QoS Predictability in V2X Communication with Machine Learning

摘要: An important use case in fifth generation systems are vehicular applications, where, reliability and low latency the main requirements. In order to determine if a application can be used one apply machine learning (ML) tools these constraints met, which open questions such as “which data is available”, features use”, “the quality of this prediction”, etc. paper we address some aspects predicting quality-of-service (QoS) cellular vehicular-to-everything scenario, where employ supervised well autoregressive integrated moving average filter predict packet delivered within desired window. Particularly, interested prediction, including generated time ahead will time. Such information essential when asserting that indeed employed safely. We show via simulation results ML prediction tool applications. For instance, QoS levels predicted two seconds with 85 % reliability.