Estimating the Likelihood of Wireless Coexistence Using Logistic Regression: Emphasis on Medical Devices.

ABSTRACT

Medical device manufacturers incorporate wireless technology in their designs to offer convenience and agility to both patients and caregivers. However, the use of unlicensed radio spectrum bands by both medical devices and other equipment raises concerns about wireless coexistence. Work by the accredited standards committee C63 of the American National Standards Institute (ANSI) to provide the community with a consensus standard for coexistence evaluation resulted in the publication of the ANSI C63.27 standard, which was later recognized by the U.S. Food and Drug Administration. Estimating the likelihood of wireless coexistence of a system under test (SUT) in a given environment is central to the evaluation and reporting of wireless coexistence, as made clear in the C63.27 standard. However, no method to perform this estimation is provided. In this paper, we propose the use of logistic regression (LR) to estimate the likelihood of wireless coexistence of a medical device in its intended environment. Radiated open environment coexistence testing was used to realize a test scenario in which the interfering network was IEEE 802.11n Wi-Fi and the SUT was ZigBee; exemplary wireless technologies for interfering network and medical device, respectively. LR model fitting was then performed to derive a model that describes the performance of SUT under a range of wireless coexistence phenomena. Finally, results were incorporated with the outcome of a spectrum survey using Monte Carlo simulation to estimate the SUT likelihood of wireless coexistence in a hospital environment.