A Real-World 5G Industrial Testbed: Hardware Simulation of Slices and Massive IoT

Abstract

Translate Article

A feature of 5G communications is high capacity conveyance of sensor information, paramount for big data applications. However the inconsistent roll-out of 5G features hinders the development of ultra-reliable low-latency communication (URLLC), massive machine-type communication (mMTC) and dynamic network slicing. This paper introduces a hardware-based 5G test-bed that emulates dynamic slicing and large-scale sensor data processing using AI-powered resource optimization. Taking advantage of current 5G infrastructure and surrogate sensor data generation, our system enables realistic big data analytics in industrial environments. An AI-based decision tree model dynamically assigns network slices, optimizing throughput costs. Our surrogate data provides good quality supplemental sensor data, preserving key statistical and spectral properties enhancing sensor-driven big data applications. Our experiments show improved data transfers, adaptive network slicing and scalable big data processing. This offers a cost-effective solution for small/medium-sized enterprises (SMEs) and research/development (R&D) teams. This work bridges the gap between theoretical 5G slicing and real-world deployment for 5G-based big data applications. It provides a repeatable framework for industrial IoT and smart manufacturing, preparing for full commercial availability of 5G features<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup><sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup>Ministry of Education Singapore, grant 2021MOE-IF-024.