Flexible Fusion Structure-Based Performance Optimization Learning for Multisensor Target Tracking.

Area of Science:

• Sensor fusion
• Network systems engineering
• Optimization algorithms

Background:

• Complex air task network systems face challenges from time-varying environments, moving nodes, and non-cooperative targets.
• Limitations in communication bandwidth and measurement distance necessitate dynamic adjustments to fusion structures.
• Fixed fusion structures offer limited adaptability to dynamic operational conditions.

Purpose of the Study:

• To design a flexible fusion algorithm using optimization learning for dynamic sensor subset selection.
• To enable systems to achieve goals under given constraints by adapting sensor configurations and fusion methods.
• To introduce and define a novel survivability index for system performance evaluation.

Main Methods:

• Development of optimization models for single and multi-target tracking incorporating system performance indexes.
• Utilization of optimization learning technology for dynamic sensor subset selection.
• Definition of system performance indexes, including a specific survivability index.

Main Results:

• The proposed flexible fusion structure demonstrates superior adjustment performance compared to fixed structures.
• Optimization models were established and solved for dynamic sensor selection in target tracking scenarios.
• Simulation examples validated the effectiveness of the proposed flexible fusion algorithms.

Conclusions:

• Flexible fusion structures with mixed methods offer significant advantages in complex air task network systems.
• Dynamic sensor subset selection is crucial for adapting to time-varying conditions and system constraints.
• The developed optimization learning approach provides a robust method for enhancing system performance and survivability.