Multi-index fusion intelligent seat measurement and evaluation method

Abstract

This invention discloses a multi-indicator fusion method for measuring and evaluating smart seats, relating to the field of smart seat measurement and evaluation technology. Specifically, it addresses the limited solutions available for multi-indicator fusion in the background technology for smart seat measurement and evaluation, which requires resolving issues of real-time evaluation and scoring rationality. This invention uses heart rate, respiration, posture, and emotional valence measured non-intrusively by the smart seat as evaluation indicators. Borrowing from the extension matter-element method framework, it constructs a user state matter-element matrix for smart seat measurement, calculates its correlation coefficients, and combines the weights of each indicator to obtain the comprehensive correlation degree. The evaluation level corresponding to the maximum comprehensive correlation degree is the user state evaluation score. This solves the problem that existing smart seat measurement and evaluation methods cannot determine a reasonable and effective evaluation framework and indicator weights during multi-indicator fusion, resulting in comprehensive scores deviating from the user's true state.

Description

Technical Field

[0001] This invention belongs to the field of intelligent seat measurement and evaluation, and specifically relates to a multi-index fusion method for intelligent seat measurement and evaluation. Background Technology

[0002] Currently, office workers primarily engage in mental labor, resulting in limited physical activity and prolonged sitting, making them highly susceptible to sub-health conditions. Real-time assessment of the condition of sedentary individuals through smart chairs can help them pay closer attention to their health and make timely improvements and adjustments. Therefore, developing a reasonable and effective evaluation framework for the measurement and evaluation of smart chairs, which integrates multiple indicators, has become one of the key issues that urgently needs to be addressed in the measurement and evaluation of smart chairs.

[0003] Currently, there are numerous evaluation and weighting methods for multi-objective optimization problems in other fields. Evaluation methods include TOPSIS, Extension Matter-Element Method, and RSR, while weighting methods include Analytic Hierarchy Process (AHP), Entropy Weight Method, and CRITIC Method. However, there are few solutions for the multi-index fusion problem in the measurement and evaluation of smart seats. This problem requires addressing the issues of real-time evaluation and the rationality of scoring. Therefore, inventing a reasonable and effective evaluation framework for optimizing the multi-index fusion problem of smart seat measurement and evaluation is of practical significance.

[0004] Therefore, there is an urgent need to design a multi-indicator fusion method for the measurement and evaluation of smart seats to solve the problem that existing smart seat measurement and evaluation methods cannot determine a reasonable and effective evaluation framework and indicator weights in the process of multi-indicator fusion, thus causing the overall score to deviate from the user's true state. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention aims to provide a multi-index fusion method for measuring and evaluating smart seats. Specifically, the method includes: Step 1. Obtaining evaluation indicators and constructing classical and section domain matrices for each indicator; Step 2. Calculating the correlation between the user's state measured by the smart seat and each level of the evaluation indicators using the classical and section domain matrices; Step 3. Generating a Bayesian network training set, training the Bayesian network using the training set, and obtaining the final evaluation indicator weights based on the trained Bayesian network; Step 4. Substituting the results of Steps 2 and 3 into the comprehensive correlation formula of the extension matter-element method to obtain the comprehensive correlation between the user's state measured by the smart seat and each evaluation level, and determining the maximum comprehensive correlation. This invention solves the problem that existing smart seat measurement and evaluation methods cannot determine a reasonable and effective evaluation framework and indicator weights during multi-index fusion, thus causing the overall score to deviate from the user's true state.

[0006] This invention also provides a multi-index fusion intelligent seat measurement and evaluation system, comprising: a data acquisition module for acquiring evaluation indicators and constructing classical domain and segment domain matrices for each evaluation indicator; a coefficient acquisition module for using the classical domain matrix and segment domain matrix to calculate the correlation between the user state measured by the intelligent seat and each level of the evaluation indicators; a weight acquisition module for generating a Bayesian network training set, using the Bayesian network training set to train the Bayesian network pairs, and obtaining the final evaluation indicator weights based on the trained Bayesian network; and an output module for substituting the results of the coefficient acquisition module and the weight acquisition module into the comprehensive correlation formula of the extension matter-element method to obtain the comprehensive correlation between the user state measured by the intelligent seat to be evaluated and each evaluation level, and determining the maximum comprehensive correlation.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A multi-index fusion method for measuring and evaluating smart seats, comprising:

[0009] Step 1. Obtain the evaluation indicators and construct the classical domain and section domain matrix of each evaluation indicator;

[0010] Step 2. Using the classical domain matrix and the section domain matrix, calculate the correlation between the user status measured by the smart seat and the various levels of the evaluation indicators;

[0011] Step 3. Generate a Bayesian network training set, use the Bayesian network training set to train the Bayesian network pairs, and obtain the final evaluation index weights based on the trained Bayesian network.

[0012] Step 4. Substitute the results of Step 2 and Step 3 into the comprehensive correlation formula of the extension matter-element method to obtain the comprehensive correlation between the user status measured by the smart seat to be evaluated and each evaluation level, and determine the maximum comprehensive correlation.

[0013] Preferably, the classical field matrix formula in step 1 is:

[0014]

[0015] In the formula, j represents the j-th evaluation level, S1, S2, S3, and S4 are the evaluation indicators for heart rate, respiration, sitting posture, and emotional valence, respectively, and v 1j v 2j v 3j v 4j These represent the value range of the evaluation index at the j-th evaluation level, [a] ij b ij ], i = 1, 2, 3, 4 are the upper and lower limits of the range of values ​​for the j-th evaluation level;

[0016] The formula for the section matrix is:

[0017]

[0018] In the formula, P represents the overall evaluation level, and v 1p v 2p v 3p v 4p Let p be the range of values ​​for the evaluation index across all evaluation levels. ip b ip ], i = 1, 2, 3, 4 are the upper and lower limits of the range of values ​​for all evaluation levels p.

[0019] Preferably, the correlation degree calculation process in step 2 includes:

[0020] Process 201. Construct the object-element matrix of the user state to be evaluated for the smart seat measurement using the extension object-element method;

[0021] Process 202. Using the measurement data of the evaluation index in the object-element matrix to be evaluated in Process 201, the value range of each level of the classical domain matrix of the evaluation index in Step 1, and the value range of all evaluation levels of the section domain matrix of the evaluation index, the correlation degree between the user state measured by the smart seat and each level of the evaluation index is calculated by using the correlation coefficient formula of each level of the extension object-element method.

[0022] Preferably,

[0023] The matrix of the objects to be evaluated in process 201 is as follows:

[0024]

[0025] In the formula, P q For the object element to be evaluated, v 1q v 2q v 3q v 4q The measurement data for the indicator to be evaluated;

[0026] The formula for calculating the correlation coefficients between the user status measured by the smart seat and the various levels of the evaluation indicators in process 202 is as follows:

[0027]

[0028] in

[0029] |v ij |=|b ij -a ij |

[0030]

[0031]

[0032] In the formula, k ij (v iq Let ) be the correlation coefficient of the i-th indicator at the j-th level, and v iq For the i-th metric, the actual value of the user's state measured by the smart seat, [a ij b ij ] represents the upper and lower limits of the value range of the i-th indicator at level j, v ip Let [a] be the range of values ​​for all evaluation levels of the i-th indicator. ip b ip ] represents the upper and lower limits of the value range of all evaluation levels for the i-th indicator.

[0033] Preferably, the process of generating the Bayesian network training set in step 3 includes:

[0034] After experts combine and compare the evaluation indicators in pairs, they combine them with fuzzy scaling to obtain multiple fuzzy judgment numbers. They then use the upper and lower bounds of each fuzzy judgment number as the boundaries for Latin hypercube sampling to form multiple fuzzy judgment number subsets. The analytic hierarchy process (AHP) is then used to calculate the weights of multiple preliminary evaluation indicators on the subsets. The subsets and the weights of the multiple preliminary evaluation indicators are used to generate a Bayesian network training set.

[0035] Let l represent the upper bound of the triangular fuzzy number, m represent the median of the triangular fuzzy number, and n represent the lower bound of the triangular fuzzy number. l and n represent the range of relative importance of the two indicators, and m represents the relative importance of the two indicators when compared.

[0036] Preferably, the Bayesian network learning in step 3 includes structure learning and parameter learning;

[0037] The structure learning is specifically as follows: taking the fuzzy judgment number as the cause and the indicator weight as the effect, the node order of the K2 learning algorithm is set to 1 to 10, where nodes 1 to 6 are expert opinion nodes and nodes 7 to 10 are indicator weight nodes.

[0038] The parameter learning process involves first dividing the nodes into intervals using the equal-width method, and then using the maximum likelihood estimation method for parameter learning. The learning sample set for a given node in the Bayesian network training set is Z = {x1, x2, ..., x...}. n The probability distribution of} is p(x) i ;θ), where θ represents the parameter to be estimated, and the search is to find the likelihood function formula The parameter θ is obtained when it is maximized.

[0039] Preferably, the process of obtaining the final evaluation index weights based on the trained Bayesian network is as follows:

[0040] Based on the trained Bayesian network, the maximum probability distribution interval of each indicator weight is obtained, and then the final evaluation indicator weight is obtained by combining the preliminary evaluation indicator weights.

[0041] Preferably, the formula for calculating the comprehensive correlation between the user status measured in step 4 and each evaluation level of the smart seat to be evaluated is as follows:

[0042]

[0043] In the formula, K j (R q ) represents the comprehensive correlation between the user state measured by the smart seat to be evaluated and the j-th evaluation level, and k represents the overall correlation between the user state measured by the smart seat to be evaluated and the j-th evaluation level. ij (v iq Let w be the correlation coefficient of the i-th indicator at the j-th level. i Let be the weight of the i-th indicator.

[0044] Preferably, the evaluation level corresponding to the maximum comprehensive correlation degree in step 4 is the evaluation score of the user status.

[0045] A multi-index fusion intelligent seat measurement and evaluation system, comprising:

[0046] The data acquisition module is used to acquire evaluation indicators and construct the classical domain and section domain matrices of each evaluation indicator.

[0047] The coefficient acquisition module is used to calculate the correlation between the user status measured by the smart seat and the various levels of the evaluation index using the classical domain matrix and the section domain matrix.

[0048] The weight acquisition module is used to generate a Bayesian network training set, train the Bayesian network pairs using the Bayesian network training set, and obtain the final evaluation index weights based on the trained Bayesian network.

[0049] The output module is used to substitute the results of the coefficient acquisition module and the weight acquisition module into the comprehensive correlation formula of the extension matter-element method to obtain the comprehensive correlation between the user status measured by the smart seat to be evaluated and each evaluation level, and to determine the maximum comprehensive correlation.

[0050] The beneficial effects of this invention are as follows: This invention discloses a multi-index fusion method for measuring and evaluating smart seats. Compared with the prior art, the improvement of this invention lies in the following: This invention uses heart rate, respiration, posture, and emotional valence measured by the smart seat without physical contact as evaluation indicators. Drawing on the extension matter-element method framework, it constructs a user state matter-element matrix for smart seat measurement, calculates its correlation coefficient, and combines the weights of each indicator to obtain the comprehensive correlation degree. The evaluation level corresponding to the maximum comprehensive correlation degree is the evaluation score of the user state. In terms of indicator weight calculation, it improves the fuzzy hierarchical analysis method based on Bayesian networks, transforming the acquisition of subjective weights into a maximum distribution probability interval sampling problem. This improves the weight bias caused by individual expert preferences, increases the consistency of expert opinions, and solves the problem that existing smart seat measurement and evaluation methods cannot determine a reasonable and effective evaluation framework and indicator weights during multi-index fusion, thus causing the comprehensive score to deviate from the user's true state. Attached Figure Description

[0051] Figure 1 This is a flowchart of the present invention; Detailed Implementation

[0052] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0053] Example:

[0054] See attached document Figure 1 As shown, a multi-index fusion method for measuring and evaluating smart seats includes:

[0055] Step 1. After the user sits in the smart chair, the user's heart rate, breathing, posture, and emotional valence are measured non-intrusively through the smart chair to obtain measurement data for these four evaluation indicators; and the classical domain and segmental domain matrices of each evaluation indicator are constructed.

[0056] Construct the classical domain and section domain matrices for each evaluation indicator in step 1; the classical domain and section domain matrices contain the value range of the evaluation indicators. The classical domain matrix of the evaluation indicators is composed of the value range of the evaluation indicators corresponding to each evaluation level, and the section domain matrix is ​​composed of the value range of the indicators corresponding to all evaluation levels.

[0057] The classical field matrix is ​​shown in the following equation:

[0058]

[0059] In the formula, j represents the j-th level of evaluation, S1, S2, S3, and s4 are the four evaluation indicators: heart rate, respiration, sitting posture, and emotional valence, respectively, and v 1j v 2j v 3j v4j These represent the value ranges of the four evaluation indicators at the j-th evaluation level, [a ij b ij ], i = 1, 2, 3, 4 are the upper and lower limits of the range of values ​​for the j-th evaluation level;

[0060] The section matrix is ​​shown in the following formula;

[0061]

[0062] In the formula, P represents the overall evaluation level; v 1p v 2p v 3p v 4p Let p be the range of values ​​for the evaluation index across all evaluation levels. ip b ip ], i = 1, 2, 3, 4 are the upper and lower limits of the range of values ​​for all evaluation levels p;

[0063] Step 2. Using the range of values ​​for each level of the classical domain matrix of the evaluation index in Step 1 and the range of values ​​for all evaluation levels of the section domain matrix of the evaluation index, calculate the correlation coefficient between the user state measured by the smart seat and each level of the evaluation index.

[0064] Process 201. Construct the object-element matrix of the user state to be evaluated for the smart seat measurement using the extension object-element method;

[0065] The matter-element matrix to be evaluated is:

[0066]

[0067] In the formula, P q For the object element to be evaluated; v 1q v 2q v 3q v 4q The measurement data for the indicator to be evaluated;

[0068] Procedure 202. Calculate the correlation coefficient for each level, and the distance P(v) between the measurement data of each evaluation index in Procedure 201 and its classical domain interval. iq v ij Therefore, by using the measurement data of the evaluation index in the object-element matrix to be evaluated in process 201 and the value range of each level of the classical domain matrix of the evaluation index in step 1 and the value range of all evaluation levels of the section domain matrix of the evaluation index, the correlation degree between the user state measured by the smart seat and each level of the evaluation index is calculated by the correlation coefficient formula of each level of the extension object-element method.

[0069] The formula for calculating the correlation coefficient between the user status measured by the smart seat and the various levels of evaluation indicators is as follows:

[0070]

[0071] in

[0072] |v ij |=|b ij -a ij |

[0073]

[0074]

[0075] In the formula, k ij (v iq Let ) be the correlation coefficient of the i-th indicator at the j-th level, and v iq For the i-th metric, the actual value of the user's state measured by the smart seat, [a ij b ij ] represents the upper and lower limits of the value range of the i-th indicator at level j, v ip Let [a] be the range of values ​​for all evaluation levels of the i-th indicator. ip b ip ] represents the upper and lower limits of the value range of all evaluation levels for the i-th indicator;

[0076] Step 3. Generate a Bayesian network training set, then use the Bayesian network training set to train the Bayesian network, and obtain the final evaluation index weights based on the trained Bayesian network.

[0077] The process of generating a Bayesian network training set includes:

[0078] The Bayesian network training set includes expert opinion data and subjective weight data. The expert opinion data is generated by experts pairwise combining and comparing all evaluation indicators, and then combining fuzzy scaling (as shown in Table 1) to obtain multiple fuzzy judgment numbers. Sampling is performed using the upper and lower bounds of each fuzzy judgment number as the boundaries of Latin hypercube sampling, thus forming multiple subsets of fuzzy judgment numbers. The subjective weight data is obtained by calculating multiple preliminary evaluation indicator weights using the analytic hierarchy process (AHP) on the subsets. The subsets and the multiple preliminary evaluation indicator weights generate the Bayesian network training set.

[0079] The fuzzy judgment matrix is ​​matrix A, and the formula for matrix A is:

[0080]

[0081] In the formula, n represents the number of indicators; Let l represent the upper bound of the triangular fuzzy number, m represent the median of the triangular fuzzy number, n represent the lower bound of the triangular fuzzy number, l and n represent the range of relative importance of the two indicators, and m represent the relative importance of the two indicators.

[0082] Table 1. Fuzzy Scales and Their Meanings

[0083]

[0084] The Bayesian network is trained using the Bayesian network training set;

[0085] Bayesian network learning is divided into structure learning and parameter learning. For structure learning, the K2 learning algorithm is selected, with the known fuzzy judgment number as the cause and the index weight as the effect. The node order of the K2 learning algorithm is set to 1 to 10, where nodes 1 to 6 are expert opinion nodes, corresponding to the 6 fuzzy judgment numbers generated by pairwise comparison of the 4 evaluation indicators; nodes 7 to 10 are index weight nodes, corresponding to the weights of the 4 evaluation indicators in step 1. Given the Bayesian network structure, the equal width method is selected to divide the nodes into intervals. Finally, the maximum likelihood estimation method is selected for parameter learning of the Bayesian network.

[0086] The maximum likelihood estimation method works as follows: The training sample set X = {x1, x2, ..., x...} of a node in the Bayesian network training set... n The probability distribution of} is p(x) i ;θ), where θ represents the parameter to be estimated, and the search is to find the likelihood function formula The parameter θ is obtained when it is maximized, where i = 1, 2, ..., n, and n is the number of learning samples of a certain node in the training set of the Bayesian network.

[0087] Based on the trained Bayesian network, the maximum probability distribution interval of each indicator weight is obtained, and then the final evaluation indicator weight is obtained by combining the preliminary evaluation indicator weights.

[0088] Specifically, starting from the cause nodes (6 fuzzy judgment numbers) of the trained Bayesian network, samples are taken sequentially according to the interval with the highest distribution probability to obtain the posterior distribution of each result node (4 evaluation index weights). Based on the maximum probability distribution interval of each index weight, the preliminary evaluation index weights within this interval are extracted to form a weight set. The final evaluation index weights are obtained by calculating the average value of each index weight in the weight set and normalizing it.

[0089] Step 4. Substitute the results of Step 2 and Step 3 into the comprehensive correlation formula in the extension matter-element method to obtain the comprehensive correlation between the user status measured by the smart seat to be evaluated and each evaluation level, and determine the maximum comprehensive correlation. The corresponding evaluation level is the evaluation score of the user status.

[0090] The formula for the comprehensive correlation between the user status measured by the smart seat to be evaluated and each evaluation level is as follows:

[0091]

[0092] Where: K j (R q ) represents the comprehensive correlation between the user state measured by the smart seat to be evaluated and the j-th evaluation level, and k represents the overall correlation between the user state measured by the smart seat to be evaluated and the j-th evaluation level. ij (v iq Let w be the correlation coefficient of the i-th indicator at the j-th level. i Let be the weight of the i-th indicator;

[0093] The formula for determining the maximum overall correlation is:

[0094] K max =max{K j (R q )}

[0095] In the formula, K max To achieve the maximum overall correlation.

[0096] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A multi-index fusion method for measuring and evaluating intelligent seats, characterized in that: include: Step 1. Obtain the evaluation indicators and construct the classical domain and section domain matrix of each evaluation indicator; Step 2. Using the classical domain matrix and the section domain matrix, calculate the correlation between the user status measured by the smart seat and the various levels of the evaluation indicators; Step 3. Generate a Bayesian network training set, use the Bayesian network training set to train the Bayesian network, and obtain the final evaluation index weights based on the trained Bayesian network. Step 4. Substitute the results of Step 2 and Step 3 into the comprehensive correlation formula of the extension matter-element method to obtain the comprehensive correlation between the user status measured by the smart seat to be evaluated and each evaluation level, and determine the maximum comprehensive correlation. Step 3, generating the Bayesian network training set, includes: After experts combine and compare the evaluation indicators in pairs, they combine them with fuzzy scaling to obtain multiple fuzzy judgment numbers. They then use the upper and lower bounds of each fuzzy judgment number as the boundaries for Latin hypercube sampling to form multiple fuzzy judgment number subsets. The analytic hierarchy process (AHP) is then used to calculate the weights of multiple preliminary evaluation indicators on the subsets. The subsets and the weights of the multiple preliminary evaluation indicators are used to generate a Bayesian network training set. The fuzzy judgment matrix is ​​matrix A, and the formula for matrix A is: In the formula, n represents the number of indicators; Let l represent the upper bound of the triangular fuzzy number, m represent the median of the triangular fuzzy number, n represent the lower bound of the triangular fuzzy number, l and n represent the range of relative importance of the two indicators, and m represent the relative importance of the two indicators. Step 3 of the Bayesian network learning includes structure learning and parameter learning; The structure learning is specifically as follows: taking the fuzzy judgment number as the cause and the indicator weight as the effect, the node order of the K2 learning algorithm is set to 1 to 10, where nodes 1 to 6 are expert opinion nodes and nodes 7 to 10 are indicator weight nodes. Parameter learning specifically involves: given the Bayesian network structure, firstly, using the equal-width method to partition the nodes into intervals, then using the maximum likelihood estimation method for parameter learning. The learning sample set X = {x1, x2, ..., x...} for a given node in the Bayesian network training set is... n The probability distribution of} is p(x) i ;θ), where θ represents the parameter to be estimated, and the search is to find the likelihood function formula The parameter θ is obtained when it is maximized, where i = 1, 2, ..., n, and n is the number of learning samples of a certain node in the training set of the Bayesian network; The process of obtaining the final evaluation index weights based on the trained Bayesian network is as follows: Based on the trained Bayesian network, the maximum probability distribution interval of each indicator weight is obtained, and then the final evaluation indicator weight is obtained by combining the preliminary evaluation indicator weights.

2. The method for measuring and evaluating a smart seat using multi-index fusion as described in claim 1, characterized in that: The classical field matrix formula for step 1 is: In the formula, j represents the j-th evaluation level, S1, S2, S3, and S4 are the evaluation indicators for heart rate, respiration, sitting posture, and emotional valence, respectively, and v 1j v 2j v 3j v 4j These represent the value range of the evaluation index at the j-th evaluation level, [a] ij ,b ij ], i = 1, 2, 3, 4 are the upper and lower limits of the range of values ​​for the j-th evaluation level; The formula for the section matrix is: In the formula, P represents the overall evaluation level, and v 1p v 2p v 3p v 4p Let p be the range of values ​​for the evaluation index across all evaluation levels. ip ,b ip ], i = 1, 2, 3, 4 are the upper and lower limits of the range of values ​​for all evaluation levels p.

3. The method for measuring and evaluating a smart seat using multi-index fusion as described in claim 1, characterized in that: Step 2, the correlation calculation process, includes: Process 201. Construct the object-element matrix of the user state to be evaluated for the smart seat measurement using the extension object-element method; Process 202. Using the measurement data of the evaluation index in the object-element matrix to be evaluated in Process 201, the value range of each level of the classical domain matrix of the evaluation index in Step 1, and the value range of all evaluation levels of the section domain matrix of the evaluation index, the correlation degree between the user state measured by the smart seat and each level of the evaluation index is calculated by using the correlation coefficient formula of each level of the extension object-element method.

4. The method for measuring and evaluating a smart seat using multi-index fusion as described in claim 3, characterized in that: The matrix of the objects to be evaluated in process 201 is as follows: In the formula, P q For the object element to be evaluated, v 1q v 2q v 3q v 4q The measurement data for the indicator to be evaluated; The formula for calculating the correlation coefficients between the user status measured by the smart seat and the various levels of the evaluation indicators in process 202 is as follows: in |v ij |=|b ij -a ij | In the formula, k ij (v iq Let ) be the correlation coefficient of the i-th indicator at the j-th level, and v iq For the i-th metric, the actual value of the user's state measured by the smart seat, [a ij b ij ] represents the upper and lower limits of the value range of the i-th indicator at level j, v ip Let [a] be the range of values ​​for all evaluation levels of the i-th indicator. ip b ip ] represents the upper and lower limits of the value range of all evaluation levels for the i-th indicator.

5. The method for measuring and evaluating a smart seat using multi-index fusion as described in claim 1, characterized in that: The formula for calculating the comprehensive correlation between the user status measured in step 4 and each evaluation level of the smart seat to be evaluated is as follows: In the formula, K j (R q ) represents the comprehensive correlation between the user state measured by the smart seat to be evaluated and the j-th evaluation level, and k represents the overall correlation between the user state measured by the smart seat to be evaluated and the j-th evaluation level. ij (v iq Let w be the correlation coefficient of the i-th indicator at the j-th level. i Let be the weight of the i-th indicator.

6. The method for measuring and evaluating a smart seat using multi-index fusion as described in claim 1, characterized in that: The rating level corresponding to the maximum comprehensive correlation in step 4 is the rating score of the user status.

7. A multi-index fusion intelligent seat measurement and evaluation system, characterized in that: include: The data acquisition module is used to acquire evaluation indicators and construct the classical domain and section domain matrices of each evaluation indicator. The coefficient acquisition module is used to calculate the correlation between the user status measured by the smart seat and the various levels of the evaluation index using the classical domain matrix and the section domain matrix. The weight acquisition module is used to generate a Bayesian network training set, train the Bayesian network pairs using the Bayesian network training set, and obtain the final evaluation index weights based on the trained Bayesian network. The output module is used to substitute the results of the coefficient acquisition module and the weight acquisition module into the comprehensive correlation formula of the extension matter-element method to obtain the comprehensive correlation between the user status measured by the smart seat to be evaluated and each evaluation level, and to determine the maximum comprehensive correlation.

Images