A Zero Velocity Detector for Foot-mounted Inertial Navigation Systems Aided by Downward-facing Range Sensor

Abstract

In this paper, we propose a zero velocity detector, referred to as the Ultrasound-Aided Stance Hypothesis Optimal dEtection (UA-SHOE), for foot-mounted Inertial Navigation Systems (INS) aided by a downward-facing range sensor. The proposed detector is derived mathematically in a Generalized Likelihood Ratio Test (GLRT) framework. We compare the effectiveness of the proposed UA-SHOE detector with two known and commonly used approaches, the Stance Hypothesis Optimal dEtection (SHOE) detector [1] and the Ultrasonic-only Stance Phase Detection (USPD) detector [2]. The Circular Error Probable (CEP) and the Root Mean Square Error (RMSE) of the navigation results are used to evaluate the three detectors. The experimental results of walking and running cases showed that the UA-SHOE detector improved errors by 27.5% and 11.3%, as compared to the USPD detector. The experimental results also illustrated that the RMSEs of the proposed UA-SHOE detector had a comparable accuracy to the commonly used SHOE detector in the walking case and outperformed the SHOE detector by more than 50% in the running experiments.