A Weighted Belief Entropy-Based Uncertainty Measure for Multi-Sensor Data Fusion

Area of Science:

• Artificial Intelligence
• Information Fusion
• Sensor Networks

Background:

• Quantifying uncertainty in sensor reports is crucial for effective data fusion.
• Existing methods in Dempster-Shafer evidence theory may lead to information loss.
• Accurate uncertainty measurement is vital for reliable sensor data fusion in real-world applications.

Purpose of the Study:

• To propose a novel weighted belief entropy for quantifying uncertainty in sensor data.
• To develop a new multi-sensor data fusion approach utilizing the proposed uncertainty measure.
• To reduce information loss during sensor data processing and fusion.

Main Methods:

• Introduced a weighted belief entropy based on Deng entropy within the Dempster-Shafer framework.
• The weighting scheme considers the relative scale of propositions within the frame of discernment (FOD).
• Developed a multi-sensor data fusion method incorporating the weighted belief entropy.

Main Results:

• The proposed weighted belief entropy quantifies uncertainty by considering both mass functions and FOD scale.
• The new fusion method demonstrates reduced information loss compared to existing approaches.
• Experimental validation on artificial data and motor rotor fault diagnosis confirms the method's effectiveness.

Conclusions:

• The weighted belief entropy offers a more comprehensive measure of uncertainty in Dempster-Shafer theory.
• The novel data fusion approach enhances accuracy and reduces information loss in sensor data processing.
• This work contributes to more robust and reliable multi-sensor data fusion systems.