The Research On System Framework And Evidence Theory Of Multi-source Data Fusion And Their Applications In Smart Healthcare

With the rapid development of sensor and IoT(Internet of Things)technology,the way people access to information has become more and more diverse.Making good use of the information obtained can be of great value.As an intelligent information processing technology,multi-source data fusion technology plays an important role and has a great development in recent years.However,the current multi-source data fusion frameworks are mostly proposed for specific fields,and a generalized fusion framework has not yet formed.Besides.the existing multi-source data fusion algorithms still have limitations in the application.For example.Dempster-Shafer(D-S)evidence theory algorithm cannot obtain correct results when dealing with highly conflictive evidences.Based on the above problems,the main research works of this thesis are as follows:(1)Research on the multi-source data fusion system frameworkMost of the frameworks used in the current multi-source data fusion systems are only suitable for specific fields,which means they cannot cope with complex and varying practical scenarios,and cannot get satisfactory fusion results when processing data with different sources and characteristics.This thesis proposes a multi-source data fusion system framework based on data evaluation and system coordination module.By introducing this module,the appropriate algorithm can be selected after analyzing the data collected by multiple sensors.It can also coordinate the work of each module in the whole system.The proposed fusion system framework makes the whole data fusion process become more flexible and efficient,which can adapt to more practical scenarios and has strong versatility.(2)Research on D-S evidence theory algorithm in multi-source data fusionAs an uncertainty reasoning method,D-S evidence theory algorithm has important applications in multi-source data fusion.However,it can usually lead to incorrect results when fusing the highly conflictive evidences.To solve the above problem,this thesis proposes an improved D-S evidence theory algorithm based on gray relational analysis.By analyzing the degree of correlation between evidences,the proposed algorithm modifies the original evidence model,and uses D-S synthesis formula to fuse the new evidence model.The fusing results are more credible compared with other improved algorithms.(3)Research on the application of the proposed system framework and algorithm in smart hea thcareThis thesis discusses the implementation of the proposed system framework in smart healthcare scenarios.and applies the improved D-S evidence theory algorithm in specific scenarios such as physical condition judgment and disease diagnosis.In addition,combining the proposed framework and the specific application of physical condition judgment.this thesis proposes an algorithm selection mechanism based on the degree of data conflict.The proposed mechanism enables the fusion system to select the more appropriate algorithm between the proposed algorithm and traditional D-S evidence theory by analyzing the characteristics of data samples,when processing the information collected by multiple physiological sensors.It can improve the accuracy rate of fusion effectively.Through the simulation,the effectiveness of the improved D-S evidence theory algorithm and the rationality of the proposed framework are verified.This thesis studies the system framework and algorithm in multi-source data fusion,and proposes a new system framework and an improved D-S evidence theory,and also discusses their applications in smart healthcare scenario.Through the above work,this thesis provides a mode for the design of general system framework and a new method for the improvement of D-S evidence theory in multi-source data fusion fields.The framework and the improved D-S evidence method proposed in this thesis can be extended to more scenarios and has wide practical significance.