Advantages and Challenges of Information Fusion Technique for Big Data Analysis: Proposed Framework

Introduction: Recently, with the surge in the availability of relevant data in various industries, the use of Information Fusion technique for data analysis is increasing. This method has several advantages, such as increased accuracy, and the use of meaningful information. In addition, there are certain challenges, including the impact of data type and analytical method on results. The goal of this study is to propose a framework for introducing the advantages and classifying the challenges of this technique. Method: We conducted a review of articles published between January 1960 and December 2017 for the design stage and from January 2018 to December 2018 for the evaluation stage. Articles were identified from various databases such as Science Direct, IEEE, Scopus, Web of Science, and Google Scholar, using the keywords decision fusion, information fusion, and symbolic fusion. We report the advantages and challenges of the methodologies described in these articles. Analysis was conducted in accordance with PRISMA guidelines. Results: A total of 132 articles were identified in the design stage and 90 articles were identified in the evaluation stage. Categories within the framework for challenges include “hardware and software requirements for processing and maintaining the process”, “data” and “data analysis method”. The categories for advantages include “value modeling”, “preferable management of uncertainty and variability”, “excellent decision making”, “comprehensive interpretation and representation”, “data management” and “simplicity of infrastructure”. Our results indicate using these two frameworks with 95% Confidence interval. Conclusion: An overall understanding of the advantages and challenges of the information fusion technique could act as a guide for the researcher for the correct usage of this technique.

Dynamic Clock Skew Estimation and Fusion Acquainted with Environmental Factors

Clock skew reflects the drift rate of a clock w.r.t. the nominal or reference clock frequency, which is the root cause of clock drifting. However, as the output of clock is largely affected by some environmental factors. Therefore, clock skew estimation is particularly difficult in wireless sensor networks (WSNs), as the working environments of WSN are usually dynamic, unpredictable or even hazard. Besides, sensors are usually powered by batteries with limited communication and computation capacity. The clock skew is found to be non-stationary containing severe measurement and process noises. Thus, we attempt to jointly consider the environmental factors into the clock skew estimation using Kalman filter. We propose to use the change of temperature and/or voltage to enhance the clock skew estimation performance. Besides, in multi-hop wireless networks, where the synchronization is done in a hierarchical procedure, one node may have the chance to receive more than one timestamps. Therefore, we propose to use the information fusion technique to dynamically combine the information contained in different timestamps to promote the clock skew estimation accuracy. Besides, we further derive the statistic lower bound of estimation errors, which can serve as a benchmark. The performance of the proposed schemes have been verified by extensive simulation results, where the root mean square error (RMSE) can be reduced by around 60% when compared with the previous solutions.

Cloud-Based Platform for Optimal Machining Parameter Selection Based on Function Blocks and Real-Time Monitoring

The way machining operations have been running has changed over the years. Nowadays, machine utilization and availability monitoring are becoming increasingly important for the smooth operation of modern workshops. Moreover, the nature of jobs undertaken by manufacturing small and medium enterprises (SMEs) has shifted from a mass production to small batch. To address the challenges caused by modern fast changing environments, a new cloud-based approach for monitoring the use of manufacturing equipment, dispatching jobs to the selected computer numerical control (CNC) machines, and creating the optimum machining code is presented. In this approach the manufacturing equipment is monitored using a sensor network and though an information fusion technique it derives and broadcasts the data of available tools and machines through the internet to a cloud-based platform. On the manufacturing equipment event driven function blocks with embedded optimization algorithms are responsible for selecting the optimal cutting parameters and generating the moves required for machining the parts while considering the latest information regarding the available machines and cutting tools. A case study based on scenario from a shop floor that undertakes machining jobs is used to demonstrate the developed methods and tools.

Songs Recognition Using Audio Information Fusion

The article presents information fusion approach for song classification with use of acoustic signal. Many acoustic features can contribute to correct identification of a song. Taking into consideration only one set of features may result in omission of relevant information. It is possible to improve the accuracy of identification process by means of the information fusion technique, in which various aspects of acoustic fingerprint are taken into consideration. Two sets of signal features were distinguished: one were based on frequency analysis (harmonic elements) and the other were based on multidimensional correlation ratios. An identification of a commercial was made with use of SVM and k-NN classifiers. The music audio signal database was used for assessing the effectiveness of the proposed solution. Results show an improved effectiveness of identification in relation to applying only one set of song features