Novel Markov Chain Based URLLC Link Adaptation Method for 5G Vehicular Networking

Abstract

Advanced driver-assistance system features in current vehicles (as precursor to fully autonomous, connected vehicular systems) will rely on 5G network links. Specifically, enabling various real-time operational aspects (e.g. emergency messaging), performance of ultra-reliable low-latency communications (URLLC) specified by 3GPP Release 16 is of direct relevance. A key challenge in achieving URLLC goals is the natural tension between simultaneously achieving high reliability (low block error rate or BLER) and low latency. When operating in a dynamic (i.e. significant time-varying channels due to vehicular mobility) environment, link adaptation is fundamental to URLLC implementation. In this paper, we introduce a new Markov Chain model for link adaptation that predicts the three key performance indicators (KPIs): end-to-end link latency, throughput and BLER as a function of the modulation and coding scheme used. The predictions from the Markov Chain model are supported by Monte Carlo simulations for various mobility scenarios, to explore how 5G network parameters may be adapted to achieve desired benchmarks, e.g. maximizing link throughput while providing strict latency and BLER bounds.