5G-TSN/DetNet Integrated Networks for Industrial IoT: A Machine Learning Framework to Balance Reliability and Bandwidth Utilization

In the growing field of the Industrial Internet of Things (IIoT), the integration of 5 G networks with TimeSensitive Networking (TSN) and Deterministic Networking (DetNet) promises unprecedented reliability in data transmission. This paper introduces a Machine Learning (ML)-based methodological framework to optimize the balance between network reliability and bandwidth utilization in 5G-TSN/DetNet integrated networks. Our framework enhances network performance by dynamically de/activating the TSN/DetNet reliability functions based on real-time analysis of 5 G radio conditions. We employ a clustering algorithm to categorize radio measurements into Green, Orange, and Red conditions, with a hysteresis logic to ensure stable state transitions and avoid premature or unnecessary TSN/DetNet reliability toggling. Proof-of-concept experimental results obtained from a 5 G testbed emulating an IIoT device with dual 5G User Equipment demonstrate the benefits of our approach. Specifically, we show that the framework can maintain high data transmission reliability with zero packet loss, even under fluctuating network conditions, while efficiently utilizing available bandwidth. The findings underscore the potential of ML in automating and enhancing decision-making processes in next-generation industrial networks, paving the way for more resilient and adaptive communication for IIOT communications. The integration of such frameworks into actual 3GPP systems is suggested as a future direction for real-world application.