Abstract
Health prediction plays an essential role in improving the reliability of a sensor network by guiding the network maintenance. However, affected by interference factors in the real operational environment, the reliability of the monitoring information about the sensor network tends to decline, which affects the health prediction accuracy. Furthermore, the lack of monitoring information and high complexity of the network increase the difficulty of health prediction. To solve these three problems, this paper proposes a new sensor network health prediction model based on the belief rule base model with attribute reliability (BRB-r). The BRB-r model is an expert system that fully considers the qualitative knowledge and quantitative data of the sensor network. In addition, it can address the fuzziness and nondeterminacy of this qualitative knowledge. In the new model, the unreliable monitoring information of the sensor network is handled by the attribute reliability mechanism. The reliability of the sensor is calculated by the average distance method. Due to the effect of the fuzziness and nondeterminacy of expert knowledge, the health status of the sensor network cannot be accurately estimated by the initial health prediction model. Consequently, the optimization model for the health prediction model is established. Finally, a case study regarding a sensor network for oil storage tanks is conducted, and the validity of this method is demonstrated.
Highlights
Health prediction plays an essential role in improving the reliability of a sensor network by guiding the network maintenance
To address the problems in health prediction for sensor networks, this paper proposes a new sensor network health prediction model based on base model with attribute reliability (BRB-r)
In the health status prediction model, the observation data are obtained by sensors, and key features should be selected as inputs for the belief rule base (BRB)-r model
Summary
The health status of a sensor network represents its comprehensive state, and it can be used to assist with decision making regarding system maintenance. The problems in engineering practice with respect to the sensor network health prediction are described, and a new health prediction model based on BRB-r is proposed. Problem formulation regarding health status prediction for sensor networks. The problems in sensor network health prediction can be formulated as follows. XM (t) are the features of the sensor network at time instant t ; r is the expert knowledge applied in the health prediction model; v is the unknown parameters. When uncertain expert knowledge is included in the health prediction model, its prediction accuracy is affected. The third problem is how to adjust the health prediction model constructed by experts. To address the problems in health prediction for sensor networks, this paper proposes a new sensor network health prediction model based on BRB-r. H(t + 1) represents the Scientific Reports | (2021) 11:2806 |
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.
