Endoscopic capsule localization with unknown signal propagation coefficients

Abstract

Wireless capsule endoscopy (WCE) localization techniques have gained increasing popularity for medical diagnostics and treatments, particularly in gastro-intestinal (GI) track, in recent years. In these localization techniques, some parameters such as relative permittivity and path loss coefficients in the human body have significant effects on the accuracy of the capsule localization. Hence, these coefficients need to be accurately known or estimated for the effectiveness of the localization algorithm. This paper proposes a new adaptive biomedical capsule localization scheme utilizing a recent geometric cooperative sensor technique to estimate the path loss coefficient for received signal strength (RSS) and the relative permittivity for time-of-flight (TOF) based range sensors. This cooperative environmental coefficient estimation technique involves use of a mobile sensor triplet in place of a single sensor, and provides instantaneous estimates. The proposed capsule localization scheme is combination of a discrete time recursive least squares (RLS) based adaptive localization scheme and the aforementioned coefficient estimation technique, with design parameters specific for the particular WCE localization application. The accuracy and effectiveness of the proposed scheme is successfully demonstrated via a set of simulation tests.