Dynamic Latency Optimization for Edge-Based Machine Learning Models in 6G-Enabled Industrial Internet of Things (IIoT)

The integration of 6G technology into the Industrial Internet of Things (IIoT) promises to redefine manufacturing through "Hyper-Reliable Low-Latency Communication" (HRLLC). However, the deployment of complex Machine Learning (ML) models at the edge remains constrained by the heterogeneous nature of industrial data and the limited computational resources of edge nodes. This article proposes a novel framework for Dynamic Latency Optimization (DLO) that leverages Deep Reinforcement Learning (DRL) for intelligent task offloading and resource allocation. By utilizing 6G's Terahertz (THz) spectrum and AI-native Network Slicing, the proposed framework dynamically adapts to fluctuating network conditions to maintain sub-millisecond latency. Our simulation results demonstrate a 42% reduction in end-to-end delay and a 30% improvement in energy efficiency compared to traditional 5G-MEC architectures. Furthermore, we explore the integration of Reconfigurable Intelligent Surfaces (RIS), Semantic Communication, and Zero-Trust Edge Security to further optimize the data-intelligence pipeline for Industry 5.0 applications, focusing on the critical synergy between human operators and autonomous systems within a resilient, sustainable, and cognitively aware industrial fabric. Keywords: 6G Networks, Industrial IoT (IIoT), Edge Intelligence, Deep Reinforcement Learning, Latency Optimization Corresponding author: Seema Patil*. 📧 Email: seemapatil.cse@bmsce.ac.in ORCID:0000-0001-6490-3020 1. Introduction: From Automation to Human-Centric Intelligence The transition from Industry 4.0 to Industry 5.0 marks a profound shift toward human-centric, resilient, and sustainable manufacturing systems. While Industry 4.0 was characterized by the digitalization of physical assets and the rise of cyber-physical systems, Industry 5.0 emphasizes the "Tactile Internet" and "Human-Robot Co-evolution." In this new paradigm, the focus shifts from pure efficiency to the seamless collaboration between humans and increasingly autonomous machines. The "Tactile Internet" concept is particularly revolutionary, as it requires a "haptic control loop"—the ability to transmit touch and feel sensations over the network with such low latency that the human brain perceives no delay. This necessitates an end-to-end latency below 1ms, encompassing both the transmission and the computational processing of sensory feedback. This evolution necessitates a communication infrastructure capable of supporting advanced applications such as ultra-responsive autonomous mobile robots (AMRs), synchronized multi-robot assembly lines, and high-fidelity haptic feedback for remote maintenance in hazardous environments. For example, a specialist surgeon operating a robotic arm in a factory cleanup of toxic waste requires instantaneous haptic feedback to "feel" the resistance of the materials being handled. If the feedback loop exceeds 10ms, the mismatch between visual and tactile input can lead to "operator sickness" or mechanical errors that jeopardize safety. Furthermore, we must consider proprioceptive alignment—the sense of self-movement and body position. In 6G-enabled IIoT, the network must act as an extension of the human nervous system, where the delay jitter is so minimal that the robotic actuator feels like a literal extension of the operator's limb. This requires not just low latency, but Isochronous Communication, where packets arrive at precisely regular intervals to maintain the temporal rhythm of human motor-sensory systems. This synchronization is critical for Tele-Operation in nanomanufacturing, where even a micro-stutter in the feedback loop can cause the robotic probe to crush a microscopic wafer. The biological threshold for "instantaneous" feedback in human motor control is roughly 1-10ms for tactile sensations and less than 1ms for the suppression of "visual-vestibular conflict." In 6G, we move into the regime of "Sub-Perceptual Jitter," where the network variance is lower than the biological noise of the human nervous system. This enables "Neuromorphic Manufacturing," where robotic limbs are controlled via high-speed neural interfaces, requiring the network to behave as a synthetic synaptic bridge. While 5G introduced Ultra-Reliable Low-Latency Communication (URLLC), providing latencies in the range of 1–5ms, 6G aims to push these boundaries into a new regime. The 6G vision targets "microsecond-level" latencies (below 0.1ms) and peak data rates exceeding 1 Tbps. This leap is not merely incremental; it enables a "Digital Twin of Everything," where the physical and digital worlds are perfectly synchronized in real-time. In a precision nanotech assembly line, a delay of just 2ms could result in a mechanical misalignment of several millimeters, potentially ruining a high-value batch of semiconductors or causing catastrophic failure in medical implant manufacturing. Furthermore, 6G is envisioned as an "AI-Native" network, where the infrastructure itself is composed of intelligent agents that optimize transmission parameters in real-time without human intervention. The core challenge in modern IIoT environments remains the "Intelligence Bottleneck"—the cumulative delay incurred when high-dimensional ML inference tasks are offloaded to resource-constrained edge nodes. In a high-speed production line where components move at several meters per second, a delay of even 5ms can lead to synchronization errors between robotic actuators. Consequently, achieving deterministic, sub-millisecond latency requires a paradigm shift: moving away from static resource management toward intelligent, context-aware management of both communication and computational resources. This involves not just faster radios, but smarter ways to decide where and how data should be processed, often referred to as Compute-Communication Integration (CCI). We term this the "Unified Intelligence-Control Loop," where the network is not an external utility but an intrinsic part of the machine's control logic. This loop is further refined by "Semantic Pruning," where AI agents at the source decide which 1% of data carries 99% of the mission-critical meaning, dramatically reducing the "Air-Time" required for transmission. 2. 6G Technical Enablers for Edge Intelligence To support the proposed DLO framework, several 6G-specific technologies are integrated into our system model, addressing the unique propagation and computation challenges of industrial halls. 2.1 Terahertz (THz) Spectrum and Sub-THz Communication The utilization of the THz band (0.1–10 THz) offers massive bandwidth (hundreds of GHz), enabling the near-instantaneous transmission of uncompressed high-definition video feeds for AI-driven inspection. While THz waves suffer from high atmospheric attenuation and limited range, they are ideal for the "pico-cell" deployments characteristic of indoor industrial environments. At these ranges, the "Oxygen Absorption" and "Water Vapor Resonance" spikes that plague long-distance THz communication are negligible, allowing for stable links. We specifically target the "Atmospheric Windows"—narrow frequency bands where molecular absorption is at its local minima—to ensure signal robustness. However, in industrial settings, we must account for Scintillation and Turbulence. Large industrial furnaces or chemical reactors can create localized pockets of high-temperature air, causing fluctuations in the air's refractive index (). This leads to "Beam Wandering," where the pencil-thin THz beam misses the target. Our model incorporates a Turbulence-Induced Fading coefficient, modeled using the Gamma-Gamma distribution, to ensure the DLO agent can predict and compensate for these environmental "flickers." The Molecular Absorption Loss () is calculated using the HITRAN database. We model the THz channel as: Where is the absorption coefficient and is the turbulence-induced fading factor. Unlike the Sub-6GHz bands, THz signals exhibit "Quasi-Optical" properties, meaning they behave more like light than traditional radio waves. Specifically, we utilize the D-band (110–170 GHz), which provides a middle ground between high data rates and manageable propagation loss. Furthermore, the short wavelengths of THz allow for Terascale MIMO—antenna arrays with tens of thousands of elements. These enable "Near-Field Beamforming," where the beam can be focused onto a specific 3D coordinate (a "focal spot") rather than just a direction. This provides unprecedented spatial reuse, allowing multiple robots in the same vicinity to operate on the same frequency without mutual interference. This "Focus-Spot" approach also minimizes the Electromagnetic Interference (EMI) on sensitive biological tissues of human workers, a key safety requirement for Industry 5.0. 2.2 Reconfigurable Intelligent Surfaces (RIS) Industrial environments are notoriously cluttered with metallic machinery that causes multi-path interference and frequent signal blockage (shadowing). To solve this, we incorporate RIS—programmable meta-surfaces that can "steer" radio waves—to bypass physical obstacles. An RIS consists of a thin sheet of electromagnetic material embedded with sensors and micro-controllers that can change the phase of an incoming wave. In our implementation, the RIS acts as a "Smart Mirror." If a large metal press blocks the direct path between a robot and an edge server, the RIS identifies the blockage and reflects the signal around the corner. The RIS consists of an N*M array of passive scattering elements. Mathematically, the RIS is modeled as a diagonal matrix. This creates "Constructive Interference" at the receiver's location. Beyond passive reflection, we integrate "Simultaneous Transmitting and Reflecting" (STAR-RIS). A STAR-RIS can reflect signals to devices in front of it while transmitting (refracting) signals to devices behind it. Thi