Distributed Estimation of Fields Using a Sensor Network with Quantized Measurements.

Area of Science:

• Sensor Networks
• Distributed Estimation
• Signal Processing

Background:

• Scalar field estimation is crucial for applications like environmental monitoring.
• Existing methods often require centralized processing, which can be inefficient for large sensor networks.
• Quantized sensor measurements introduce challenges in distributed estimation.

Purpose of the Study:

• To develop and evaluate distributed estimation algorithms for scalar fields using sensor networks with quantized measurements.
• To propose two novel schemes: measurement diffusion and estimate diffusion.
• To assess the performance of these algorithms, including their ability to handle time-varying fields.

Main Methods:

• Distributed estimation algorithms where sensors locally share information with neighbors.
• Measurement diffusion: sensors broadcast received measurements.
• Estimate diffusion: sensors broadcast local estimates and Hessians.
• Iterative combination of information at each sensor.

Main Results:

• Both measurement diffusion and estimate diffusion schemes effectively estimate scalar fields.
• The algorithms can handle time-varying scalar fields.
• Numerical studies show steady-state performance comparable to centralized estimation.

Conclusions:

• Distributed estimation with quantized sensor data is feasible using the proposed diffusion schemes.
• These methods offer a viable alternative to centralized estimation, particularly for large-scale sensor networks.
• The algorithms demonstrate robustness and efficiency in estimating scalar fields.