Evaluation of the Influence of Measurement Noise and Adaptive Gain Adjustment in Kalman Filtering for Estimating Angular Coordinates of Randomly Maneuvering Targets

Abstract

Accurate state estimation in nonlinear dynamic systems is a fundamental requirement in many engineering applications, particularly when system noise characteristics vary over time. Conventional Kalman filters (KF) often suffer from performance degradation under time-varying noise conditions due to their reliance on fixed noise covariance assumptions. The objective of this paper is to present a comprehensive performance analysis of conventional KF and adaptive KF (AKF) for nonlinear state estimation in a fourth-order dynamic system. Methodologically, the study employs a comparative quantitative simulation design, focusing on evaluating estimation accuracy, convergence behavior, and statistical consistency under different process and measurement noise conditions. The proposed framework incorporates an adaptive mechanism for updating noise covariance matrices based on innovation statistics. Simulation results demonstrate that the adaptive KF significantly improves estimation accuracy and stability compared to the conventional KF, particularly in scenarios with rapidly changing noise characteristics. These findings confirm the effectiveness of adaptive filtering strategies and highlight their potential for practical applications in control and navigation systems.