A Weighted Decision-Level Fusion Architecture for Ballistic Target Classification in Midcourse Phase

Area of Science:

• Missile Defense Systems
• Radar Signal Processing
• Artificial Intelligence

Background:

• Ballistic missile defense faces challenges in accurately recognizing warheads during the midcourse phase.
• Sensor data heterogeneity and varying reliability complicate target identification.

Purpose of the Study:

• To develop an advanced fusion architecture for improved warhead recognition.
• To address sensor data differences and reliability variations in multi-sensor systems.

Main Methods:

• Proposed a weighted decision-level fusion architecture utilizing data from multiple radar sensors.
• Implemented an online feature reliability evaluation method to generate dynamic sensor weight coefficients.
• Tested the model with a simulation dataset involving multiple sensors and bandwidths.

Main Results:

• The weighted decision-level fusion method significantly enhances warhead recognition rates compared to single-sensor approaches.
• The model effectively accounts for environmental and locational influences on sensor performance reliability.
• Demonstrated compatibility with diverse classifiers, including traditional and ensemble learning algorithms.

Conclusions:

• The proposed weighted fusion architecture offers a robust solution for improving warhead recognition in ballistic midcourse phases.
• Dynamic adjustment of sensor weights based on reliability is crucial for effective multi-sensor data fusion.
• The method shows promise for enhancing the overall performance and accuracy of missile defense systems.