Tracking Multiple Marine Ships via Multiple Sensors with Unknown Backgrounds.

Area of Science:

• Multi-target tracking
• Sensor fusion
• Signal processing

Background:

• Accurate multi-target tracking relies on understanding background parameters like clutter and detection probability.
• These parameters are often assumed constant but are actually unknown and time-varying, leading to tracking errors.
• Existing algorithms suffer performance degradation when these background characteristics are misspecified.

Purpose of the Study:

• To develop an online method for tracking multiple targets using multiple sensors.
• To jointly adapt to unknown and time-varying clutter rate and probability of detection.
• To improve the accuracy and robustness of multi-target tracking algorithms in dynamic environments.

Main Methods:

• Proposed a novel filtering approach for online multi-target tracking.
• Developed a method for parallel estimation of unknown clutter rate and detection probability.
• Integrated estimated parameters into the generalized labeled multi-Bernoulli filter.

Main Results:

• The proposed method effectively adapts to slowly-varying clutter and detection probability.
• Numerical studies using multistatic Doppler data demonstrate the validity and improved performance.
• The joint estimation enhances the accuracy of multi-target tracking compared to fixed-parameter assumptions.

Conclusions:

• The developed online adaptive filter significantly improves multi-target tracking performance.
• Jointly estimating unknown background parameters is crucial for robust tracking systems.
• The method shows promise for real-world applications with dynamic environmental conditions.