China
Africa
Explore chapters and articles related to this topic
Heterogeneous Data Fusion for Healthcare Monitoring: A Survey
Published in Rashmi Agrawal, Marcin Paprzycki, Neha Gupta, Big Data, IoT, and Machine Learning, 2020
Shrida Kalamkar, Geetha Mary A
Sensor data fusion is an essential and integral part of the Internet of Things (IoT). Sensors are used in variety of applications, such as climate monitoring, smart mobile devices, healthcare, automotive systems, industrial control, traffic control. Data in IoT is dynamic and heterogeneous, which leads to inadequacy of simple single-source analysis methods. Data fusion integrates multiple data and knowledge into a consistent, accurate and useful representation which makes data fusion to provide high-quality information for a reliable decision support. Data fusion also leads to an increase in the accuracy of information generated from multiple sources by reducing the uncertainty of data. The fusion of complementary information generated from multiple sources can provide more accurate information instead of single-source information. Confidence can be increased when multiple independent measurements are made on the same event dataset. Therefore, the result is more reliable. In the world of the IoT, as the size of data increases, handling these large volumes of streaming and historical data, which can vary from structured to unstructured and numerical to microblog data streams, is challenging because its volume is heterogeneous and highly dynamic. Hence, techniques and methodology for understanding and resolving issues about data fusion in the IoT needs to be investigated. The data sources for a fusion process are not specified to originate from identical sensors. Therefore, the data fusion process is categorised into two types, as shown in Figure 9.1.
Similarity-Based Artificial Intelligence
Published in Mark Chang, Artificial Intelligence for Drug Development, Precision Medicine, and Healthcare, 2020
Data fusion is the process of integrating multiple data sources to produce more consistent, accurate, and useful information beyond that which any individual data source can provide. A simple sample will be integrated with the data from various clinical trials and published clinical trial data. Such data usually are a mix of individual patient data and trial summary data (such as means, medians, confidence intervals, standard errors, sample sizes and p-values). Interestingly, humans constantly use data fusion in comprehending the surrounding world. As humans, we rely heavily on our vision, smell, taste, hearing, and physical movement. We rely on a fusion of smelling, tasting, and touching food to ensure that it is edible (or not). Similarly, we rely on our ability to see, hear, and control the movement of our body to walk or drive and to perform most of our daily tasks. Our brain performs fusional processing based on individual knowledge at instants in time, and we take the appropriate action.
Artificial Neural Network-Based Modeling and Controlling of Drying Systems
Published in Alex Martynenko, Andreas Bück, Intelligent Control in Drying, 2018
Mortaza Aghbashlo, Soleiman Hosseinpour, Arun S. Mujumdar
The combination of ANN technology with real-time measurement tools like bio-sensing, imaging, spectral, acoustical, and electrical techniques can be one of the interesting subjects for real-time monitoring and control of drying systems in future studies. Figure 9.8 shows a comprehensive flowchart for real-time monitoring and control of drying systems using this combination. The required information about the process can be captured using some independent sources such as color CCD camera, digital microscope, infrared camera, electronic nose and tongue, and so on. The obtained data can then be subjected to preprocessing and processing techniques like image and signal processing tools in order to enhance the quality of signals and to extract more appropriate features about the process. Afterward, the features extracted from different independent sources can be merged into a single features vector employing some data fusion techniques such as evidence theory (Dempster-Shafer theory). After finding an appropriate feature vector, modeling should be carried out to establish a proper model for correlating the feature vector to unmeasured physicochemical and thermodynamic characteristics such as moisture ratio, drying rate, chemical attributes, and energy consumption. Using such a methodology, the weighing system can be eliminated. Furthermore, there is no need for sophisticated and expensive instruments for measuring chemical attributes.
A Multimodal Human-Computer Interaction for Smart Learning System
Published in International Journal of Human–Computer Interaction, 2023
Tareq Mahmod Alzubi, Jafar A. Alzubi, Ashish Singh, Omar A. Alzubi, Murali Subramanian
The data fusion module is an essential process that integrates and correlates information from multiple sources to achieve better decision-making results. The process of data fusion involves combining multiple interrelated datasets, which can be classified into low-data, intermediate, high-data, and sensor fusion types, depending on the nature of the input and output data. The proposed model employs a Data in-data out (DAI-DAO) multi-modality approach for the fusion of inter-related datasets collected from multiple sources. The fusion process is initiated immediately after data collection. The raw data are fused to provide more reliable and accurate information. This type of fusion reduces the amount of data by retaining only useful information for high-level processes, resulting in a more efficient and effective data analysis. The fusion process can be represented by Equation (8). Where Dataset1 to DatasetN are the different types of input data and f is the fusion function that correlates and merges the information to produce a single output.
A novel evaluation method for pavement distress based on impact of ride comfort
Published in International Journal of Pavement Engineering, 2022
Yishun Li, Chenglong Liu, Yuchuan Du, Shengchuan Jiang
Multi-source data fusion, or multi-sensors fusion, is the process of integrating multiple data sources to produce more consistent, accurate, and useful information than that provided by any individuals. The method of data fusion is used in many fields of detection and evaluation to achieve higher accuracy or more comprehensive assessment. Du et al. (2019) achieved rapid detection of pavement anti-skid performance by fusing three-dimensional space and reflection intensity information. Huang et al. (2014) used the Dempster-Shafer (D-S) evidence theory to detect pavement crack combining 2D grey-scale image analysis and 3D laser scanning information. 2D grey-scale image processing can help detect the edge, however, such technology cannot distinguish the dark areas caused by tyre marks, oil spills, shadows, or repairs from the real pavement cracks. 3D laser scanning analysis can accurately quantify the crack in case of shadows and tyre marks, while it is overly sensitive to noise. The fusion of two sources of data can improve the detection results. Rostami Shahrbabaki et al. (2018) fused a spot detector and connected vehicle data for real-time traffic state estimation in urban signalised links.
Risk-based, sensor-fused detection of flooding casualties for emergency response
Published in Ships and Offshore Structures, 2021
Kristian Bertheussen Karolius, Jakub Cichowicz, Dracos Vassalos
Data fusion is the process of combining information from a number of different sources to provide a robust and complete description of an environment or process of interest. (Durant-Whyte and Henderson 2016)The process of fusing, or aggregating, scattered information from a range of independent sensors is important in applications where large amount of data must be combined, to obtain information of quality adequate for supporting decision-making. Data fusion is applied across a wide range of industries, such as military systems, surveillance and monitoring systems, process control systems, and information systems. It also plays a central role in autonomous systems and robotics because it allows essential measurements and information to be combined to generate knowledge with high level of confidence (or lack of uncertainty), to enable decisions to be executed autonomously. Due to the complex and multivariate nature of the problem, multi-sensor data fusion process is highly relevant for obtaining strengthened information for optimised decision making in flooding emergences.