Multi-Vehicle Cooperative Target Tracking with Time-Varying Localization Uncertainty via Recursive Variational...

Area of Science:

• Robotics and Autonomous Systems
• Sensor Fusion
• Estimation Theory

Background:

• Cooperative target tracking using multiple vehicles enhances state estimation.
• Accuracy relies on relative localization, often degraded by dynamic environments and satellite navigation unreliability.
• Time-varying and uncertain localization noise challenges existing cooperative tracking methods.

Purpose of the Study:

• To develop a robust recursive Bayesian framework for cooperative target tracking.
• To jointly estimate target state, cooperative vehicle state, and time-varying localization noise parameters.
• To ensure reliable cooperative tracking even with uncertain noise characteristics.

Main Methods:

• A recursive Bayesian framework is proposed for joint state and noise parameter estimation.
• An online variational Bayesian inference algorithm is developed for efficient recursive estimation.
• The framework handles time-varying and uncertain statistical characteristics of relative localization noise.

Main Results:

• Simulation results demonstrate the algorithm's effectiveness in boosting target tracking accuracy.
• The proposed method successfully adapts to dynamically changing localization noise.
• Accurate estimation of both target and vehicle states is achieved under challenging noise conditions.

Conclusions:

• The developed recursive Bayesian framework reliably improves cooperative target tracking.
• The online variational Bayesian inference algorithm provides efficient and accurate state estimation.
• This approach is effective for cooperative tracking in dynamic environments with uncertain localization noise.