A Situation Awareness Multi-Index Quality Management System and Method Based on Data Fusion

Abstract

The present invention discloses a scenario-aware multi-index quality management system based on data fusion, which includes a quality management module for acquisition devices, a quality management module for scenario information, a quality management module for services, a quality management module for user experience, and a quality management module for driving decision-making; the quality management module for acquisition devices, the quality management module for scenario information, the quality management module for services, and the quality management module for driving decision-making form a closed-loop connection, and the quality management module for services, the quality management module for driving decision-making, and the quality management module for user experience are interconnected. The present invention adopts a scenario-aware multi-index quality management system and method based on data fusion, realizes the adaptive adjustment and interaction of multiple system quality indicators, comprehensively measures the quality level of the scenario-aware system, and utilizes the mutual restriction relationship between different indicators to realize the adaptive adjustment and optimization of the system under the condition of lacking real user feedback.

Description

Technical Field

[0001] The present invention relates to the field of context awareness technology, and in particular to a context awareness multi-index quality management system and method based on data fusion. Background Art

[0002] As one of the core applications of ubiquitous computing, the context-aware system emphasizes transparent services, can actively perceive the relevant context information of the service entity in real time, and provide users with appropriate services to meet user or system needs after integration, reasoning and decision-making. Context-aware technology has been widely used in scenarios such as smart homes, smart healthcare and smart cities, realizing the transformation from perceptual intelligence to cognitive intelligence, greatly improving the user experience.

[0003] With the development of context-aware technology, different context-aware system frameworks have become richer, and it has become an inevitable trend to establish a quality assessment system with strong universality for different systems. Existing researchers have proposed different system quality indicators for context-aware systems, such as quality of device (QoD), quality of context (QoC), quality of service (QoS) and quality of experience (QoE). Various quality indicators measure system quality from different aspects, such as collection equipment to user experience. Although they have different focuses, they are closely related to each other. The integration of different categories of quality indicators to obtain comprehensive quality indicators and apply them to context-aware systems has achieved good results.

[0004] However, as the system becomes more and more intelligent, more and more devices are connected, and the system requires less and less active user intervention, so the system quality management method that is biased towards the user side is no longer completely applicable. The user of a system may also be the next-level module or system, so it is necessary to consider the feedback and association between systems and establish quality indicators for system drive and decision-making. Secondly, the quality indicators of the situational awareness system have not yet fully exerted their own role and the role of mutual cooperation between the indicators. There is a lack of a unified management mechanism, and it is impossible to fully exert the role of each quality indicator other than evaluation. It is necessary to establish interactive management of multiple indicators from information acquisition to system drive execution so that the system can be adaptively optimized. Therefore, how to further establish the interaction between the various quality management indicators from the perception end to the decision-making end and design a multi-indicator situational awareness system quality management system is a research problem that needs to be solved urgently. Summary of the invention

[0005] The object of the present invention is to provide a scenario-aware multi-index quality management system based on data fusion, which realizes the adaptive adjustment and interaction of multiple system quality indexes, comprehensively measures the quality level of the scenario-aware system through the interactive management of multiple quality indexes, and at the same time utilizes the mutual restriction relationship between different indexes to realize the adaptive adjustment and optimization of the system under the condition of lack of real user feedback. Another object of the present invention is to provide a working method of a scenario-aware multi-index quality management system based on data fusion.

[0006] To achieve the above object, the present invention provides a scenario-aware multi-index quality management system based on data fusion, including a collection device quality management module, a scenario information quality management module, a service quality management module, a user experience quality management module, and a driving decision quality management module;

[0007] The collection device quality management module, the scenario information quality management module, the service quality management module, and the driving decision quality management module form a closed-loop connection, the scenario information quality management module is connected to the user experience quality management module, and the service quality management module and the driving decision quality management module are respectively connected to the user experience quality management module in an interactive manner;

[0008] The service quality management module includes a knowledge graph-based service recommendation unit, a subjective QoS acquisition unit, an objective QoS acquisition unit, a QoS fusion sorting and optimization unit, and a subjective and objective threshold screening unit connected in sequence. The subjective and objective threshold screening unit includes a subjective QoS threshold and an objective QoS threshold;

[0009] The user experience quality management module includes a user and system feedback unit, a subjective QoE acquisition unit, and an objective QoE acquisition unit.

[0010] Preferably, the collection device quality management module includes a sensing device information configuration unit, a QoD parameter acquisition unit, and a low-level scenario information acquisition unit connected in sequence;

[0011] The sensing device information configuration unit is used to receive feedback information in real time and adjust the information configuration of the sensing device;

[0012] The QoD parameter acquisition unit is used to acquire the QoD parameters of physical sensors;

[0013] The low-level scenario information acquisition unit is used to collect data using intelligent devices that meet the system requirements, avoid collecting useless scenario information, intuitively quantify the matching degree of each intelligent sensing device with the current processing requirements, save storage space and computing resources, and transmit the acquired low-level scenario information to the scenario information quality management module for further processing.

[0014] Preferably, the scenario information quality management module includes a scenario information preprocessing unit, a scenario information repository unit, a QoC parameter weighted fusion unit, a QoC-based scenario information uncertainty elimination unit, and an advanced scenario information acquisition unit that are connected in sequence;

[0015] The scenario information preprocessing unit is used to standardize homogeneous scenario information, classify heterogeneous scenario information, and transmit the original scenario information and the preprocessed scenario information to the scenario information repository unit;

[0016] The scenario information repository unit is used to store the original and preprocessed scenario information;

[0017] The QoC parameter weighted fusion unit is used to calculate each type of QoC parameter value of the scenario information, and set corresponding parameter weights according to system and user requirements for weighted fusion;

[0018] The QoC-based scenario information uncertainty elimination unit is used to handle the incompleteness, imprecision, and inconsistency of the scenario information;

[0019] The advanced scenario information acquisition unit is used to model, reason, and fuse the low-level scenario information after uncertainty elimination to obtain advanced scenario information, providing necessary information support for service recommendation.

[0020] Preferably, the knowledge graph-based service recommendation unit is used to send a request for the required service according to the advanced scenario information;

[0021] The subjective QoS acquisition unit is used to determine subjective QoS parameters and corresponding weights based on user expectations and preferences statistics. For QoS parameters related to user preferences, the weights are determined according to user preferences, and for QoS parameters not related to user preferences, they are processed according to the default weights. The subjective QoS parameters are optional parameters;

[0022] The objective QoS acquisition unit is used to obtain objective QoS parameters that meet the required service and corresponding weights. The objective QoS parameters are mandatory parameters;

[0023] The QoS fusion sorting and optimization unit is used to classify and fuse the subjective and objective QoS weights to obtain subjective and objective QoS values respectively, and sort them in descending order;

[0024] The subjective and objective threshold screening unit is used to calculate the subjective and objective QoS thresholds to screen the services recommended by the system. The subjective QoS threshold is determined according to user subjective preferences. When user preference feedback information cannot be obtained, the subjective QoS threshold is the same as the objective QoS threshold.

[0025] Preferably, the driving decision quality management module includes a decision optimization unit, a driving control unit, and a decision adaptive perception unit. The decision adaptive perception unit includes a subjective decision perception subunit, an objective decision perception subunit, and a quality index adaptive adjustment subunit;

[0026] The decision optimization unit and the driving control unit are connected in sequence. Both the subjective decision perception subunit and the objective decision perception subunit are connected to the quality index adaptive adjustment subunit;

[0027] The decision optimization unit is used to judge the degree to which the service meets the expectations according to the actuator information and the environment information, and further optimize the service after screening by the subjective QoS threshold and the objective QoS threshold to adapt to the operation requirements of the actual actuator;

[0028] The driving control unit is used to directly control the actuators serving the environment, the system, and the customers, avoid unnecessary responses or actions of the physical actuators directly serving the users and the environment, and perform stable control and management on the physical or virtual terminals according to the self-adjustable fault detection and calibration mechanism, and provide the information reflecting the decision execution level to the objective QoE acquisition unit to increase the feedback information when there is no real user feedback;

[0029] The decision adaptive perception unit is used to automatically perceive the execution situation of the current system decision;

[0030] The subjective decision perception subunit is used to count the decision execution situations from user feedback and correction and the decision execution situations of the cascaded system feedback. The subjective decision perception subunit is an optional item;

[0031] The objective decision perception subunit is used to actively perceive whether the impact of the decision execution on the entity meets the expectations. The objective decision perception subunit is a mandatory item;

[0032] The quality index adaptive adjustment subunit is used to perform adaptive adjustment operations on the system according to the interaction performance of the quality indexes at all levels of the system and the decision execution situation, send the QoD and QoC information that needs to be adjusted to the perception device information configuration unit, and the corresponding information included will be extracted layer by layer in different modules and used to adjust the quality indexes of each module.

[0033] Preferably, the user and system feedback unit is connected to the subjective QoE acquisition unit;

[0034] The subjective QoE acquisition unit is used to capture the QoE actively provided by the user, send the subjective QoE information to the subjective QoS acquisition unit, and jointly assist in determining the subjective QoS parameters and the corresponding weights;

[0035] The objective QoE acquisition unit is used to predict the objective QoE by fusing the comprehensive QoC parameters, QoS parameters, and decision execution status, and send the result to the objective QoS acquisition unit for quality metric adaptive adjustment and service optimization.

[0036] The working method of the above-mentioned scenario-aware multi-metric quality management system based on data fusion includes the following steps:

[0037] Step S01: Configure the sensing device information, and preset and adjust the relevant configuration information of the intelligent sensing device according to the specific application requirements of the scenario-aware system;

[0038] Step S02: Obtain the QoD parameters, and obtain the QoD parameters of the intelligent sensing device in the applicable state in the current scenario;

[0039] Step S03: Obtain the low-level scenario information, and select the intelligent sensing device whose QoD parameters meet the system requirements to collect the scenario information;

[0040] Step S04: Preprocess the scenario information, classify and preprocess the collected scenario information. For homogeneous scenario information, the data format is unified into a preset standard format; for heterogeneous data, extract the effective information and recombine it into a unified data format and allow default values. For simple data and scenarios, it is usually preprocessed into boolean data;

[0041] Step S05: Store the scenario information, store the collected original scenario information, the preprocessed scenario information, and the system preset parameter threshold information in the database;

[0042] Step S06: Obtain the QoC parameters, and calculate the quality of the obtained scenario information, including reliability, update degree, correctness, and scenario information correlation parameters;

[0043] Step S07: Weighted fusion of QoC parameters, analyze the nature of the information source of the QoC parameters, match the appropriate QoC parameters, and perform intelligent fusion on the QoC parameters to obtain the comprehensive quality of the scenario information;

[0044] Step S08: Eliminate the uncertainty of the scenario information based on QoC, and perform inconsistency, incompleteness, and imprecision elimination processing on the scenario information with the weighted fusion QoC parameters;

[0045] Step S09: Obtain the high-level scenario information, further process the scenario information after uncertainty elimination, and fuse the high-quality scenario information through an inference fusion method. The inference fusion method adopts the D-S evidence theory and the Bayesian method;

[0046] Step S10: Knowledge graph service recommendation. According to the obtained high-level scenario information, use the recommendation system algorithm based on the knowledge graph to send a request for the currently required service;

[0047] Step S11: Obtaining subjective QoS. Obtain subjective QoS based on user preferences and subjective QoE. For QoS parameters related to user preferences, determine the QoS parameter weights according to the user. For QoS parameters not related to user preferences, process them according to the default weights;

[0048] Step S12: Obtaining objective QoS. Calculate the QoS parameter values through the current service and objective QoE;

[0049] Step S13: QoS fusion, ranking, and selecting the best. Assign different weights to subjective and objective QoS to obtain the comprehensive QoS score and rank the eligible services; The subjective weight is determined according to user preferences and the fuzzy AHP method, and the objective weight is determined using the information entropy method. When the subjective weight is λ, the objective weight is 1 - λ, the subjective QoS value is W O , the objective QoS value is W S , the comprehensive QoS value is λ×W O +(1 - λ)×W S ;

[0050] Step S14: Obtaining subjective and objective QoS thresholds. Set the QoS thresholds according to user preferences and QoE. The subjective threshold is set according to the service level required by the user. Select the services that meet the user's requirements and save computing and storage resources from the services sorted in step S13. The objective threshold is set according to the general threshold preset by the system. When the subjective QoS threshold cannot be obtained, make it consistent with the objective QoS threshold;

[0051] Step S15: Determine whether the QoS exceeds the threshold. Compare the QoS of the obtained service with the threshold set in step S14. If it does not exceed the threshold, select the service with the highest score within the threshold. Otherwise, even if there is a service with a higher score outside the threshold, it will not be considered, and re-enter step S10;

[0052] Step S16: Obtaining service availability and consistency. According to the service provided in step S15, combine with the drive control unit to obtain the availability and consistency of the service, further optimize the service, enhance the reliability and effectiveness of the service, and improve the system decision-making accuracy;

[0053] Step S17: Drive control. Based on the self-adjustable fault detection and calibration mechanism, perform stable control and management on the physical or virtual terminal, and provide the information reflecting the actuator execution decision to the objective QoE acquisition unit;

[0054] Step S18: Obtaining user and system feedback, obtaining the true feedback of the user, which is fundamentally different from the predicted objective QoE here, and at the same time obtaining the feedback information from the cascaded system and storing it;

[0055] Step S19: Judging whether the feedback is from the user, judging whether the feedback in step S18 is from a real user. If it is from a real user, go to step S21; otherwise, go to step S22;

[0056] Step S20: Obtaining objective QoE, calculating the objective QoE by combining QoC parameters, drive control feedback data, and the QoS parameter values of the current service, and feeding the result back to the objective QoS obtaining unit in step S12;

[0057] Step S21: Obtaining subjective QoE, capturing the QoE actively provided by the user, and feeding the result back to the subjective QoS obtaining unit in step S11;

[0058] Step S22: Obtaining QoME, calculating the quality of machine experience from the system rather than the user;

[0059] Step S23: Subjective decision perception, obtaining the degree to which the current actuator's execution decision meets the expectation according to the QoME of the cascaded system;

[0060] Step S24: Objective decision perception, adaptively judging the degree to which the actuator's execution decision meets the expectation according to the execution unit and feeding it back to the quality index adaptive adjustment unit;

[0061] Step S25: Adaptive adjustment of quality indicators, performing adaptive adjustment operations on the system according to the interaction performance of the quality indicators at each level of the system and the decision execution situation.

[0062] The advantages and positive effects of the scenario-aware multi-index quality management system and method based on data fusion described in the present invention are:

[0063] 1. It can actively sense scenario information, monitor the changes in the environmental state in real time, and actively provide appropriate services for users through the processing of scenario information.

[0064] 2. It can continuously adjust and optimize the system performance by utilizing the relationship between each management module, and can realize the adaptive adjustment of the configuration of application programs or sensing devices.

[0065] 3. Using multiple indicators such as QoD, QoC, QoS, QoAD, and QoE to manage the quality of the scenario-aware system, reducing computing resources while improving the system's fault tolerance rate and the correctness of decisions, making the system efficient.

[0066] 4. For application scenarios lacking user feedback, a driving decision quality management module is added. On the one hand, it can further optimize system services, and on the other hand, it can strengthen the feedback to the next-level system, so that when there is no real user feedback or little feedback, a high system quality can still be maintained.

[0067] 5. Calculate QoS from the perspectives of users and the system. By setting an adaptive QoS threshold, the services provided by the system can be more reasonable, effective, and consume less resources while ensuring that user needs are met.

[0068] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0069] Figure 1 It is a structural framework diagram of a scenario-aware multi-index quality management system based on data fusion according to the present invention;

[0070] Figure 2 It is a composition and connection relationship diagram of a scenario-aware multi-index quality management system based on data fusion according to the present invention;

[0071] Figure 3 It is a schematic diagram of the working process of a scenario-aware multi-index quality management system based on data fusion according to the present invention. Detailed Embodiments

[0072] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0073] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the field to which the present invention belongs.

[0074] Embodiment 1

[0075] A scenario-aware multi-index quality management system based on data fusion, as Figure 1 shown, includes a collection device quality management module, a scenario information quality management module, a service quality management module, a user experience quality management module, and a driving decision quality management module.

[0076] The collection device quality management module, the scenario information quality management module, the service quality management module, and the driving decision quality management module form a closed-loop connection. The scenario information quality management module is connected to the user experience quality management module, and the service quality management module and the driving decision quality management module are respectively connected to the user experience quality management module in an interactive manner.

[0077] The acquisition device quality management module is used to manage and adjust the information configuration of the sensing devices, obtain the relevant QoD parameters of the sensing devices, select the compliant sensing devices according to the feedback information from the driving decision quality management module, and send the acquired low-level scenario information to the scenario information quality management module. The QoD parameters refer to the parameters describing the quality of the acquisition devices, and the low-level scenario information refers to the simple scenario information directly from sensors or user presets.

[0078] The scenario information quality management module is used to calculate the QoC parameters and eliminate the uncertainty of the scenario information accordingly, obtain the high-level scenario information through fusion reasoning, and send the high-level scenario information to the service quality management module for service recommendation. The QoC parameters refer to the parameters used to describe the quality of the scenario information, mainly including update degree, reliability, correctness, and integrity, etc. The uncertainty of the scenario information refers to the imprecision, inconsistency, and incompleteness between the information, and the high-level scenario information refers to the information that can reflect the entity state obtained by fusing and reasoning the low-level scenario information.

[0079] The service quality management module is used to make intelligent recommendations for relevant services according to the high-level scenario information obtained by fusion reasoning. After calculating the QoS parameters of the recommended services or service combinations, it filters the appropriate services according to the QoS parameter values and pushes them to the driving decision quality management module at the same time. The QoS parameter values refer to the parameters describing the performance of the system services, mainly including response time, packet loss rate, and cost, etc.

[0080] The driving decision quality management module is used to further optimize the services recommended by the system, directly act on the physical or virtual actuators, and sense the correctness and compliance with expectations after the actuators execute the decisions. It adaptively updates and adjusts the quality indicators according to the user and system feedback and sends them to the acquisition device quality management module. The physical actuator refers to the device that can drive through electricity, magnetism, etc. and can cause actual changes to objects or the environment, and the virtual actuator refers to network services, etc.

[0081] The user experience quality management module is used to obtain information such as the feedback from users and the system and the user preference statistics, evaluate and calculate the QoE indicator values and send them to the service quality management module and the driving decision quality management module.

[0082] The acquisition device quality management module includes a sensing device information configuration unit, a QoD parameter acquisition unit, and a low-level scenario information acquisition unit connected in sequence.

[0083] The sensing device information configuration unit is used to receive the feedback information in real time and adjust the information configuration of the sensing devices.

[0084] The QoD parameter acquisition unit is used to acquire the QoD parameters of physical sensors, including sensor accuracy, sampling frequency, maximum effective measurement distance, service life, etc., and acquire the mathematical model accuracy of virtual sensors.

[0085] The low-level scenario information acquisition unit is used to collect data using intelligent devices that meet the system requirements, avoid collecting useless scenario information, intuitively quantify the matching degree of each intelligent sensing device with the current processing requirements, save storage space and computing resources, and transmit the acquired low-level scenario information to the scenario information quality management module for further processing.

[0086] The scenario information quality management module includes a scenario information preprocessing unit, a scenario information repository unit, a QoC parameter weighted fusion unit, a QoC-based scenario information uncertainty elimination unit, and a high-level scenario information acquisition unit connected in sequence.

[0087] The scenario information preprocessing unit is used to standardize homogeneous scenario information, classify heterogeneous scenario information, and transmit the original scenario information and the preprocessed scenario information to the scenario information repository unit.

[0088] The QoC parameter weighted fusion unit is used to calculate the QoC parameter values of each category of scenario information, including update degree, reliability, correctness, etc., and set the corresponding parameter weights for weighted fusion according to system and user requirements; for example, when the system has particularly high requirements for real-time performance, the weight of the update degree takes precedence.

[0089] The scenario information repository unit is used to store the original and preprocessed scenario information.

[0090] The QoC-based scenario information uncertainty elimination unit is used to handle the incompleteness, imprecision, and inconsistency of scenario information.

[0091] The high-level scenario information acquisition unit is used to model, reason, and fuse the low-level scenario information after uncertainty elimination to obtain high-level scenario information, providing necessary information support for service recommendation.

[0092] The service quality management module includes a knowledge graph-based service recommendation unit, a subjective QoS acquisition unit, an objective QoS acquisition unit, a QoS fusion sorting and optimization unit, and a subjective and objective threshold screening unit connected in sequence.

[0093] The knowledge graph-based service recommendation unit is used to send requests for the required services according to the high-level scenario information.

[0094] The subjective QoS acquisition unit is used to determine subjective QoS parameters and corresponding weights based on user expectations and preference statistics. For QoS parameters related to user preferences, weights are determined according to user preferences. For QoS parameters not related to user preferences, they are processed according to default weights. Subjective QoS parameters are optional parameters.

[0095] The objective QoS acquisition unit is used to obtain objective QoS parameters that meet the required services and their corresponding weights. Default objective QoS parameters include system response time, latency, and cost, etc. Objective QoS parameters are mandatory parameters.

[0096] The QoS fusion sorting and optimization unit is used to classify and fuse the subjective and objective QoS weights respectively to obtain subjective and objective QoS values, and sort them in descending order. The fusion of subjective QoS and objective QoS can adopt the Analytic Hierarchy Process (AHP).

[0097] The subjective and objective threshold screening unit is used to calculate subjective and objective QoS thresholds to screen the services recommended by the system, including subjective QoS thresholds and objective QoS thresholds. The subjective QoS threshold is determined according to user subjective preferences. Different times and different users will affect the determination of the subjective QoS threshold. When feedback information such as user preferences cannot be obtained, the subjective QoS threshold is consistent with the objective QoS.

[0098] The drive decision-making quality management module includes a decision optimization unit, a drive control unit, and a decision self-adaptive perception unit. The decision self-adaptive perception unit includes a subjective decision perception subunit, an objective decision perception subunit, and a quality index self-adaptive adjustment subunit.

[0099] The decision optimization unit and the drive control unit are connected in sequence. The subjective decision perception subunit and the objective decision perception subunit are both connected to the quality index self-adaptive adjustment subunit.

[0100] The decision optimization unit is used to judge the degree to which the service meets the expectations according to the actuator information and environmental information, and further optimize the service after screening by the subjective QoS threshold and the objective QoS threshold to adapt to the operation requirements of the actual actuator.

[0101] The drive control unit is used to directly control the actuators serving the environment, system, and customers, to avoid unnecessary responses or actions of the physical actuators directly serving users and the environment, and to stably control and manage the physical or virtual terminals based on a self-adjustable fault detection and calibration mechanism, and provide information reflecting the decision execution level to the objective QoE acquisition unit to increase the feedback information when there is no real user feedback.

[0102] The decision self-adaptive perception unit is used to automatically perceive the execution situation of the current system decision.

[0103] The subjective decision perception subunit is used to count the decision execution situations from user feedback and correction, as well as the decision execution situations feedback by the cascaded system. The subjective decision perception subunit is an optional item.

[0104] The objective decision perception subunit is used to actively perceive whether the impact of decision execution on the entity meets the expectation. The objective decision perception subunit is a mandatory item.

[0105] The quality index adaptive adjustment subunit is used to perform adaptive adjustment operations on the system according to the interaction performance of the quality indexes at all levels of the system and the decision execution situation, send the QoD and QoC information that needs to be adjusted to the perception device information configuration unit, and the corresponding information contained will be extracted layer by layer in different modules and used to adjust the quality indexes of each module.

[0106] The user experience quality management module includes a user and system feedback unit, a subjective QoE acquisition unit, and an objective QoE acquisition unit.

[0107] The user and system feedback unit is connected to the subjective QoE acquisition unit.

[0108] The user and system feedback unit is used to record user expectations, preference statistics, etc., and at the same time obtain the Quality of Machine Experience (QoME) of the cascaded system. QoME essentially measures the impact of the machine on the end-user QoE, and the feedback information is provided to the subjective decision perception subunit for evaluating the decision execution level.

[0109] The subjective QoE acquisition unit is used to capture the subjective QoE actively provided by the user, send the subjective QoE information to the subjective QoS acquisition unit, and jointly determine the subjective QoS parameters and the corresponding weights.

[0110] The objective QoE acquisition unit is used to predict the objective QoE by fusing the comprehensive QoC parameters, QoS parameters, and decision execution situation, and send the result to the objective QoS acquisition unit for quality index adaptive adjustment and service optimization.

[0111] Embodiment 2

[0112] A working method of a scenario-aware multi-index quality management system based on data fusion, as Figure 3 shown. Taking the smart home system as an example, the system obtains user location information through Wi-Fi, Bluetooth, sound sensors, and computer activities, and intelligently adjusts the air conditioners and lights in different rooms, etc. The specific steps include:

[0113] Step S01: Perception device information configuration

[0114] The perception device information configuration unit presets and adjusts relevant configuration information for the intelligent perception device according to the specific application requirements of the scenario perception system.

[0115] Step S02: Obtaining QoD parameters

[0116] The QoD parameter obtaining unit obtains the QoD parameters of the intelligent perception device in the applicable state in the current scenario. Two main QoD parameters are obtained: accuracy and sampling interval.

[0117] Step S03: Obtaining low-level scenario information

[0118] The low-level scenario information obtaining unit selects the intelligent perception devices whose QoD parameters meet the system requirements for scenario information collection. The original scenario information collected is all location information.

[0119] Step S04: Preprocessing scenario information

[0120] The scenario information preprocessing unit classifies and preprocesses the collected scenario information. For homogeneous scenario information, the data format is unified into a preset standard format; for heterogeneous data, the effective information is extracted and recombined into a unified data format and default values are allowed; for simple data and scenarios, it is usually preprocessed into boolean data. The data is binarized, being "1" if in the specified location and "0" otherwise.

[0121] Step S05: Storing scenario information

[0122] The scenario information storage unit stores the collected original scenario information, the preprocessed scenario information, and system preset parameter thresholds and other information into the database.

[0123] Step S06: Obtaining QoC parameters

[0124] The QoC parameter obtaining unit calculates the quality of the obtained scenario information, including parameters such as reliability, update degree, correctness, and scenario information relevance. Three parameters, reliability, update degree, and correctness, are adopted.

[0125] Step S07: Weighted fusion of QoC parameters

[0126] The QoC parameter weighted fusion unit analyzes the nature of the information source for the QoC parameters, matches appropriate QoC parameters, and performs intelligent fusion on the QoC parameters to obtain the comprehensive quality of scenario information.

[0127] Step S08: Eliminating the uncertainty of scenario information based on QoC

[0128] The QoC-based scenario information uncertainty elimination unit uses the QoC parameter scenario information after weighted fusion to perform inconsistency, incompleteness, and imprecision elimination processing. The D-S evidence theory is used to handle the inconsistency of scenario information, and the probability statistics method is used to handle the incompleteness of scenario information.

[0129] Step S09: Obtaining high-level scenario information

[0130] The high-level scenario information acquisition unit further processes the scenario information after uncertainty elimination, and fuses the high-quality scenario information through an inference fusion method. The inference fusion method adopts the D-S evidence theory and the Bayesian method.

[0131] Step S10: Knowledge graph service recommendation

[0132] The knowledge graph service recommendation unit sends a request for the currently required service (service combination) according to the high-level scenario information, using the recommendation system algorithm based on the knowledge graph.

[0133] Step S11: Obtaining subjective QoS

[0134] The subjective QoS acquisition unit obtains the subjective QoS based on user preferences and subjective QoE. For the QoS parameters related to user preferences, the QoS parameter weights are determined according to the user, and for the QoS parameters not related to user preferences, they are processed according to the default weights. The subjective QoS is affected by the user's manual modification control of the lights and air conditioners.

[0135] Step S12: Obtaining objective QoS

[0136] The objective QoS acquisition unit calculates the QoS parameter values through the current service and objective QoE.

[0137] Step S13: QoS fusion, ranking, and selecting the best

[0138] The QoS fusion, ranking, and selecting the best unit assigns different weights to the subjective and objective QoS, obtains the comprehensive QoS score, and ranks and grades the eligible services (service combinations). The subjective weight is determined according to user preferences and the fuzzy AHP method, and the objective weight is determined using the information entropy method.

[0139] Step S14: Obtaining subjective and objective QoS thresholds

[0140] The subjective and objective QoS threshold acquisition unit sets the QoS thresholds according to user preferences and QoE. The subjective threshold is set according to the service level required by the user. The objective threshold is set according to the general threshold preset by the system. When the range is [0, 1], the general threshold is set to [0.45, 0.85].

[0141] Step S15: Judging whether QoS exceeds the threshold

[0142] Compare the QoS of the obtained service with the QoS threshold set in step S14. If it does not exceed the threshold, select the service with the highest score within the threshold. Otherwise, even if there is a service with a higher score outside the threshold, it will not be considered, and re-enter step S10.

[0143] Step S16: Obtain service availability and consistency

[0144] According to the service provided in step S15, combined with the drive control unit, obtain the availability and consistency of the service, further optimize the service, strengthen the reliability and effectiveness of the service, and improve the system decision accuracy.

[0145] Step S17: Drive control

[0146] The drive control unit stably controls and manages the physical or virtual terminal based on the self-adjustable fault detection and calibration mechanism, and provides the information reflecting the actuator execution decision to the objective QoE obtaining unit. The drive control is manifested as the adjustment of the light switch and the air conditioner temperature.

[0147] Step S18: Obtain user and system feedback

[0148] The user and system feedback obtaining unit obtains the true feedback of the user, which is fundamentally different from the predicted objective QoE here, and at the same time obtains the feedback information from the cascaded system and stores it.

[0149] Step S19: Determine whether the feedback is from the user

[0150] Determine whether the feedback in step S18 is from a real user. If it is from a real user, enter step S21; otherwise, enter step S22.

[0151] Step S20: Obtain objective QoE

[0152] The objective QoE obtaining unit calculates the objective QoE by combining the QoC parameters, the drive control feedback data, and the QoS parameter values of the current service, and feeds the result back to the objective QoS obtaining unit in step S12.

[0153] Step S21: Obtain subjective QoE

[0154] The subjective QoE obtaining unit captures the QoE actively provided by the user, and feeds the result back to the subjective QoS obtaining unit in step S11.

[0155] Step S22: Obtain QoME

[0156] The QoME obtaining unit calculates the machine experience quality from the system rather than the user.

[0157] Step S23: Subjective decision perception

[0158] The subjective decision perception unit obtains the degree to which the current actuator's execution decision meets the expectation according to the QoME of the cascaded system.

[0159] Step S24: Objective decision perception

[0160] The objective decision perception unit adaptively judges the degree to which the current actuator's execution decision meets the expectation according to the execution unit and feeds it back to the quality index adaptive adjustment unit.

[0161] Step S25: Quality index adaptive adjustment

[0162] The quality index adaptive adjustment subunit adaptively adjusts the system according to the interaction performance of each quality index of the system and the decision execution situation. It sends information such as QoD, QoS, and QoC that need to be adjusted to the sensing device information configuration unit, and the corresponding information contained will be extracted layer by layer in different modules and used to adjust the quality indexes of each module. The adjustment of the QoD index is carried out in step S01, the adjustment of the QoS index is carried out in step S12, and the adjustment of the QoC index is carried out in step S06. If the user score of the QoE index reaches more than 0.55 points, the system quality index will not be adjusted. If the user score of the QoE index is below 0.55, the relevant information data of the sensing device information configuration unit and the scenario information storage unit need to be adjusted.

[0163] Therefore, the present invention adopts the above-mentioned scenario-aware multi-index quality management system based on data fusion, realizes the adaptive adjustment and interaction of multiple system quality indexes, comprehensively measures the quality level of the scenario-aware system through the interactive management of multiple quality indexes, and at the same time utilizes the mutual restriction relationship between different indexes to realize the adaptive adjustment and optimization of the system under the condition of lacking real user feedback.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A scenario-aware multi-index quality management system based on data fusion, characterized in that: It includes a collection device quality management module, a scenario information quality management module, a service quality management module, a user experience quality management module, and a driving decision quality management module; The collection device quality management module, the scenario information quality management module, the service quality management module, and the driving decision quality management module form a closed-loop connection. The scenario information quality management module is connected to the user experience quality management module. The service quality management module and the driving decision quality management module are respectively interconnected with the user experience quality management module; The service quality management module includes a knowledge graph-based service recommendation unit, a subjective QoS acquisition unit, an objective QoS acquisition unit, a QoS fusion sorting and optimization unit, and a subjective and objective threshold screening unit connected in sequence. The subjective and objective threshold screening unit includes a subjective QoS threshold and an objective QoS threshold; The user experience quality management module includes a user and system feedback unit, a subjective QoE acquisition unit, and an objective QoE acquisition unit; The knowledge graph-based service recommendation unit is used to send a request for the required service according to the high-level scenario information; The subjective QoS acquisition unit obtains the subjective QoS based on the user preferences and the subjective QoE. It is used to determine the subjective QoS parameters and the corresponding weights according to the statistics of the user expectations and preferences. The QoS parameters related to the user preferences determine the weights according to the user preferences, and the QoS parameters not related to the user preferences are processed according to the default weights. The subjective QoS parameters are optional parameters; The objective QoS acquisition unit calculates the QoS parameter values through the current service and the objective QoE. It is used to obtain the objective QoS parameters that meet the required service and the corresponding weights. The objective QoS parameters are mandatory parameters; The QoS fusion sorting and optimization unit is used to classify and fuse the subjective and objective QoS weights to obtain the subjective and objective QoS values respectively, and sort them in descending order; The subjective and objective threshold screening unit sets the QoS threshold according to the user preferences and the QoE. It is used to calculate the subjective and objective QoS thresholds and screen the services recommended by the system. The subjective QoS threshold is determined according to the user's subjective preferences. When the user preference feedback information cannot be obtained, the subjective QoS threshold is the same as the objective QoS threshold.

2. The multi-index quality management system for situation awareness based on data fusion according to claim 1, characterized in that: The collection device quality management module includes a sensing device information configuration unit, a QoD parameter acquisition unit, and a low-level scenario information acquisition unit connected in sequence; The sensing device information configuration unit is used to receive the feedback information in real time and adjust the information configuration of the sensing device; The QoD parameter acquisition unit is used to acquire the QoD parameters of the physical sensors; The low-level scenario information acquisition unit is used to collect data using intelligent devices that meet the system requirements and conform to the QoD parameters, intuitively quantify the matching degree of each intelligent sensing device with the current processing requirements, and transmit the acquired low-level scenario information to the scenario information quality management module for further processing.

3. The multi-index quality management system for situation awareness based on data fusion according to claim 1, wherein: The scenario information quality management module includes a scenario information preprocessing unit, a scenario information repository unit, a QoC parameter weighted fusion unit, a QoC-based scenario information uncertainty elimination unit, and an advanced scenario information acquisition unit, which are connected in sequence; The scenario information preprocessing unit is used to standardize homogeneous scenario information, classify heterogeneous scenario information, and transmit the original scenario information and the preprocessed scenario information to the scenario information repository unit; The scenario information repository unit is used to store the original and preprocessed scenario information; The QoC parameter weighted fusion unit is used to calculate each type of QoC parameter value of the scenario information, and perform weighted fusion according to the system and user requirements by setting corresponding parameter weights; The QoC-based scenario information uncertainty elimination unit is used to handle the incompleteness, imprecision, and inconsistency of the scenario information; The advanced scenario information acquisition unit is used to model, reason, and fuse the low-level scenario information after uncertainty elimination to obtain advanced scenario information.

4. A scenario-aware multi-index quality management system based on data fusion according to claim 1, characterized in that: The drive decision quality management module includes a decision optimization unit, a drive control unit, and a decision adaptive perception unit. The decision adaptive perception unit includes a subjective decision perception subunit, an objective decision perception subunit, and a quality index adaptive adjustment subunit; The decision optimization unit and the drive control unit are connected in sequence. Both the subjective decision perception subunit and the objective decision perception subunit are connected to the quality index adaptive adjustment subunit; The decision optimization unit is used to judge the degree to which the service meets the expectations according to the actuator information and the environment information, and further optimize the service after screening by the subjective QoS threshold and the objective QoS threshold to adapt to the operation requirements of the actual actuator; The drive control unit is used to directly control the actuators serving the environment, system, and customers, avoid unnecessary responses or actions of the physical actuators directly serving users and the environment, and perform stable control and management of physical or virtual terminals based on a self-regulating fault detection and calibration mechanism, and provide the information reflecting the decision execution level to the objective QoE acquisition unit to increase the feedback information when there is no real user feedback; The decision adaptive perception unit is used to automatically perceive the execution situation of the current system decision; The subjective decision perception subunit is used to count the decision execution situations from user feedback and correction and the decision execution situations of the cascaded system feedback. The subjective decision perception subunit is an optional item; The objective decision perception subunit is used to actively perceive whether the impact of the decision execution on the entity meets the expectations. The objective decision perception subunit is a mandatory item; The quality index adaptive adjustment subunit is used to perform adaptive adjustment operations on the system according to the interaction performance of the quality indexes at all levels of the system and the decision execution situation, and send the QoD and QoC information that needs to be adjusted to the perception device information configuration unit. The corresponding information included will be extracted layer by layer in different modules and used to adjust the quality indexes of each module.

5. The multi-index quality management system for situation awareness based on data fusion according to claim 1, characterized in that: The user and system feedback unit is connected to the subjective QoE acquisition unit; The subjective QoE acquisition unit is used to capture the QoE actively provided by users, and send the subjective QoE information to the subjective QoS acquisition unit to jointly assist in determining the subjective QoS parameters and their corresponding weights; The objective QoE acquisition unit is used to predict the objective QoE by fusing the comprehensive QoC parameters, QoS parameters and decision execution status, and send the result to the objective QoS acquisition unit for adaptive adjustment of quality indicators and optimization of services.

6. A working method of the data fusion-based multi-index quality management system for situation awareness according to any one of claims 1-5, characterized in that, It includes the following steps: Step S01: Configure the information of the sensing device, and preset and adjust the relevant configuration information of the intelligent sensing device according to the specific application requirements of the scenario sensing system; Step S02: Obtain the QoD parameters, and obtain the QoD parameters of the intelligent sensing device in the applicable state in the current scenario; Step S03: Obtain the low-level scenario information, and select the intelligent sensing device whose QoD parameters meet the system requirements to collect the scenario information; Step S04: Preprocess the scenario information, classify and preprocess the collected scenario information. For homogeneous scenario information, the data format is unified into a preset standard format; For heterogeneous data, extract the effective information and recombine it into a unified data format; Step S05: Store the scenario information, store the collected original scenario information, the preprocessed scenario information and the system preset parameter threshold information in the database; Step S06: Obtain the QoC parameters, and calculate the quality of the obtained scenario information, including reliability, update degree, correctness, and scenario information correlation parameters; Step S07: Weighted fusion of QoC parameters, analyze the nature of the information source of the QoC parameters, match the appropriate QoC parameters, and perform intelligent fusion on the QoC parameters to obtain the comprehensive quality of the scenario information; Step S08: Eliminate the uncertainty of the scenario information based on QoC, and perform inconsistency, incompleteness and imprecision elimination processing on the scenario information with the weighted fusion QoC parameters; Step S09: Obtain the high-level scenario information, further process the scenario information after uncertainty elimination, and fuse the high-quality scenario information through the inference fusion method. The inference fusion method adopts the D-S evidence theory and the Bayesian method; Step S10: Recommend the knowledge graph service, and send a request for the currently required service according to the obtained high-level scenario information by using the recommendation system algorithm based on the knowledge graph; Step S11: Obtain the subjective QoS, and obtain the subjective QoS according to the user preference and subjective QoE. For the QoS parameters related to the user preference, determine the QoS parameter weights according to the user, and for the QoS parameters not related to the user preference, process them according to the default weights; Step S12: Obtain the objective QoS, and calculate the QoS parameter value through the current service and the objective QoE; Step S13: QoS fusion sorting and optimization. Different weights are given to subjective and objective QoS to obtain the comprehensive QoS score and rank the eligible services. The subjective weight is determined according to user preferences and the fuzzy AHP method, and the objective weight is determined by the information entropy method. When the subjective weight is , the objective weight is , the subjective QoS value is , and the objective QoS value is , the comprehensive QoS value is ; Step S14: Obtaining QoS subjective and objective thresholds. Set the QoS thresholds according to user preferences and QoE. The subjective threshold is set based on the service level required by the user. Select the services that meet the user's requirements and save computing and storage resources from the services sorted in step S13. The objective threshold is set according to the general threshold preset by the system. When the subjective QoS threshold cannot be obtained, make it consistent with the objective QoS threshold; Step S15: Determine whether the QoS exceeds the threshold. Compare the QoS of the obtained service with the threshold set in step S14. If it does not exceed the threshold, select the service with the highest score within the threshold. Otherwise, even if there are services with higher scores outside the threshold, do not consider them and re-enter step S10; Step S16: Obtaining service availability and consistency. According to the service provided in step S15, combine with the drive control unit to obtain the availability and consistency of the service; Step S17: Drive control. Based on the self-adjustable fault detection and calibration mechanism, perform stable control and management on the physical or virtual terminal, and provide the information reflecting the execution decision of the actuator to the objective QoE acquisition unit; Step S18: Obtaining user and system feedback. Obtain the real feedback of the user, which is fundamentally different from the predicted objective QoE here. At the same time, obtain the feedback information from the cascaded system and store it; Step S19: Determine whether the feedback is from the user. Determine whether the feedback in step S18 is from a real user. If it is from a real user, enter step S21; otherwise, enter step S22; Step S20: Obtaining objective QoE. Calculate the objective QoE by combining the QoC parameters, drive control feedback data, and the QoS parameter values of the current service, and feedback the result to the objective QoS acquisition unit in step S12; Step S21: Obtaining subjective QoE. Capture the QoE actively provided by the user, and feedback the result to the subjective QoS acquisition unit in step S11; Step S22: Obtaining QoME. Calculate the machine experience quality from the system rather than the user; Step S23: Subjective decision perception. According to the QoME of the cascaded system, obtain the degree to which the current actuator execution decision meets the expectation; Step S24: Objective decision perception. According to the adaptive judgment of the execution unit, obtain the degree to which the actuator execution decision meets the expectation and feedback it to the quality index adaptive adjustment unit; Step S25: Adaptive adjustment of quality index. Perform adaptive adjustment operations on the system according to the interaction performance of the quality indexes at all levels of the system and the decision execution situation.

Images