Risk-Aware and Energy-Efficient AoI Optimization for Multi-Connectivity WNCS with Short Packet Transmissions

TL;DR

This work addresses AoI optimization in multi-connectivity WNCSs with short packets by introducing the EE-PAoI ratio to jointly optimize information freshness and energy. It derives closed-form expressions for average and PAoI-violation performance across non-retransmission, ARQ, and repetition schemes, revealing that more links are not always better and that an SNR-dependent threshold governs connectivity choice. A risk-aware connectivity strategy is proposed, modeling PAoI distribution via a CLFP framework and solving with Dinkelbach iterations to find the optimal number of connections. Simulations in a smart-grid control scenario show substantial EE-PAoI gains (up to 15x) at low SNR and demonstrate the approach’s effectiveness in avoiding risky plant states. The results provide practical guidelines for designing energy-efficient, timely WNCSs under finite-blocklength constraints and fading channels.

Abstract

Age of Information (AoI) has been proposed to quantify the freshness of information for emerging real-time applications such as remote monitoring and control in wireless networked control systems (WNCSs). Minimization of the average AoI and its outage probability can ensure timely and stable transmission. Energy efficiency (EE) also plays an important role in WNCSs, as many devices are featured by low cost and limited battery. Multi-connectivity over multiple links enables a decrease in AoI, at the cost of energy. We tackle the unresolved problem of selecting the optimal number of connections that is both AoI-optimal and energy-efficient, while avoiding risky states. To address this issue, the average AoI and peak AoI (PAoI), as well as PAoI violation probability are formulated as functions of the number of connections. Then the EE-PAoI ratio is introduced to allow a tradeoff between AoI and energy, which is maximized by the proposed risk-aware, AoI-optimal and energy-efficient connectivity scheme. To obtain this, we analyze the property of the formulated EE-PAoI ratio and prove the monotonicity of PAoI violation probability. Interestingly, we reveal that the multi-connectivity scheme is not always preferable, and the signal-to-noise ratio (SNR) threshold that determines the selection of the multi-connectivity scheme is derived as a function of the coding rate. Also, the optimal number of connections is obtained and shown to be a decreasing function of the transmit power. Simulation results demonstrate that the proposed scheme enables more than 15 folds of EE-PAoI gain at the low SNR than the single-connectivity scheme.

Figures (11)

• Figure 1: Multi-connectivity enabled WNCS with short packets.
• Figure 2: The evolution of AoI and queueing process in the multi-connectivity WNCS with transmission diversity. The red 'x' symbol on the left-hand side represents that the packet was dropped due to multiple links being busy. The right red 'x' symbol on the right-hand side represents that the packet transmission failed.
• Figure 3: The average AoI/PAoI vs. transmit power of each connection with different number of connections.
• Figure 4: The PDF and CDF of PAoI with the transmit power $P_\textrm{t}=35$ dBm and $K=4$ connections.
• Figure 5: Impact of the number of connections on the PAoI violation probability with the transmit power $P_\textrm{t}=35$ dBm and different PAoI thresholds.