Design and Implementation of Time-Sensitive Wireless IoT Networks on Software-Defined Radio

TL;DR

This paper demonstrates that with proper synchronization and delay management, software-defined radio can be effectively used for time-sensitive industrial IoT applications requiring low latency and precise timing.

Contribution

The authors designed a time-slotted wireless system on USRP SDR platform with a novel Just-in-time algorithm for synchronization, enabling SDR to meet industrial IoT timing requirements.

Findings

90% of time slots synchronized within ±0.05μs

End-to-end latency reduced to 3.75ms

SDR suitable for applications like sensor data collection and AGV control

Abstract

Time-sensitive wireless networks are an important enabling building block for many emerging industrial Internet of Things (IoT) applications. Quick prototyping and evaluation of time-sensitive wireless technologies are desirable for R&D efforts. Software-defined radio (SDR), by allowing wireless signal processing on a personal computer (PC), has been widely used for such quick prototyping efforts. Unfortunately, because of the \textit{uncontrollable delay} between the PC and the radio board, SDR is generally deemed not suitable for time-sensitive wireless applications that demand communication with low and deterministic latency. For a rigorous evaluation of its suitability for industrial IoT applications, this paper conducts a quantitative investigation of the synchronization accuracy and end-to-end latency achievable by an SDR wireless system. To this end, we designed and implemented a time-slotted wireless system on the Universal Software Radio Peripheral (USRP) SDR platform. We developed a time synchronization mechanism to maintain synchrony among nodes in the system. To reduce the delays and delay jitters between the USRP board and its PC, we devised a {\textit{Just-in-time}} algorithm to ensure that packets sent by the PC to the USRP can reach the USRP just before the time slots they are to be transmitted. Our experiments demonstrate that $90\%$ ($100\%$) of the time slots of different nodes can be synchronized and aligned to within $ \pm 0.5$ samples or $ \pm 0.05\mu s$ ($ \pm 1.5$ samples or $ \pm 0.15\mu s$), and that the end-to-end packet delivery latency can be down to $3.75ms$. This means that SDR-based solutions can be applied in a range of IIoT applications that require tight synchrony and moderately low latency, e.g., sensor data collection, automated guided vehicle (AGV) control, and Human-Machine-Interaction (HMI).