosomes. In this way, the new chromosome may inherit the excellent characteristics of the parent chromosome and further explore new solution spaces to generate better resource allocation solutions. The crossover operation can generate more new possibilities by combining existing resource allocation solutions, thereby increasing the chance of finding the optimal solution.

[0083] Mutation operation: Mutation operation is used in genetic algorithms to prevent the population from falling into the local optimal solution. By randomly selecting certain genes (resource allocation parameters) in the offspring chromosomes for fine-tuning, more possible solutions can be explored and the chance of finding the global optimal solution can be increased. For example, randomly increasing or decreasing the bandwidth allocation of a certain channel, or changing the proportion of CPU resource allocation. Ensuring that the population has sufficient diversity can prevent the algorithm from falling into the local optimal solution and ensure the global search capability of the genetic algorithm.

[0084] Iteration and termination conditions: After the algorithm completes the selection, crossover, and mutation operations in each generation, it will continue to iterate and generate new populations until the preset termination conditions are met (such as reaching the specified number of iterations, or the fitness value of the optimal solution found has not changed significantly in consecutive generations). When the termination conditions are met, the chromosome with the highest fitness is selected from the final population, and this chromosome is the optimal resource allocation solution. Through continuous iteration, the genetic algorithm can gradually optimize the population and eventually find a resource allocation solution that can minimize the total time to establish a data sharing channel.

[0085] The final resource allocation plan is a combination of parameters that represents the resources allocated to each data sharing channel. For example: Channel 1: Bandwidth 50Mbps, CPU 10%, Memory 500MB; Channel 2: Bandwidth 30Mbps, CPU 15%, Memory 600MB; Channel 3: Bandwidth 70Mbps, CPU 20%, Memory 700MB; This resource allocation plan can minimize the total establishment time of all data sharing channels, ensure that each channel is established as soon as possible within the specified time threshold, and effectively utilize system resources.

[0086] Embodiment 2: An AI IoT multi-domain data sharing system, comprising:

[0087] The pre-pairing module obtains the target domain data and calculates the correlation between the data in different domains by using a weight fusion method based on relevance, and obtains the correlation vector after the target domain data is pre-paired with the data in multiple different domains;

[0088] The secondary pairing module continues to perform correlation analysis based on the correlation vector after pairing, and performs correlation level judgment according to the correlation level analysis result, and determines whether to establish a pairing relationship between the target domain data and the different domain data according to the correlation level obtained by judgment, and summarizes all the established pairing relationships to obtain a pairing relationship table between the target domain data and multiple different domain data;

[0089] The data sharing channel establishment module sets the actual time threshold for establishing the data sharing channel corresponding to the target pairing relationship based on the pairing relationship form, and then uses the genetic algorithm to allocate resources for channel establishment, obtains the final resource allocation plan and applies it.

[0090] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0091] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0092] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0093] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0094] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An AI Internet of Things multi-domain data sharing method, characterized in that: The following steps are involved: Obtain target domain data and use a weight fusion method based on relevance to calculate the correlation between data in different domains, and obtain a correlation vector after the target domain data is pre-paired with multiple different domain data; Continue to perform correlation analysis based on the paired correlation vectors, and make correlation level judgment based on the correlation analysis results; Determine whether to establish a pairing relationship between the target domain data and the different domain data according to the correlation level obtained by judgment, summarize all established pairing relationships, and obtain a pairing relationship form between the target domain data and multiple different domain data; The actual time threshold for establishing the data sharing channel corresponding to the target pairing relationship is set based on the pairing relationship form, and then the genetic algorithm is used to allocate resources for channel establishment, and the final resource allocation plan is obtained and applied.

2. According to claim 1, the AI ​​Internet of Things multi-domain data sharing method is characterized in that: The weight fusion method based on correlation calculates the correlation between data in different fields, which means: Get the correlation coefficient r between the pre-paired target domain data and the target different domain data x，yi ; i represents the number corresponding to the pre-pairing, yi represents the type of the target different field corresponding to the pre-pairing number i, and x represents the type of the target field; Then assign weights based on the importance of the fields: N represents the total number of pre-pairing pairs, and wi represents the allocation weight corresponding to the pre-pairing number i.

3. The AI ​​IoT multi-domain data sharing method according to claim 2 is characterized in that: The correlation vector after pre-pairing refers to the correlation coefficient r corresponding to the pre-pairing number i x，yi Summarize the importance distribution weights wi to obtain the relevance vector Gi = [r x，yi , wi].

4. The AI ​​IoT multi-domain data sharing method according to claim 3 is characterized in that: Correlation coefficient r x，yi It is measured by Pearson's correlation coefficient.

5. The AI ​​IoT multi-domain data sharing method according to claim 4 is characterized in that: Continuing the correlation analysis based on the paired correlation vector means: Get the correlation vector Gi after pairing = [r x，yi , wi], extract the correlation coefficient r x，yi And do the following conversion: R1i=|r x，yi |*C; R2i=|r x，yi | α *C; R3i=ln(1+|r x，yi |)*C; α is greater than 1, C is a positive constant, and R1i represents the correlation coefficient r x，yi The corresponding type I conversion value, R2i represents the correlation coefficient r x，yi The corresponding type II conversion value, R3i represents the correlation coefficient r x，yi The corresponding three-type conversion value, and then all types of conversion values ​​​​are r x，yi After replacement, we obtain the vector Fi=[R1i, R2i, R3i, wi] under the correlation analysis.

6. The AI ​​IoT multi-domain data sharing method according to claim 5 is characterized in that: Judging the relevance level based on the relevance analysis results means: The vector Fi=[R1i, R2i, R3i, wi] under the correlation analysis is taken as the input variable, and the correlation level of the pre-pairing number i is taken as the output variable. The input variable is fuzzified and the value of the input variable is converted into a fuzzy set. The output variable is fuzzified and the output variable is converted into a fuzzy set. Fuzzy rules are formulated to describe the correlation levels under different data type combinations. The fuzzified input variables are inferred through fuzzy rules to obtain the correlation level of the pre-pairing number i.

7. The AI ​​IoT multi-domain data sharing method according to claim 6 is characterized in that: Determining whether to establish a pairing relationship between the target domain data and the different domain data based on the correlation level obtained by judgment refers to: Obtain the relevant level of the pre-pairing number i, and compare it with the preset level threshold. If the relevant level of the pre-pairing number i is greater than or equal to the preset level threshold, the pairing relationship of the pre-pairing number i is retained; if the relevant level of the pre-pairing number i is less than the preset level threshold, the pairing relationship of the pre-pairing number i is deleted.

8. The AI ​​IoT multi-domain data sharing method according to claim 7 is characterized in that: The logic for obtaining the actual time threshold for establishing a data sharing channel is: Di represents the relevant level corresponding to the retained pairing relationship i, fi represents the preset urgency adjustment coefficient in the current scenario, T0 represents the preset basic duration, Zi represents the preset load index data of the data fusion module in the current scenario, ZB represents the standard value corresponding to the load index data of the data fusion module in the current scenario, and Si represents the actual time threshold for establishing the data sharing channel corresponding to the pairing relationship i.

9. The AI ​​IoT multi-domain data sharing method according to claim 8, characterized in that: Resource allocation for channel establishment using genetic algorithms refers to: Encoding and initial population: Encode the channel resource allocation parameter group, i.e., the decision variable, into a chromosome form and randomly generate m chromosomes as the initial population; Fitness evaluation: The constraint condition is defined as the establishment time of each data sharing channel cannot exceed its actual time threshold. At the same time, the total establishment time of all data sharing channels is calculated. The total establishment time is the fitness value of the chromosome. The goal is to minimize the total establishment time of all data sharing channels. Selection operation: Use the roulette wheel selection method to select offspring as new parents; Crossover operation: randomly exchange data in chromosomes of different parents; Mutation operation: randomly select data of different offspring chromosomes for adjustment; Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal resource allocation plan.

10. An AI Internet of Things multi-domain data sharing system, implemented based on an AI Internet of Things multi-domain data sharing method according to any one of claims 1 to 9, characterized in that: include: The pre-pairing module obtains the target domain data and calculates the correlation between the data in different domains by using a weight fusion method based on relevance, and obtains the correlation vector after the target domain data is pre-paired with the data in multiple different domains; The secondary pairing module continues to perform correlation analysis based on the correlation vector after pairing, and performs correlation level judgment according to the correlation level analysis result, and determines whether to establish a pairing relationship between the target domain data and the different domain data according to the correlation level obtained by judgment, and summarizes all the established pairing relationships to obtain a pairing relationship table between the target domain data and multiple different domain data; The data sharing channel establishment module sets the actual time threshold for establishing the data sharing channel corresponding to the target pairing relationship based on the pairing relationship form, and then uses the genetic algorithm to allocate resources for channel establishment, obtains the final resource allocation plan and applies it.

Images