Information screening optimization method and system for sensing communication

Abstract

The invention discloses an information screening optimization method for sensing communication. The method comprises the following steps of S1, information acquisition and preprocessing, S2, information feature model establishment, S3, screening rule and weight setting, S4, information screening and evaluation, and S5, dynamic screening strategy adjustment. An information screening optimization system for sensing communication comprises an information acquisition module, a preprocessing module, a feature extraction module, a rule setting module, a screening evaluation module and a strategy adjustment module. According to the method, noise and error data are removed through preprocessing, fuzzy comprehensive evaluation, the D-S evidence theory and the genetic algorithm are combined, information is evaluated from multiple dimensions, multi-aspect attributes of the information are fully considered, comprehensive evaluation is achieved, the quality of screening results is improved, meanwhile, multi-source evidence information is fused through the D-S evidence theory, and the screening efficiency is improved. The internal association between different information is considered, the deep value contained in the information combination is mined, and the information utilization rate is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of sensing communication, and in particular, to an information screening and optimization method and system for sensing communication. Background Art

[0002] In the modern field of sensing communication, with the continuous development and popularization of sensor technology, the types and quantities of sensors are increasing day by day. From the equipment status monitoring in industrial production, to the meteorological parameter collection in environmental monitoring, and then to the perception of various environmental and device information in smart homes, a large amount of sensing information is continuously generated.

[0003] The existing sensing information screening methods have many limitations. Most of the methods have simple principles and only screen information based on a single dimension or a few indicators, lacking a comprehensive consideration and comprehensive evaluation of the multi-faceted attributes of information. In the face of complex and diverse application scenarios and changing actual needs, these methods are difficult to make flexible and effective adjustments. In addition, the inherent relevance between information is often ignored, and the deep value contained in the information combination cannot be excavated. The uncertainty existing in the information itself is also insufficiently considered, making the screening results greatly reduced in accuracy and reliability, and it is difficult to effectively meet the requirements for high-quality sensing information in actual applications.

[0004] Therefore, this application proposes an information screening and optimization method and system for sensing communication. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose an information screening and optimization method and system for sensing communication.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] An information screening and optimization method for sensing communication, comprising the following steps:

[0008] S1. Information collection and preprocessing: Collect sensing information through various sensors, and preprocess the collected information;

[0009] S2. Establish an information feature model: Extract features from the preprocessed information to establish an information feature model. The feature extraction uses a statistics-based method to convert the sensing information into a representative feature vector for subsequent analysis and screening;

[0010] S3. Set screening rules and weights: According to different application scenarios and requirements, set corresponding information screening rules and assign weights to each rule;

[0011] S4. Information Screening and Evaluation: By comprehensively applying fuzzy comprehensive evaluation, D-S evidence theory, and genetic algorithm, respectively obtain the fuzzy matching degree between information and screening rules, the final evaluation result of information, and the optimal screening rules and weight combinations. Based on this, screen the sensing information, calculate the comprehensive score, retain the high-score information and process the low-score information. At the same time, evaluate the screening result. If the requirements are not met, adjust the rules and weights according to the optimal combination and re-screen.

[0012] S5. Dynamically Adjust the Screening Strategy: According to the real-time environmental changes and changes in application requirements, dynamically adjust the screening rules and weights. By real-time monitoring and analyzing the changes in environmental changes and application requirements, timely adjust the screening strategy.

[0013] Preferably, in the step S1, the preprocessing includes data cleaning, denoising, and normalization operations; the data cleaning is to remove the obviously incorrect or abnormal data in the collected information, the denoising process uses a filtering algorithm to remove noise interference, and the normalization process is to convert data in different ranges into a unified range.

[0014] Preferably, in the step S2, the feature extraction method based on statistics includes calculating at least one statistic such as the average value, maximum value, minimum value, standard deviation, and change rate of the sensing information to construct a feature vector.

[0015] Preferably, in the step S3, the information screening rules are set based on the importance, timeliness, and relevance of the information, and the weights assigned to each rule are determined according to the influence degree of the rule on the overall screening goal.

[0016] Preferably, in the step S4, the fuzzy comprehensive evaluation determines the evaluation factor set, the comment set, and the factor weight vector, calculates the membership matrix of each information under each comment level through the membership function, and then obtains the fuzzy matching degree;

[0017] The D-S evidence theory assigns basic probability assignments to each evidence source, and uses the evidence combination rule to fuse the basic probability assignments of multiple evidence sources to determine the final evaluation result of the information;

[0018] The genetic algorithm encodes the screening rules and weights into chromosomes, randomly generates an initial population, defines a fitness function, and evolves the population through selection, crossover, and mutation operations until the termination condition is met to obtain the optimal combination.

[0019] Preferably, in the step S5, the real-time monitoring of environmental change parameters and application requirement indicators, when the fluctuation of the environmental change parameters or application requirement indicators exceeds the preset threshold, dynamically adjust the screening rules and weights.

[0020] An information screening and optimization system for sensing communication, comprising:

[0021] An information acquisition module, which is used to collect sensing information through various sensors and transmit the collected information to the preprocessing module;

[0022] A preprocessing module, connected to the information acquisition module, which is used to preprocess the collected information, including data cleaning, denoising and normalization operations, and send the preprocessed information to the feature extraction module;

[0023] A feature extraction module, connected to the preprocessing module, which is used to extract features from the preprocessed information, convert the sensing information into a representative feature vector by using a statistics-based method, establish an information feature model, and transmit the established information feature model to the rule setting module and the screening and evaluation module;

[0024] A rule setting module, connected to the feature extraction module, which is used to set corresponding information screening rules according to different application scenarios and requirements, assign weights to each rule, and send the set screening rules and weights to the screening and evaluation module;

[0025] A screening and evaluation module, respectively connected to the feature extraction module and the rule setting module, which is used to comprehensively apply fuzzy comprehensive evaluation, D-S evidence theory and genetic algorithm to obtain the fuzzy matching degree between the information and the screening rules, the final evaluation result of the information, and the optimal combination of screening rules and weights respectively, screen the sensing information accordingly, calculate the comprehensive score, retain the high-score information and process the low-score information, and at the same time evaluate the screening result. If the screening result does not meet the requirements, re-screen after adjusting the rules and weights according to the optimal combination, and send the screening result to the strategy adjustment module;

[0026] A strategy adjustment module, connected to the screening and evaluation module, which is used to dynamically adjust the screening rules and weights according to the real-time environmental changes and changes in application requirements, timely adjust the screening strategy by real-time monitoring and analyzing the environmental changes and changes in application requirements, and feedback the adjusted screening rules and weights to the screening and evaluation module

[0027] The present invention has the following beneficial effects:

[0028] 1. Remove noise and incorrect data through preprocessing, and comprehensively evaluate the information from multiple dimensions by combining fuzzy comprehensive evaluation, D-S evidence theory and genetic algorithm, fully consider the multi-faceted attributes of the information, achieve a comprehensive and comprehensive evaluation, improve the quality of the screening result, and at the same time fuse multi-source evidence information through D-S evidence theory, consider the internal correlation between different information, explore the deep value contained in the information combination, improve the information utilization rate, and provide more valuable information for practical applications.

[0029] 2. By flexibly setting the screening rules and weights according to different application scenarios and actual requirements, and dynamically adjusting the screening strategy to respond in real time to changes in the environment and requirements, the screening method can effectively adapt to complex and diverse scenarios and meet the ever-changing needs.

[0030] 3. Through preprocessing operations such as data cleaning and denoising, and by comprehensively applying a variety of algorithms to handle information uncertainty, the information error and the impact of uncertainty are reduced, the accuracy and reliability of the screening results are improved, and the requirements for high-quality sensing information in practical applications are effectively met. Brief Description of the Drawings

[0031] Figure 1 It is the overall flowchart of an information screening and optimization method for sensing communication proposed by the present invention;

[0032] Figure 2 It is the ER diagram of the smart home sensing information processing in the first embodiment of the present invention;

[0033] Figure 3 It is the ER diagram of the industrial equipment sensing information processing in the second embodiment of the present invention;

[0034] Figure 4 It is the ER diagram of the intelligent transportation sensing information processing in the third embodiment of the present invention. Detailed Embodiments

[0035] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0036] An information screening and optimization method for sensing communication includes the following steps:

[0037] S1. Information collection and preprocessing: Sensing information is collected through various sensors. The sensors can convert physical quantities, chemical quantities, etc. into sensing information in the form of electrical signals or digital signals. However, due to factors such as the accuracy limitation of the sensors themselves and environmental interference, the collected information may contain errors, outliers, or noise, which will affect the subsequent data processing and analysis results.

[0038] It is necessary to preprocess the collected information. The preprocessing includes data cleaning, denoising, and normalization operations; data cleaning is to remove the obviously wrong or abnormal data in the collected information. When data outside the normal range appears, it can be determined as abnormal data and excluded.

[0039] The denoising process uses the mean filtering algorithm to remove noise interference, which can smooth the signal, remove noise interference, and improve the quality of the data;

[0040] Normalization is to unify data in different ranges into a specific interval. The advantage of doing this is to eliminate the dimensional differences between different data features, making subsequent analysis and calculations more accurate and efficient;

[0041] S2. Establish an information feature model: Extract features from the preprocessed information to establish an information feature model. Feature extraction uses a statistics-based method. Statistics-based feature extraction methods include calculating at least one statistic among the average value of sensing information (the result obtained by dividing the sum of all data in a set of data by the number of data in this set, which reflects the central tendency of the data and can represent the overall level of this set of sensing information), the maximum value (the largest numerical value in a set of data, which reflects the upper limit of the data and can reflect the maximum degree that the sensing information may reach), the minimum value (the smallest numerical value in a set of data, and the minimum value reflects the lower limit of the data, which helps to understand the lowest level that the sensing information may appear), the standard deviation (the arithmetic square root of the variance, which measures the degree of dispersion of a set of data relative to the average value and can reflect the fluctuation of the data. The larger the standard deviation, the more dispersed the data and the greater the fluctuation; the smaller the standard deviation, the more concentrated the data and the smaller the fluctuation), and the change rate (indicating the change speed of the data over time or other variables, which can reflect the dynamic change of the sensing information and is very useful for monitoring the change trend and abnormal conditions of the system) to construct a feature vector, and convert the sensing information into a representative feature vector for subsequent analysis and screening;

[0042] S3. Set screening rules and weights: According to different application scenarios and requirements, set corresponding information screening rules and assign weights to each rule. The information screening rules are set based on the importance, timeliness, and relevance of the information, and the weights assigned to each rule are determined according to the impact degree of the rule on the overall screening goal;

[0043] S4. Information screening and evaluation: Comprehensively use fuzzy comprehensive evaluation, D-S evidence theory, and genetic algorithm to obtain the fuzzy matching degree between the obtained information and the screening rules, the final evaluation result of the information, and the optimal screening rule and weight combination;

[0044] Fuzzy comprehensive evaluation determines the evaluation factor set, comment set, and factor weight vector, calculates the membership matrix of each information at each comment level through the membership function, and then obtains the fuzzy matching degree, which can more comprehensively consider various features and uncertainties of the information and improve the accuracy and rationality of screening;

[0045] The D-S evidence theory assigns basic probability assignments to each evidence source and uses the evidence combination rule to fuse the basic probability assignments of multiple evidence sources to determine the final evaluation result of the information;

[0046] The genetic algorithm encodes the screening rules and weights into chromosomes, randomly generates an initial population, defines a fitness function, and evolves the population through selection, crossover, and mutation operations until the termination condition is met to obtain the optimal combination;

[0047] Based on this, the sensing information is screened and the comprehensive score is calculated, the high-score information is retained and the low-score information is processed. At the same time, the screening result is evaluated. If the requirements are not met, the rules and weights are adjusted according to the optimal combination and then screened again;

[0048] S5. Dynamically adjust the screening strategy: According to the real-time environmental changes and changes in application requirements, when the fluctuations of environmental change parameters or application requirement indicators exceed the preset threshold, dynamically adjust the screening rules and weights. By real-time monitoring and analyzing the environmental changes and changes in application requirements, timely adjust the screening strategy.

[0049] An information screening and optimization system for sensing communication, including:

[0050] An information acquisition module, which is used to collect sensing information through various sensors and transmit the collected information to the preprocessing module;

[0051] A preprocessing module, connected to the information acquisition module, which is used to preprocess the collected information, including data cleaning, denoising, and normalization operations, and send the preprocessed information to the feature extraction module;

[0052] A feature extraction module, connected to the preprocessing module, which is used to extract features from the preprocessed information, convert the sensing information into a representative feature vector by using a statistics-based method, establish an information feature model, and transmit the established information feature model to the rule setting module and the screening evaluation module;

[0053] A rule setting module, connected to the feature extraction module, which is used to set corresponding information screening rules according to different application scenarios and requirements, assign weights to each rule, and send the set screening rules and weights to the screening evaluation module;

[0054] A screening evaluation module, respectively connected to the feature extraction module and the rule setting module, which is used to comprehensively apply fuzzy comprehensive evaluation, D-S evidence theory, and genetic algorithm to obtain the fuzzy matching degree between the information and the screening rules, the final evaluation result of the information, and the optimal combination of screening rules and weights. Based on this, the sensing information is screened and the comprehensive score is calculated, the high-score information is retained and the low-score information is processed. At the same time, the screening result is evaluated. If the screening result does not meet the requirements, the rules and weights are adjusted according to the optimal combination and then screened again, and the screening result is sent to the strategy adjustment module;

[0055] A strategy adjustment module, connected to the screening and evaluation module, is used to dynamically adjust the screening rules and weights according to real-time environmental changes and changes in application requirements. By monitoring and analyzing environmental changes and changes in application requirements in real time, the screening strategy is adjusted in a timely manner, and the adjusted screening rules and weights are fed back to the screening and evaluation module.

[0056] Example 1:

[0057] Step 1: Information collection and preprocessing

[0058] In a smart home environment monitoring system, sensing information of the indoor environment is collected through temperature sensors, humidity sensors, light intensity sensors, etc., and this information is preprocessed. When cleaning data, the normal temperature range is set as [10°C, 30°C], the normal humidity range is [30%, 70%], and the normal light intensity range is [0, 1000 Lux]. Data outside this range is determined as abnormal data and excluded. The mean filter algorithm is used to denoise the data, smooth the signal, and remove noise interference. Data such as temperature, humidity, and light intensity is normalized to the [0, 1] interval to eliminate the dimension difference and improve the accuracy and efficiency of subsequent analysis and calculation.

[0059] Step 2: Establish an information feature model

[0060] Feature extraction is performed on the preprocessed information such as temperature, humidity, and light intensity. Calculate the average value, maximum value, minimum value, standard deviation, and change rate of temperature;

[0061] Average value calculation formula:

[0062] where xi represents the i-th data and n is the number of data.

[0063] Standard deviation calculation formula:

[0064] Change rate calculation formula (assuming time as a variable, xt represents the data at time t):

[0065]

[0066] For example, the average temperature reflects the average temperature level indoors, and the change rate reflects the rate of change of temperature. Similar statistics of humidity and light intensity are calculated to construct a feature vector and establish an indoor environment information feature model for subsequent analysis and screening.

[0067] Step 3: Set screening rules and weights

[0068] According to the application requirements of smart homes, set information screening rules. When the indoor temperature is too high (exceeding 28°C) and the humidity is low (below 40%), it is considered that the environmental comfort is poor. This rule has a high importance and is assigned a high weight. When the light intensity is too strong (exceeding 800 Lux) and the time is at night, it is considered that there is an abnormal situation. Set the corresponding rules and assign a certain weight. Determine the weight of each rule according to the degree of influence of the rule on the overall screening goal (such as ensuring indoor environmental comfort and safety).

[0069] Step Four: Information Screening and Evaluation

[0070] Determine the evaluation factor set as the relevant characteristics of temperature, humidity, and light intensity (such as average value, change rate, etc.) through the fuzzy comprehensive evaluation method. The comment set is "comfortable", "relatively comfortable", "uncomfortable", "abnormal". Determine the factor weight vector according to expert experience and actual needs. The weight of the relevant characteristics of temperature is relatively high because temperature has a greater impact on comfort. Calculate the membership matrix of each environmental information under each comment level through the membership function. For example, for temperature information, calculate the membership under comment levels such as "comfortable", "relatively comfortable", "uncomfortable", "abnormal" according to its characteristic value and membership function, and then obtain the fuzzy matching degree between the information and the screening rule;

[0071] Let the evaluation factor set U = {u1, u2 ···, u n},the comment set V = {v1, v2 ···, v m},the factor weight vector W = (w1, w2, ···, w n ), where

[0072] The factor weight vector μij(x) represents the membership degree of factor u i to comment v j . The membership matrix where r ij = μij(xi).

[0073] Fuzzy comprehensive evaluation result where is the fuzzy composition operator. Common ones are the maximum-minimum composition operator (∧ represents taking the minimum, ∨ represents taking the maximum).

[0074] Screen the sensing information according to the fuzzy matching degree, calculate the comprehensive score, retain the high-score information (such as a high comprehensive score indicates that the environmental state is good or meets expectations), process the low-score information (such as issuing an alarm or prompting the user to make adjustments), and at the same time evaluate the screening result. If the requirements are not met (such as the user feedback that the environment is still uncomfortable), then fine-tune the factor weight vector according to the actual situation, etc., and re-screen.

[0075] Step Five: Dynamically Adjust the Screening Strategy

[0076] Monitor the changes in the indoor environment in real time, such as seasonal changes and changes in user activity patterns. When the season changes from summer to winter, the preset threshold of the environmental temperature may need to be adjusted. At this time, when the fluctuation of the environmental temperature change parameter or application requirement index (such as the user's requirement for temperature comfort) exceeds the preset threshold, dynamically adjust the screening rules and weights. The threshold for too high temperature can be lowered in winter, and at the same time, the weight for the appropriate temperature range can be increased to timely adjust the screening strategy to adapt to the new environment and requirements.

[0077] Example Two:

[0078] Step One: Information Collection and Preprocessing

[0079] In the industrial equipment operation monitoring system, collect the sensing information of equipment operation through vibration sensors, temperature sensors, pressure sensors, etc., and perform preprocessing. During data cleaning, according to the normal operation parameter range of the equipment, eliminate the abnormal data outside the range. Use the mean filtering algorithm to denoise and smooth the signal. Normalize the data such as vibration amplitude, temperature, and pressure to a unified interval to eliminate the dimension difference and prepare for subsequent processing.

[0080] Step Two: Establish an Information Feature Model

[0081] Extract the features of the preprocessed vibration, temperature, pressure and other information. Calculate the average value, maximum value, minimum value, standard deviation and change rate of the vibration information (where the calculation formulas for the average value, standard deviation and change rate are the same as those in Example One), which reflect the intensity, stability and change trend of equipment vibration; calculate similar statistics for temperature and pressure information, construct a feature vector, and establish an information feature model of equipment operation status for subsequent analysis and screening;

[0082] Step Three: Set Screening Rules and Weights

[0083] According to the application scenarios and requirements of industrial equipment operation monitoring, set information screening rules. When the vibration amplitude of the equipment exceeds the normal range and the temperature rises, it may indicate potential equipment failures. Set this rule and assign a higher weight; when the pressure fluctuates abnormally, set the corresponding rule and assign a weight according to its impact on equipment operation. Determine the weight of each rule according to its impact on the overall screening goal (such as timely detecting equipment failures and ensuring the normal operation of equipment).

[0084] Step Four: Information Screening and Evaluation

[0085] Using the D-S evidence theory, vibration sensors, temperature sensors, pressure sensors, etc. are regarded as different evidence sources, and basic probability assignments are assigned to each evidence source. For example, according to the historical data and reliability of the sensors, different basic probability assignments are assigned to propositions such as normal operation, minor faults, and serious faults of the equipment for the vibration sensor.

[0086] Let m i (A) represent the basic probability assignment of the i-th evidence source to proposition A, and the basic probability assignments of multiple evidence sources are fused using the evidence combination rule. When the vibration sensor detects abnormal vibration amplitude and the temperature sensor detects an increase in temperature, the basic probability assignments of the two are fused through the evidence combination rule to determine the final evaluation result of the equipment operation state, that is, the probability that the equipment is in normal operation, minor faults, or serious faults.

[0087] Evidence combination rule (m1 and m2 are two basic probability assignment functions, A is a proposition):

[0088]

[0089] where is the conflict coefficient.

[0090] Based on this, the sensing information is screened. Information indicating that the equipment may have faults (such as a high probability of serious faults) in the evaluation results is retained and further analyzed, and information indicating normal operation of the equipment is appropriately processed. At the same time, the screening results are evaluated. If it is found that the screening results do not match the actual operation of the equipment (such as the equipment actually has a fault but is not accurately screened out), the basic probability assignments of each evidence source are readjusted, and the information screening and evaluation are carried out again.

[0091] Step Five: Dynamically adjust the screening strategy

[0092] Monitor the changes in the industrial production environment in real time, such as factors like production task adjustment and equipment aging. When the production task change leads to an increase in the equipment operation load, the preset threshold of the equipment operation parameters may need to be changed. At this time, when the fluctuations of the environmental change parameters or application requirement indicators exceed the preset threshold, the screening rules and weights are dynamically adjusted. For example, increase the screening rule weights for abnormal changes in key equipment parameters (such as vibration amplitude and temperature), and timely adjust the screening strategy to adapt to the changes in the equipment operation state.

[0093] Example Three:

[0094] Step One: Information collection and preprocessing

[0095] In the intelligent traffic flow monitoring system, sensing information of traffic flow is collected through flow sensors, speed sensors, vehicle type sensors, etc., and preprocessed. When cleaning data, abnormal data is removed according to the normal range of traffic flow and the reasonable parameter range of vehicle driving. The mean filter algorithm is used for denoising to make the data smoother. The flow data, speed data, etc. are normalized to a suitable interval to eliminate the dimension difference for subsequent processing.

[0096] Step 2: Establish an information feature model

[0097] Feature extraction is performed on the preprocessed information such as flow, speed, and vehicle type. Calculate the average value, maximum value, minimum value, standard deviation, and change rate of the flow information (where the calculation formulas for the average value, standard deviation, and change rate are the same as those in Embodiment 1), which reflect the overall level, peak situation, fluctuation degree, and change trend of traffic flow; calculate similar statistics of the speed information, construct a feature vector, and establish a traffic flow information feature model to prepare for information screening.

[0098] Step 3: Set screening rules and weights

[0099] According to the application scenarios and requirements of intelligent traffic flow monitoring, set information screening rules. For example, when the traffic flow of a certain section exceeds a certain threshold and the vehicle speed is generally low, it is considered that there may be congestion in this section, set this rule and assign a higher weight; when the flow of a specific type of vehicle (such as a large truck) increases abnormally in certain sections, set the corresponding rule and assign a weight according to its impact on the traffic condition. Determine the weight of each rule according to its influence on the overall screening goal (such as accurately monitoring traffic congestion and optimizing traffic flow).

[0100] Step 4: Information screening and evaluation

[0101] The genetic algorithm is used to encode the screening rules and weights into chromosomes. For example, binary encoding is used to represent the opening and closing of the rules and the magnitude of the weights. Randomly generate an initial population, define a fitness function, and use the accuracy (such as the proportion of correctly identified congested sections) and effectiveness (such as the contribution to traffic flow optimization) of the screening results as indicators.

[0102] The fitness function F(x), where x represents the chromosome (combination of screening rules and weights), and calculate its fitness value according to specific indicators.

[0103] Through the selection operation, select the chromosomes with higher fitness; perform the crossover operation to exchange some genes of the chromosomes to generate new chromosomes; perform the mutation operation to randomly change some genes of the chromosomes.

[0104] A common method for the selection operation is the roulette wheel selection method. Let the population size be N, and the fitness of individual i be F(x i) Then the probability that individual i is selected is:

[0105] Crossover operation (taking single-point crossover as an example): Randomly select a crossover point and exchange the gene segments after the crossover point of the two parental chromosomes to generate offspring chromosomes;

[0106] Mutation operation: With a certain mutation probability p m , flip the genes on the chromosome (when in binary coding) and through multiple generations of evolution, when the termination condition is met (such as the fitness no longer improves or the preset number of generations of evolution is reached), the optimal screening rules and weight combinations are obtained. According to the optimal combination, the sensing information is screened, the comprehensive score is calculated, the high-score information (such as information closely related to traffic congestion) is retained, and the low-score information (such as information with little impact on traffic conditions) is processed. At the same time, the screening results are evaluated. If the screening results cannot meet the requirements of traffic management (such as still unable to effectively relieve congestion), the genetic algorithm is re-run, the screening rules and weights are adjusted, and screening and evaluation are performed again.

[0107] Step Five: Dynamically adjust the screening strategy

[0108] Real-time monitor the changes in traffic flow and the changes in traffic management requirements, such as abnormal traffic flow caused by holidays, special events, etc. When the fluctuations of the traffic flow change parameters or application requirement indicators (such as the adjustment of traffic control measures) exceed the preset threshold, dynamically adjust the screening rules and weights. During holidays, increase the weights of the traffic flow screening rules for main roads and roads around scenic spots, and timely adjust the screening strategy to adapt to the traffic conditions during special periods.

[0109] It should be noted that in the comparative examples, Comparative Example 1 uses the traditional simple threshold method, and its accuracy in screening information is poor because its screening rules are relatively simple and it is unable to comprehensively consider various characteristics and uncertainties of the information, making it difficult to deeply understand that the accuracy rate is 65%. Comparative Example 2 uses the single statistic screening method, but it is not comprehensive enough in dealing with the characteristics and uncertainties of the information, and its accuracy rate is 72%. While in Example 1, through the algorithm based on fuzzy comprehensive evaluation, and by using the membership function to comprehensively consider multiple factors to determine the matching degree between the information and the screening rules, its accuracy rate reaches 85%. Example 2 uses the D-S evidence theory to take different sensors as evidence sources and fuse probability assignments, and the accuracy rate is 82%. Example 3 uses the genetic algorithm, by encoding the screening rules and weights, and using selection, crossover, and mutation operations to evolve the population to find the optimal combination, which can effectively screen the information, so that the accuracy rate of information screening is 84%. Specifically, as shown in Table 1:

[0110] Table 1: Comparison table of information screening, complex scenario information, and screening results for examples and comparative examples

[0111]

[0112] It should be noted that in the comparative examples, the information accuracy rates of Examples 1, 2, and 3 in the complex scenario information accuracy are 90%, 88%, and 85% respectively in the complex scenario, which are also higher than 80% of Comparative Example 1 and 82% of Comparative Example 2. This further reflects the superiority of the method in the examples when facing complex scenarios, and it can screen information more accurately, reduce misjudgments. Among them, the advantage of Example 1 in the complex scenario is more obvious, indicating that the fuzzy comprehensive evaluation method has a good effect in dealing with the information fuzziness and uncertainty in complex scenarios;

[0113] In terms of the stability of the screening results, the stability of the screening results of Examples 1, 2, and 3 are 97%, 96%, and 95% respectively, all higher than 92% of Comparative Example 1 and 94% of Comparative Example 2. This means that the methods of the three examples can maintain a relatively stable screening effect in multiple experiments, with small result fluctuations and high reliability. Among them, the stability of Example 1 is the highest, indicating that after determining the evaluation factor set, comment set, and weight vector, the fuzzy comprehensive evaluation method can perform information screening relatively stably.

[0114] It should be noted that in the comparative examples, the accuracy rate of Comparative Example 1 when part of the data is missing is 55%, and it may not be able to effectively use limited information for accurate screening. The accuracy rate of Comparative Example 2 is 50%, and it is highly sensitive to data loss and lacks an effective mechanism to cope with data incompleteness;

[0115] The accuracy rate of Example 1 is 68%, indicating that in the case of partial data loss, the information screening optimization method based on fuzzy comprehensive evaluation can still maintain a relatively high accuracy rate;

[0116] The accuracy rate of Example 2 is 65%, and the method based on D-S evidence theory also has a certain effect in dealing with partial missing data;

[0117] The accuracy rate of Example 3 is 60%, and the method based on genetic algorithm has a relatively low accuracy rate when part of the data is missing, as shown in Table 2 specifically:

[0118] Table 2: Comparison table of examples and comparative examples under partial data loss, information dynamic change, and sudden change information

[0119]

[0120] It should be noted that in terms of the information dynamic change tracking error rate, the error rate of the first embodiment is 20%. Fuzzy comprehensive evaluation can adjust the membership degree and evaluation results in a timely manner, and the tracking effect is good. The error rate of the second embodiment is 25%. Although the D-S evidence theory can fuse multi-source information, there is a certain lag when dealing with rapidly changing information. The error rate of the third embodiment is 30%. The evolution process of the genetic algorithm is difficult to adapt to rapid changes in a timely manner, resulting in a large error. The error rate of Comparative Example 1 is 40%, and that of Comparative Example 2 is 35%. Traditional methods have great limitations in tracking information dynamic changes.

[0121] In terms of the accuracy rate under sudden change information, the accuracy rate of the first embodiment is 70%. The flexibility of fuzzy comprehensive evaluation enables it to quickly adapt to mutations. The second embodiment performs best, reaching 75%. The D-S evidence theory can effectively fuse evidence source information to cope with sudden situations. The accuracy rate of the third embodiment is 60%. Due to the slow evolution process of the genetic algorithm, it is difficult to adjust quickly. The accuracy rate of Comparative Example 1 is only 40%, and that of Comparative Example 2 is 45%. Traditional methods can hardly maintain a high accuracy rate during sudden changes and are difficult to meet actual needs.

[0122] Specifically, as Figure 2 shown, the relationship expression of the smart home environment monitoring system in the first embodiment is relatively clear and accurate. The "1:N" identifier indicates that the sensor entity and the sensing information have a one-to-many relationship, that is, one sensor entity (such as a temperature and humidity sensor, a light sensor) can continuously collect and generate multiple pieces of sensing information (temperature and humidity values at different time points, light intensity values, etc.).

[0123] The many-to-many relationship between the sensing information and the screening rules is represented by "M:N", which means that one piece of sensing information (such as the temperature sensing information at a certain moment) may be related to multiple screening rules (such as the rule of too high temperature, the rule of abnormal temperature change, etc.); conversely, one screening rule (such as the rule of too high temperature) can also be applied to multiple different pieces of sensing information (temperature sensing information in different rooms and at different times).

[0124] The sensing information is connected to the evaluation result and the environmental parameter threshold through "1:1" respectively. The "1:1" relationship between the sensing information and the evaluation result reflects that after a piece of sensing information (such as the humidity sensing information at a specific moment) is analyzed and processed, it will only correspond to one evaluation result (such as "humidity is comfortable" or "humidity is uncomfortable"); the "1:1" between the sensing information and the environmental parameter threshold means that a certain piece of sensing information (such as the current light intensity sensing information) will be compared with the corresponding environmental parameter threshold (such as the suitable light intensity range) to determine whether it is in a normal state. Through these relationship identifiers and connections in the whole figure, the associations and operation logics among the entities in the smart home environment monitoring system in the first embodiment are effectively presented.

[0125] Specifically, as Figure 3As shown, the second embodiment is accurately expressed. The annotation of "1:N" clearly reflects the one-to-many relationship between the sensor entity and the sensing information, that is, one sensor entity (such as vibration, temperature sensors, etc.) can continuously generate multiple pieces of sensing information (vibration amplitudes, temperature values, etc. at different times).

[0126] The sensing information is associated with other entities through different paths: it is connected to the screening rule through "feature extraction and condition matching", which means that according to the preset screening rule, features are extracted from the sensing information (such as extracting features such as frequency and amplitude from vibration sensing information) and matched with the conditions to judge whether there is an abnormality in the equipment operation; it is connected to the equipment parameter threshold through "threshold comparison", that is, the value in the sensing information (such as temperature value) is compared with the parameter threshold for normal operation of the equipment (such as normal temperature range) to judge whether it exceeds the normal range; between the sensing information and the equipment operation state, through the "analysis and comparison judgment" node, the results of the screening rule and the threshold comparison are comprehensively considered to determine the operation state of the equipment (normal, faulty, etc.). In addition, as an evidence source, the sensor entity determines the equipment operation state through the process of "assignment, using evidence combination and fusing information", which accurately presents the core logic of fusing multiple sensor information based on the D-S evidence theory to judge the equipment state in the second embodiment. Overall, the construction of the elements and relationships in the figure effectively expresses the association and operation mechanism of each entity in the industrial equipment operation monitoring system in the second embodiment.

[0127] Specifically, as Figure 4 shown, this figure clearly presents the relevant logic of the third embodiment. The "sensor entity" is connected to multiple entities of the same type in a "1:N" relationship, indicating that one main sensor entity can be associated with multiple slave sensor entities, or a sensor category covers multiple instances for collecting various information in the traffic field.

[0128] The "screening rule" is in the core position. The "chromosome" is connected to the screening rule through "providing optimization parameters and strategies", and the "population" is connected to the screening rule through "chromosome set", implying the use of genetic algorithms (involving the concepts of chromosome and population) to optimize the screening rule. The screening rule comprehensively processes the information collected by multiple sensor entities, and then obtains the "traffic condition assessment result", and uses the "traffic parameter threshold" as the evaluation and judgment standard, which completely presents the association and operation logic among the elements in the traffic condition assessment system in the third embodiment.

[0129] Generally speaking, the embodiments of the present invention fuse through a unique innovative information screening algorithm, evaluate information from multiple dimensions, fully consider various attributes of the information, achieve comprehensive and overall evaluation, improve the quality of the screening results, and can provide more accurate and reliable monitoring and analysis results in practical applications.

[0130] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, making equivalent substitutions or changes, shall be covered by the protection scope of the present invention.

Claims

1. An information screening and optimization method for sensing communication, characterized in that Including the following steps: S1. Information acquisition and preprocessing: Collect sensing information through various sensors and preprocess the collected information; S2. Establish an information feature model: Extract features from the preprocessed information, establish an information feature model. The feature extraction uses a statistics-based method to convert the sensing information into a representative feature vector for subsequent analysis and screening; S3. Set screening rules and weights: According to different application scenarios and requirements, set corresponding information screening rules and assign weights to each rule; S4. Information screening and evaluation: Comprehensively use fuzzy comprehensive evaluation, D-S evidence theory, and genetic algorithm to obtain the fuzzy matching degree between the information and the screening rules, the final evaluation result of the information, and the optimal combination of screening rules and weights respectively. Based on this, screen the sensing information, calculate the comprehensive score, retain the high-score information and process the low-score information, and at the same time evaluate the screening result. If the requirements are not met, adjust the rules and weights according to the optimal combination and rescreen; S5. Dynamically adjust the screening strategy: According to the real-time environmental changes and changes in application requirements, dynamically adjust the screening rules and weights. By real-time monitoring and analyzing the changes in environmental changes and application requirements, adjust the screening strategy in a timely manner.

2. The information screening and optimization method for sensing communication according to claim 1, wherein In step S1, the preprocessing includes data cleaning, denoising, and normalization operations; the data cleaning is to remove the obviously incorrect or abnormal data in the collected information, the denoising process uses a filtering algorithm to remove noise interference, and the normalization process is to convert data in different ranges into a unified range.

3. An information screening and optimization method for sensing communication according to claim 1, characterized in that In step S2, the statistics-based feature extraction method includes calculating at least one statistic such as the average value, maximum value, minimum value, standard deviation, and change rate of the sensing information to construct a feature vector.

4. The information screening and optimization method for sensing communication according to claim 1, characterized in that In step S3, the information screening rules are set based on the importance, timeliness, and relevance of the information, and the weights assigned to each rule are determined according to the influence degree of the rule on the overall screening goal.

5. The information screening and optimization method for sensing communication according to claim 1, wherein In step S4, the fuzzy comprehensive evaluation determines the evaluation factor set, comment set, and factor weight vector, calculates the membership matrix of each information under each comment level through the membership function, and then obtains the fuzzy matching degree; The D-S evidence theory assigns basic probability assignments to each evidence source and uses the evidence combination rule to fuse the basic probability assignments of multiple evidence sources to determine the final evaluation result of the information; The genetic algorithm encodes the screening rules and weights into chromosomes, randomly generates an initial population, defines a fitness function, and evolves the population through selection, crossover, and mutation operations until the termination condition is met to obtain the optimal combination.

6. The information screening and optimization method for sensing communication according to claim 1, wherein In step S5, the environmental change parameters and application requirement indicators are monitored in real time. When the fluctuations of the environmental change parameters or application requirement indicators exceed the preset threshold, the screening rules and weights are dynamically adjusted.

7. An information screening and optimization system for sensing communication, characterized in that, Including: An information acquisition module for collecting sensing information through various sensors and transmitting the collected information to the preprocessing module; A preprocessing module, connected to the information acquisition module, for preprocessing the acquired information, including data cleaning, denoising, and normalization operations, and sending the preprocessed information to the feature extraction module; A feature extraction module, connected to the preprocessing module, for extracting features from the preprocessed information, converting the sensing information into representative feature vectors using a statistics-based method, establishing an information feature model, and transmitting the established information feature model to the rule setting module and the screening and evaluation module; A rule setting module, connected to the feature extraction module, for setting corresponding information screening rules according to different application scenarios and requirements, assigning weights to each rule, and sending the set screening rules and weights to the screening and evaluation module; A screening and evaluation module, respectively connected to the feature extraction module and the rule setting module, for comprehensively applying fuzzy comprehensive evaluation, D-S evidence theory, and genetic algorithm to obtain the fuzzy matching degree between the information and the screening rules, the final evaluation result of the information, and the optimal combination of screening rules and weights, screening the sensing information accordingly, calculating the comprehensive score, retaining the high-score information and processing the low-score information, and evaluating the screening result at the same time. If the screening result does not meet the requirements, re-screen after adjusting the rules and weights according to the optimal combination, and send the screening result to the strategy adjustment module; A strategy adjustment module, connected to the screening and evaluation module, for dynamically adjusting the screening rules and weights according to the real-time environmental changes and changes in application requirements, adjusting the screening strategy in a timely manner by monitoring and analyzing the real-time environmental changes and changes in application requirements, and feeding back the adjusted screening rules and weights to the screening and evaluation module.

Images