Adaptive Resource Allocation and Mode Switching for D2D Networks With Imperfect CSI in AGV-Based Factory Automation

Safiu A. Gbadamosi, Gerhard P. Hancke, Adnan M. Abu‐MahfouzIEEE Open Journal of Vehicular Technology 2024 · DOI

Abstract

In industrial factory automation and control systems, reliable communication for automated guided vehicles (AGVs) in dynamic, interference-laden factory settings is essential, particularly for real-time operations. Device-to-device (D2D) technology can enhance industrial network performance by offloading traffic and improving resource utilization. However, deploying D2D-enabled networks presents challenges such as interference control and imperfect channel state information (ICSI). In this paper, we investigate an adaptive resource allocation and mode switching strategy (ARAMS) in D2D-enabled industrial small cell (SC) networks with ICSI to maximize system throughput and address reuse interference for AGVs. The ARAMS scheme integrates mode switching (MS), channel-quality factor (CQF), and power control (PC) within a bi-phasic resource-sharing (RS) algorithm to lower computational complexity. Numerical results show that as channel estimation error variance (CEEV) increases, the ARAMS scheme consistently outperforms other approaches in maximizing system throughput, with the exception of the semi-centralized AIMS scheme.

Introduction

In the bustling environment of a modern "smart factory," Automated Guided Vehicles (AGVs) are the lifeblood of logistics. Maintaining connectivity for these robots in metal-heavy, interference-rich environments is a significant challenge for wireless engineers. In Industry 4.0, AGVs require ultra-reliable, low-latency communication for collision avoidance and task coordination. While device-to-device (D2D) technology—allowing devices to communicate directly—can offload traffic from the main network, it introduces substantial interference challenges.

A critical hurdle in real-world deployment is Imperfect Channel State Information (ICSI). Most existing research assumes "Perfect CSI," where the base station has an exact, real-time view of signal behavior. In dynamic factories with moving machinery, measurements are inherently noisy. This paper proposes the ARAMS (Adaptive Resource Allocation and Mode Switching) scheme to optimize communication even when the network's view of the environment is fuzzy.

Methodology: The ARAMS Scheme

The ARAMS scheme addresses the NP-hard optimization problem of maximizing system throughput under interference constraints by breaking it into two manageable phases:

Phase 1: Mode Switching and Partner Selection

The system first decides the operational mode for each AGV based on the Channel Gain Ratio (CGR). An AGV can operate in:

  • Cellular Mode: Communicating directly with the small-cell base station.
  • Reuse Mode: Communicating directly with another device while sharing the spectrum.

To minimize mutual interference, a novel Channel Quality Factor (CQF) is calculated for each cell to identify the optimal reuse partner (pairing D2D users with cellular users).

Phase 2: Distributed Power Control

Once modes and pairings are established, the system employs Lagrangian dual decomposition to determine the optimal transmit power for every device. This distributed approach allows individual small-cell base stations to make local decisions, reducing signaling overhead and latency—critical for real-time AGV operations.

Performance Evaluation

The study utilized numerical simulations in MATLAB to model a 500m macro-cell area containing six orthogonal small cells. AGVs and cellular users were distributed via a Homogeneous Poisson Point Process (PPP). The simulation incorporated WINNER I (Line-of-Sight) models for D2D links and varying levels of Channel Estimation Error Variance (CEEV) to test robustness against imperfect information.

Key Results

  • Throughput Gains: ARAMS outperformed standard "Random" and "Power Control (PC)" schemes by 7.14% to 68.90% across various scenarios.
  • Resilience to Error: As channel data became more "imperfect" (higher CEEV), ARAMS maintained a higher system Signal-to-Interference-plus-Noise Ratio (SINR) than most competing approaches.
  • Benchmarking: While the semi-centralized AIMS scheme achieved approximately 10% higher throughput, ARAMS offers a more practical solution for real-world deployment due to its lower computational complexity and distributed nature.

Discussion and Future Work

A key insight from the study is that ARAMS effectively manages the "power boost" often seen as channel estimation error increases, turning a channel imperfection into a managed variable. Furthermore, the mode-switching capability acts as a "safety valve," preserving Quality of Service (QoS) by flipping AGVs to cellular mode when direct link interference becomes excessive.

Future research will look to expand the model beyond one-to-one resource mapping to accommodate high-density environments where multiple D2D pairs share the same resource block. Additionally, incorporating specific AGV movement patterns (e.g., warehouse aisles) and moving toward hardware validation using Software Defined Radios (SDRs) are essential next steps.

Conclusion

The ARAMS scheme provides a robust framework for D2D-enabled industrial networks. By intelligently balancing mode switching and power control, it ensures high throughput and reliability for AGVs in the challenging, dynamic environments of modern automated factories.

Summary updated

Abstract

In industrial factory automation and control system, reliable communication for automated guided vehicles (AGVs) in dynamic, interference laden factory settings are essential particularly for real-time operations. Device-to-device (D2D) technology can enhance industrial network performance by offloading traffic and improving resource utilization. However, deploying D2D-enabled networks presents challenges such as interference control and imperfect channel state information (ICSI). In this paper, we investigate an adaptive resource allocation and mode switching strategy (ARAMS) in D2D-enabled industrial small cell (SC) networks with ICSI to maximize the system throughput and address reuse interference for AGVs. The ARAMS scheme integrates mode switching (MS), channel-quality factor (CQF), and power control (PC) within a bi-phasic resource-sharing (RS) algorithm to lower the computational complexity. In the initial phase, the operational mode for each D2D user (DU) per cell is adaptively selected based on the channel gain ratio (CGR). Subsequently, it computes the CQF for each cell with a reuse DU to identify an optimal reuse partner. The final phase employs the Lagrangian dual decomposition method to decide the DU's and industrial cellular users (CUs) optimum distributed power to maximize the system throughput under the interference constraints. The numerical results show that as channel estimation error variance (CEEV) increases, the ARAMS scheme consistently outperforms other approaches in maximizing system throughput, except for the AIMS scheme.
BibTeX
@article{gbadamosi2024adaptive,
  author = {Safiu A. Gbadamosi and Gerhard P. Hancke and Adnan M. Abu‐Mahfouz},
  title = {{Adaptive Resource Allocation and Mode Switching for D2D Networks With Imperfect CSI in AGV-Based Factory Automation}},
  journal = {IEEE Open Journal of Vehicular Technology},
  year = {2024},
  doi = {10.1109/ojvt.2024.3519135},
}
Figure