The Complexity-Performance Tradeoff in Resource Allocation for URLLC Exploiting Dynamic CSI

TL;DR

This work tackles the complexity-performance tradeoff in URLLC resource allocation for dense IIoT by integrating dynamic, time-correlated CSI with a graph-based allocation framework. It extends the Graph-Based Allocation Algorithm (GBA) to account for in-band pilot signaling and proposes a dynamic pilot transmission strategy to adapt the CSI age per device and channel, guided by a best-improvement heuristic. Through analysis and simulations, the authors show that GBA substantially outperforms a greedy baseline in dense networks and that dynamic pilot scheduling yields additional gains in fairness and spectrum efficiency, even when accounting for computational delays. The findings indicate that scalable, CSI-aware scheduling is feasible for URLLC in large-scale industrial environments, enabling more reliable, low-latency communications with rich channel information.

Abstract

The challenging applications envisioned for the future Internet of Things networks are making it urgent to develop fast and scalable resource allocation algorithms able to meet the stringent reliability and latency constraints typical of the Ultra Reliable, Low Latency Communications (URLLC). However, there is an inherent tradeoff between complexity and performance to be addressed: sophisticated resource allocation methods providing optimized spectrum utilization are challenged by the scale of applications and the concomitant stringent latency constraints. Whether non-trivial resource allocation approaches can be successfully applied in large-scale network instances is still an open question that this paper aims to address. More specifically, we consider a scenario in which Channel State Information (CSI) is used to improve spectrum allocation in a radio environment that experiences channel time correlation. Channel correlation allows the usage of CSI for longer time before an update, thus lowering the overhead burden. Following this intuition, we propose a dynamic pilot transmission allocation scheme in order to adaptively tune the CSI age. We systematically analyze the improvement of this approach applied to a sophisticated, recently introduced graph-based resource allocation method that we extend here to account for CSI. The results show that, even in very dense networks and accounting for the higher computational time of the graph-based approach, this algorithm is able to improve spectrum efficiency by over 12% as compared to a greedy heuristic, and that dynamic pilot transmissions allocation can further boost its performance in terms of fairness, while concomitantly further increase spectrum efficiency of 3-5%. \

Figures (11)

• Figure 1: Example of resource allocation on a channel partially used for pilot transmission. Red squares are already allocated RUs, hence $\beta_c=3$ for this channel. Grey squares are the pRUs reserved to pilot signaling, and cannot be allocated. We assume that for user $D_i$ the number of required RUs is $\mathcal{F}(c,i,\rho)=4$, and that the data packet is issued at time slot $t_i=3$. Hence, the RUs up to slot 9 are assigned to $D_i$, with $\zeta=6$.
• Figure 2: Conditional pdf of $|h_{i,c,m+t}|^2$ when $|h_{i,c,m}|^2=1.5$, for different values of the temporal distance $t$.
• Figure 3: Needed resources as a function of the CSI age, for different values of the measured channel fading coefficient. Here the device is placed at a distance $d=40$ m from the AP, and the parameters in Table \ref{['tab:Sympara']} are used.
• Figure 4: Example of pRUs allocation. Here, each cycle has $T=8$ slots, and $\eta=0.25$, meaning that $M=2$ devices can send a pilot over each channel at a given cycle in a round robin fashion, as shown with the numbers, using the pRUs (reported in colors). The computational time $\omega_c$ is lower than a cycle duration, so $W=2$ and the information collected at cycle $m$ is used for the allocation of the data RUs (in white) on cycle $m+2$.
• Figure 5: Fraction of served devices as a function of the fraction $\eta$ of pilot RUs, for various values of the computational delay $W$. Here, $N=100$, $\gamma=0.95$, $T=50$, $C=5$, $\Delta = 25$.

Theorems & Definitions (2)

• Lemma 1
• proof