Sequential fusion for multi-rate multi-sensor nonlinear dynamic systems with heavy-tailed noise and missing...

Area of Science:

• Control Systems Engineering
• Signal Processing
• Statistical Inference

Background:

• Sequential fusion estimation is critical for multi-sensor nonlinear dynamic systems.
• Heavy-tailed noise and missing measurements pose significant challenges to traditional filtering methods.
• Existing Bayesian inference techniques often struggle with non-stationary noise and computational errors.

Purpose of the Study:

• To develop a robust sequential fusion estimation algorithm for multi-rate, multi-sensor nonlinear systems.
• To address challenges posed by heavy-tailed noise and missing measurements.
• To improve the accuracy and reliability of state estimation in dynamic systems.

Main Methods:

• A sequential Student's t-based unscented Kalman filter (SSTUKF) and its square-root form (SR-SSTUKF) were developed.
• The unscented transform was employed for calculating Student's t weighted integrals.
• Adaptive factors, based on the t-test, were introduced to mitigate uncertainties from non-stationary noise and computational errors.

Main Results:

• The computational complexity and convergence of the SR-SSTUKF were analyzed.
• The proposed method demonstrated validity and robustness in an agile target tracking scenario.
• Simulation results confirmed that the SR-SSTUKF with adaptive factors significantly enhances estimation accuracy.

Conclusions:

• The proposed SR-SSTUKF provides a reliable and accurate sequential fusion estimation method for challenging nonlinear systems.
• Adaptive factors effectively suppress uncertainties, leading to improved performance.
• The method shows promise for applications requiring robust state estimation in dynamic environments.