A game theory approach to target tracking in sensor networks

Area of Science:

• Robotics
• Control Systems
• Signal Processing

Background:

• Sensor networks are crucial for target tracking.
• Moving targets can exhibit intelligent, evasive behavior.
• Traditional tracking methods struggle against adversarial targets.

Purpose of the Study:

• To develop a robust moving-target tracking algorithm for sensor networks.
• To address challenges posed by intelligent, evasive targets.
• To create a distributed tracking solution for multi-node sensor systems.

Main Methods:

• Formulated target estimation as a zero-sum game.
• Employed a minimax filter to minimize worst-case estimation error.
• Developed a distributed minimax filter using neighbor communication.

Main Results:

• The proposed distributed minimax filter effectively tracks moving targets.
• The algorithm demonstrates robustness against intelligent adversaries.
• Simulation results confirm satisfactory performance in complex scenarios.

Conclusions:

• The distributed minimax filter offers a viable solution for robust adversarial target tracking.
• This approach enhances the capabilities of sensor networks in dynamic environments.
• Future work may involve real-world deployment and further algorithm optimization.