Object target tracking using the alpha sliding innovation filter

The Sliding Innovation Filter (SIF) is an estimation technique developed in 2020 to provide a robust method for estimating a system’s parameters and states in the presence of high modeling uncertainties. This filter ensures that the estimates remain close to the true trajectories. The Alpha-SIF (aSIF), a variant of the SIF introduced in 2022, aims to further smooth the estimates by mitigating the effects of measurement noise. In this work, the aSIF is used to track a ground vehicle navigating within a 2D environment. The angle of maneuver is also estimated using a linearized model that differs from the actual nonlinear model, highlighting the modeling uncertainties. Both measurement and system noise are considered significant, resulting in low signal-to-noise ratios ranging from 15 to 52. The results are compared with the original SIF in terms of Root Mean Square Error (RMSE), Maximum Absolute Error (MAE), and Simulation Time (ST). Findings indicate significant improvements of 12.56% to 49.11% in RMSE and 18.07% in ST, while an 8.07%-13.51% improvement is observed in MAE for the first three states after excluding the error in the initial values.