Research on a Particle Filtering Multi-Target Tracking Algorithm for Distributed Systems.

Area of Science:

• Robotics and Autonomous Systems
• Signal Processing
• Computer Vision

Background:

• The expansion of unmanned aerial vehicle (UAV) applications necessitates sophisticated multi-target tracking systems.
• Traditional tracking methods often assume independent measurements, which is insufficient for distributed systems generating coupled measurements with target relationship data.

Purpose of the Study:

• To propose a novel distributed particle filtering algorithm that incorporates coupled measurements for enhanced multi-target tracking.
• To improve the accuracy and robustness of tracking systems in low-altitude economy applications.

Main Methods:

• Introduced a novel distributed particle filtering algorithm by integrating coupled measurements into the standard particle filtering framework.
• Developed a method to fuse direct and coupled measurements through optimization.
• Constructed a cost function to optimize particle weights based on fused measurements.

Main Results:

• The proposed distributed particle filtering algorithm demonstrated superior performance compared to conventional particle filtering and unscented Kalman filtering.
• Achieved over 7% accuracy improvement across various motion models, noise levels, and target counts.
• Exhibited significant robustness against measurement noise and an increasing number of targets.

Conclusions:

• The novel distributed particle filtering algorithm effectively leverages coupled measurements for improved multi-target tracking.
• The method offers a robust and accurate solution for UAV applications in the low-altitude economy.
• This advancement addresses key limitations of traditional tracking approaches in complex environments.