Abstract
The paradigm of multisensor data fusion has been evolved from a centralized architecture to a decentralized or distributed architecture along with the advancement in sensor and communication technologies. These days, distributed state estimation and data fusion has been widely explored in diverse fields of engineering and control due to its superior performance over the centralized one in terms of flexibility, robustness to failure and cost effectiveness in infrastructure and communication. However, distributed multisensor data fusion is not without technical challenges to overcome: namely, dealing with cross-correlation and inconsistency among state estimates and sensor data. In this paper, we review the key theories and methodologies of distributed multisensor data fusion available to date with a specific focus on handling unknown correlation and data inconsistency. We aim at providing readers with a unifying view out of individual theories and methodologies by presenting a formal analysis of their implications. Finally, several directions of future research are highlighted.
Highlights
Multisensor data fusion refers to the process of utilizing additional and complementary data from multiple sources to achieve inferences that are not feasible/possible from an individual data source operating independently
In Reference [32], the Cholesky decomposition model of unknown cross-correlation is applied to Bar-Shalom Campo (BC) formula, and the fused solution is iteratively approximated based on min-max optimization function for unknown correlation coefficient ρ
Identification and subsequent removal of inconsistent and abnormal data Uses residuals generated between the modeled outputs and Characteristics actual sensor measurements to identify inconsistency
Summary
Multisensor data fusion refers to the process of utilizing additional and complementary data from multiple sources to achieve inferences that are not feasible/possible from an individual data source operating independently. Multisensor data fusion is to obtain a more meaningful and precise estimate of a state by combining data from multiple sensors and model-based predictions. In a distributed architecture where the assumption of statistical independence is not applicable, filtering without taking the cross-correlation into account may lead to divergence due to the inconsistency in fused mean and covariance [9]. We denote the Gaussian distribution as paper is concluded and several future directions12of research in distributed data fusion are x ∼ N ( x, P), with mean xand covariance P. Localarchitectures sensors may pre-process the data before transmitting it to the central node, the term ‘raw data’ signify sensorFusion measurements or pre-processed data without filtering or local fusion
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
