Data Reconstruction Using Smart Sensor Placement.

Area of Science:

• Geospatial analysis
• Data science
• Control theory

Background:

• Spatio-temporal field estimation is crucial for various applications but faces challenges with sensor placement and inherent uncertainties.
• Existing methods often struggle to optimally place sensors and account for positional uncertainties during data acquisition.

Purpose of the Study:

• To develop an efficient sensor placement strategy for spatio-temporal field estimation.
• To determine the optimal number of sensors for capturing key field features.
• To enhance field reconstruction robustness against environmental and model uncertainties using a data-driven control method.

Main Methods:

• QR decomposition for efficient sensor placement and determining optimal sensor count.
• Kriged Kalman Filter (KKF) for uncertainty-aware field reconstruction.
• Integration of positional uncertainty from data acquisition platforms into the KKF estimator.

Main Results:

• Demonstrated efficacy of the dynamic sensor placement strategy.
• Successful uncertainty-aware field estimation using the integrated KKF.
• Numerical results validate the proposed approach for robust spatio-temporal field reconstruction.

Conclusions:

• The proposed dynamic sensor placement strategy, combined with the KKF, provides an effective solution for spatio-temporal field estimation.
• The method enhances resilience to uncertainties by integrating positional data acquisition errors.
• This approach optimizes sensor utilization and improves the accuracy of field reconstruction.