Recursive Error Division Kalman Filter for Advanced Driver Assistance Radar Systems

Abstract

This paper presents a time-domain analysis for digital Infinite Impulse Response filters used in tracking applications. It provides information on how such a filter may be used as discrete-time estimator, and even more so, offer valuable information on the rate of change of the input signal. An example of an application where the proposed acquisition-synchronous filter is useful, is in automotive radar-based driver assistance systems. The filter is responsible for estimating the velocity and acceleration of the tracked targets, which are then fed into a range- only Kalman Filter Tracker. The proposed algorithm fusion not only reduces computational complexity, but also improves performance. Unlike alpha-beta-gamma filters, the present method can also be extended to non-Gaussian distributions for Extended Kalman Filters.