Multi-Sensor Adaptive Weighted Data Fusion Based on Biased Estimation.

Area of Science:

• Signal Processing
• Data Fusion
• Estimation Theory

Background:

• Optimal weighting factors in data fusion can lose optimality.
• Unbiased estimators may still have reducible estimation error.
• Existing data fusion methods like least squares and batch estimation have limitations.

Purpose of the Study:

• To propose a multi-sensor adaptive weighted data fusion algorithm using biased estimation.
• To analyze the reasons for the loss of optimality in weighting factors.
• To reduce the estimation error of unbiased estimators in data fusion.

Main Methods:

• Proving that an unbiased estimator can further optimize estimation error.
• Developing a method to construct a biased estimation value from an unbiased one.
• Calculating the optimal weighting factor using estimation error.
• Comparing performance via simulation tests against least squares and batch estimation.

Main Results:

• Biased estimation data fusion demonstrates superior accuracy.
• The proposed algorithm shows enhanced stability in data fusion.
• Biased estimation data fusion exhibits greater noise resistance compared to other methods.

Conclusions:

• The proposed biased estimation data fusion algorithm effectively overcomes the loss of optimality.
• This approach significantly reduces estimation error, improving overall data fusion performance.
• Biased estimation offers a robust and accurate solution for multi-sensor data fusion applications.