Experimental Evaluation of the Impact of Physical Space Information on Wireless Link Quality Estimation

Abstract

Predicting wireless link quality (LQ) is a promising approach to enhancing Quality of Experience (QoE). This paper examines the impact of physical space on LQ; it uses machine learning to predict RSSI and throughput from physical space information. Experiments show that the position/orientation of user equipment (UE) are crucial, with orientation affecting RSSI, and velocity significantly impacting throughput. It is revealed that these LQ parameters have different spatial factor dependencies. An experiment is conducted in a static indoor environment with clear line of sight to simplify the problem. We focus on deriving LQ from current physical space information rather than forecasting future quality.