Intelligent Bi-directional Relaying Communication for Edge Intelligence based Industrial IoT Networks

Jun Liu, Yuwei Zhang, Jing Wang, Tao Cui, Lin Zhang, Chao Li, Kai Chen, Huang Huang, Xuan Zhou, Wei Zhou, Zhao Wang, Li Sun, Suili Feng, Dongqing Xie, Dahua Fan, Jianghong Ou, Jiangtao Ou, Yun Li, Haige Xiang, Kaimeno Dube, Abbarbas Muazu, Nakilavai Rono, Yajuan TangEAI 2022 · DOI

Abstract

This research investigates the performance of two-way relay networks (TWRN) utilizing multiple amplify-and-forward (AF) relays in industrial IoT (IIoT) environments. We specifically examine the "best relay selection" strategy under the realistic constraint of outdated channel state information (CSI). By deriving closed-form lower bounds for outage probability and symbol error rate (SER), and performing asymptotic analysis at high signal-to-noise ratios (SNR), we demonstrate a critical performance bottleneck: diversity collapse. Our analysis shows that even slight staleness in CSI reduces the system's diversity order to unity, effectively negating the reliability benefits of multiple relays. These theoretical findings are corroborated by Monte Carlo simulations, providing essential design constraints for edge intelligence in dynamic industrial environments.

Introduction

In the rapidly evolving landscape of the Industrial Internet of Things (IIoT), "Edge Intelligence" relies on fast, reliable data exchange between sensors, actuators, and controllers. Two-way relay networks (TWRN) have emerged as a popular solution because they allow two source nodes to exchange information simultaneously via an intermediate relay, significantly improving spectral efficiency compared to traditional one-way relaying.

However, most theoretical models assume "perfect" Channel State Information (CSI)—meaning the system knows exactly how the signal will fade at the moment it picks a relay. In real-world industrial environments, factors like Doppler shifts (caused by moving machinery), feedback delays, and estimation errors mean that by the time a relay is selected and used, the channel conditions have already changed. This paper quantifies exactly how much this "staleness" or "outdated CSI" degrades system performance, specifically focusing on outage probability and symbol error rate (SER).

System Model and Relay Selection

The authors model a system with two source nodes (A and B) and $N$ Amplify-and-Forward (AF) relays. To maximize efficiency, the system employs a "best relay selection" strategy, where the relay that provides the highest minimum signal-to-noise ratio (SNR) for both communication directions is chosen.

Figure 1: System model of a two-way relay network illustrating the two-phase communication process between nodes A and B via a selected relay $R_n$.

As shown in Figure 1, the communication occurs in two distinct phases. In the first phase, both source nodes transmit their signals to the relay cluster simultaneously. In the second phase, the selected relay $R_n$ amplifies the received signal and broadcasts it back to both sources. Each source then subtracts its own transmitted signal to decode the message from the other node.

Methods

To analyze the impact of outdated CSI, the authors employed a rigorous mathematical framework:

  • Channel Correlation Modeling: They used a correlation coefficient ($\varpi$) to represent the relationship between the CSI at the time of selection and the CSI at the time of actual transmission. A value of $\varpi = 1$ represents perfect CSI, while $\varpi = 0$ represents completely obsolete information.
  • Derivation of Lower Bounds: The team developed closed-form mathematical expressions for the lower bounds of both Outage Probability and Symbol Error Rate (SER) under Rayleigh fading conditions.
  • Asymptotic Analysis: By evaluating the system's behavior as the transmit power ($P$) approaches infinity, they calculated the "diversity order"—a key metric for how robust a system is against fading.
  • Monte Carlo Simulations: These simulations were used to validate the theoretical derivations, ensuring the mathematical lower bounds were tight and accurate across various SNR ranges.

Key Insights: The Diversity Collapse

The most significant finding of this study is the phenomenon of diversity collapse. In a perfect world, having $N$ relays should provide a diversity order of $N$, meaning the error rate drops sharply as power increases. However, the authors prove that if the CSI is even slightly outdated ($\varpi < 1$), the diversity order drops to 1.

This implies that in the presence of feedback delays, adding more relays does not improve the slope of the reliability curve at high power; it only provides a marginal shift in performance. When CSI is completely obsolete ($\varpi = 0$), the selection process becomes essentially random, and the performance of a 5-relay system becomes indistinguishable from a single-relay system.

Key Results

  • Outage Probability: The paper provides a complex but exact formula for the Cumulative Distribution Function (CDF) of the network gains, which is used to calculate the probability that the system fails to meet a required SNR threshold.
  • Asymptotic Performance: At high transmit power, the outage probability and SER are shown to be inversely proportional to the SNR ($P$) rather than $P^N$, confirming the loss of diversity.
  • Tightness of Bounds: The derived lower bounds are remarkably rigid, aligning almost perfectly with simulation results at high SNR. This provides engineers with reliable formulas to predict system failure without needing exhaustive simulations.

Conclusion and Future Work

This study highlights a critical vulnerability in relay-based IIoT networks. The collapse of diversity order due to outdated CSI suggests that edge intelligence systems must prioritize low-latency feedback or more robust selection algorithms to maintain high reliability.

While this study assumes a Rayleigh fading environment, future research may explore Rician fading to better model "Line-of-Sight" components common on industrial floors. Additionally, the impact of outdated CSI on Decode-and-Forward (DF) relays and multi-way directional relaying remains an open area for investigation.

Summary updated

Abstract

Within this specific record, our group study the two-way interact body (TWRN) that has a number of amplify-and-forward (AF) relays. In that, the best one is actually really got to help the info communication among sources. A interact option is actually really according to the obsolete channel problem information (CSI) in addition to our group analyze its own very personal effect on the system effectiveness in the Rayleigh fading atmospheres. Especially, we extremely preliminary acquire a restricted decreased connected for the outage opportunity and afterward current an asymptotic assessment for greater signal-to-noise ratio (SNR). Our group extra acquire a restricted decreased connected along with an asymptotic result on the authorize error cost (SER). Originating got via these results, our group easily quickly obtain that body system range order remain at unity offered that the CSI is actually really obsolete. Relative results reveal the rigidness on the effectiveness bounds along with the effects of obsolete interact option on the body system effectiveness. Simulation outcomes are likewise offered to corroborate the scholastic evaluation.
BibTeX
@inproceedings{liu2022intelligent,
  author = {Jun Liu and Yuwei Zhang and Jing Wang and Tao Cui and Lin Zhang and Chao Li and Kai Chen and Huang Huang and Xuan Zhou and Wei Zhou and Zhao Wang and Li Sun and Suili Feng and Dongqing Xie and Dahua Fan and Jianghong Ou and Jiangtao Ou and Yun Li and Haige Xiang and Kaimeno Dube and Abbarbas Muazu and Nakilavai Rono and Yajuan Tang},
  title = {{Intelligent Bi-directional Relaying Communication for Edge Intelligence based Industrial IoT Networks}},
  booktitle = {EAI},
  year = {2022},
  doi = {10.4108/eetinis.v9i32.1909},
}
Figure