Physical Limitations to the Accuracy of Range-based Localization and Sensing

Abstract

As we approach higher and higher localization and sensing accuracies with 6G technologies, the radio capabilities and signal processing will cease to be the main accuracy limitations, and other physical phenomena start to play an increasingly critical role. In this paper, we investigate how radio-based physical measurements, including radar, wireless localization and Integrated Sensing and Communication (ISAC), can be affected by environmental conditions, and at what level of accuracy and dimensional and temporal scales these effects become significant. Specifically, we produce and publish a dataset where the long term stability of radio-based range measurements is empirically assessed, showing peak-to-peak variations of 0.3 mm over a 1 m range (which scales linearly, e.g. 3 mm over 10 m or 3 cm over 100 m) over 6 months. Further, we provide and substantiate a mechanistic explanation of these variations by showing how the combination of changes in refractive index of the air and thermal expansion of the room matches the measured data. We further demonstrate how such effects can be compensated for by measuring the atmospheric conditions and exploiting a combination of physical modeling and machine learning.