Smart Timing Synchronization for Small Data Transmission

TL;DR

This paper proposes machine learning-based timing synchronization methods for mobile IoT devices in 5G systems, enabling reliable small data transmission without traditional random access procedures.

Contribution

It introduces novel ML-aided techniques for validating and predicting timing advance in mobile UEs, enhancing the applicability of configured grant small data transmission.

Findings

High accuracy in TA prediction across various environments

Effective validation of timing advance for mobile UEs

Improved readiness for small data transmission in 5G IoT

Abstract

Cellular Internet-of-things (C-IoT) user equipments (UEs) typically transmit periodic but small amounts of uplink data to the base station. To avoid undergoing a traditional random access procedure prior to every transmission, 5th generation (5G) and newer systems use configured grants for small data transmission (CG-SDT), which is equivalent to its long-term evolution (LTE) counterpart of preconfigured uplink resources (PURs)-based transmission. CG-SDT configures uplink resources to UEs in advance for transmission without a random access procedure. A prerequisite for CG-SDT is that the UEs must use a valid timing advance (TA). This is done by validating a previously held TA before CG-SDT. While this validation is trivial for stationary UEs, mobile UEs often encounter conditions where the previous TA is no longer valid and a new one is to be requested by falling back to legacy random access procedures. This limits the applicability of CG-SDT in mobile UEs. To this end, we propose UE-native smart timing synchronization techniques to counter this drawback and ensure a near-universal adoption of CG-SDT. We introduce new machine learning-aided solutions for validation and prediction of TA for UEs with any type of mobility. We perform comprehensive simulation evaluations across different types of communication environments to demonstrate the effectiveness of our proposed solution in predicting the TA.