Online Bias Estimation for Single-Platform Airborne Radar Using Bias-Subspace Information-Guided MAP-EKF

Area of Science:

• Aerospace Engineering
• Signal Processing
• Control Systems

Background:

• Systematic measurement biases degrade airborne radar target tracking accuracy.
• Online bias estimation is challenging in single-platform operations due to coupled states and nonlinear models.
• Existing methods like direct augmentation or fixed batch optimization have limitations in convergence and update reliability.

Purpose of the Study:

• To develop an effective online bias estimation method for single-platform airborne radar.
• To improve target state and radar bias estimation, particularly under weakly observable conditions.
• To enhance the reliability and efficiency of bias updates in recursive filtering.

Main Methods:

• Proposed a Bias Subspace Information-Guided MAP-EKF (BI-MAP-EKF) algorithm.
• Augmented the state to include target motion and range, azimuth, and elevation biases.
• Utilized posterior Cramér-Rao lower bound (PCRLB) and Schur complement for bias subspace extraction and update reliability assessment.
• Implemented a scheduler with cooldown and maximum-window safeguards for adaptive updates.

Main Results:

• The BI-MAP-EKF significantly improved azimuth bias and horizontal position estimation in weakly informative geometries.
• The method reduced the number of accepted MAP refinements compared to fixed-period MAP and MHE baselines.
• In stronger maneuvering scenarios, it offered a competitive accuracy-cost trade-off.

Conclusions:

• The proposed BI-MAP-EKF scheduler is effective for online bias estimation in single-platform airborne radar.
• It demonstrates particular strength in weakly observable geometries, improving tracking performance.
• The method provides a balanced approach to accuracy and computational cost in various maneuvering conditions.