Regret Analysis with Almost Sure Convergence for OBF-ARX Filter

This paper considers the output prediction problem for an unknown Linear Time-Invariant (LTI) system. In particular, we focus our attention on the OBF-ARX filter, whose transfer function is a linear combination of Orthogonal Basis Functions (OBFs), with the coefficients determined by solving a least-squares regression. We prove that the OBF-ARX filter is an accurate approximation of the Kalman Filter (KF) by quantifying its online performance. Specifically, we analyze the average regret between the OBF-ARX filter and the KF, proving that the average regret over N time steps converges to the asymptotic bias at the speed of $O\left(N^{-0.5+\epsilon}\right)$ almost surely for all $\epsilon \gt 0$. Then, we establish an upper bound on the asymptotic bias, demonstrating that it decreases exponentially with the number of OBF bases, and the decreasing rate $\tau(\boldsymbol{\lambda}, \mu)$ explicitly depends on the poles of both the KF and the OBF. Numerical results on diffusion processes validate the derived bounds.