BROWSE
Detail
Journal Article
Enhancing digital twin efficiency in indoor environments: Virtual sensor-driven optimization of physical sensor combinations
Cited 7 time in 
Share 
Share
- Authors
- Hakjong Shin, Younghoon Kwak
- Issue Date
- 2024-05
- Citation
- Automation in Construction, v.161, pp.1-15
- ISSN
- 0926-5805
- Publisher
- Elsevier BV
- Language
- English
- Type
- Journal Article
- Abstract
- Multiple sensor nodes are preferred for the development of digital twins in indoor environments; however, it poses significant financial burdens. Therefore, considering both cost-effectiveness and information richness, optimal physical sensor placement and efficient virtual sensor operation must be balanced. Data-driven techniques for sensor placement prioritize locations via statistical metrics, information fidelity, and signal strength. When extending spatial sensing through virtual sensors, the indirectly determined physical sensor sites encounter challenges like multicollinearity and overfitting, hampering the reliability of predictive models (virtual sensors). To address this, this study introduces an algorithm that selects optimal physical sensor combinations based exclusively on virtual sensor performance. The results highlight the superiority of our proposed algorithm over conventional methods for virtual sensor development. Notably, our approach reduces computational workload compared to brute-force methods. This algorithm provides an economical means to manage extensive data and holds substantial promise for advancing digital twin capabilities in indoor environments.
- KSP Keywords
- Brute-force, Conventional methods, Cost-effectiveness, Data-driven techniques, Digital Twin, Indoor environment, Information richness, Predictive model, Sensor Performance, Sensor Placement, Sensor development