Blind Robust Multi-Horizon EFIR Filter for Tightly Integrating INS and UWB

Abstract

In this paper, a novel model to localize a human accurately in indoor industrial environments by integrating the inertial navigation system (INS) technique with ultra wide band (UWB) technology is present. The distances between the target human and the reference node measured from INS and UWB are used by the fusion filter. To improve accuracy, the robust extended finite impulse response (EFIR) filtering algorithm is developed to operate on multi horizons (MHs), thereby providing blind suboptimal estimates. Experimental testing has shown that the real-time blind MH EFIR algorithm is almost as accurate as the EFIR filter, which requires a preliminarily determined optimal horizon.