Variable structure bidimensional reference pattern-based anti-bias track to track association

Abstract

Aiming at distributed sensors with nonlinear biases, this paper proposes an anti-bias track-to-track association algorithm named variable structure bidimensional reference pattern (VSBRP). Firstly, we theoretically analyze the limitations of the original reference-pattern (REP) in dealing with sensor nonlinear bias. Subsequently, to eliminate the adverse effects of sensor bias on track-to-track association, we reformulate the reference-pattern feature by organizing the position and velocity information, with the aim of estimating the relative bias using multiple dimensions of information. Additionally, we propose to regulate the structure of REP in real-time according to missed and false tracks, thereby improving the robustness of the target tracking system. Numerical simulations are performed to demonstrate the effectiveness and superiority of the proposed method.