The advent of sixth-generation (6G) mobile networks forecasts the widespread deployment of edge artificial intelligence (AI), where AI inference tasks are offloaded from resource-constrained edge devices to edge servers. This paradigm promises low latency and high-efficiency processing, yet faces critical challenges in meeting the stringent latency requirements of emerging applications such as autonomous driving and real-time robotics. Traditional ultra-reliable and low-latency communication (URLLC) paradigms fall short in the context of edge inference, where the high-dimensional nature of extracted features introduces a fundamental trade-off between reliability and latency. In this paper, we exploit the inherent robustness of AI models to channel distortions to design an ultra-low-latency edge inference framework that jointly optimizes computation and communication resources. We focus on both multi-snapshot (sequential sensing) and multi-view (distributed sensing) scenarios. For each, we derive upper bounds on inference accuracy as a function of the number of snapshots/views and their average bit error rate (BER). These bounds guide the formulation of optimization problems that minimize total system latency while satisfying accuracy constraints. The joint optimization is decomposed into subproblems involving snapshot/view selection, transmit power control, and allocation of computation and communication time. To solve this efficiently, we develop a low-complexity iterative algorithm. Experimental results on synthetic and real-world datasets validate our approach, demonstrating significant latency reductions while maintaining inference performance. Our findings provide a foundation for robust and efficient edge AI design in future 6G networks.
Zhi-Feng Wang, Qunsong Zeng, Ren-Zhi Yuan et al.· IEEE Transactions on Wireles...· 0 citations
The integration of sensing capabilities into 6G wireless networks, known as integrated sensing and communication (ISAC), introduces severe privacy risks by enabling unauthorized user localization through the eavesdropping reference signals. Existing physical-layer defense strategies, such as artificial noise or conventional artificial multipath, suffer from critical limitations including significant communication performance degradation, reliance on unrealistic assumptions about adversarial channel state information (CSI), and increased system overhead. To address these challenges, this paper proposes a novel transparent artificial multipath (TAM) framework. It employs constant-envelope precoding and combining matrices to transparently encrypt reference signals in the frequency domain. This design ensures that legitimate base stations can perform channel estimation and localization without any performance loss, while adversaries receive a deliberately distorted channel impulse response, preventing accurate user localization and tracking. The proposed method is power-efficient, requires no prior knowledge of the adversarial CSI, and incurs negligible additional dynamic signaling overhead. Furthermore, we present a scalable multi-user extension that allows sharing the time-frequency resources without interference. Both theoretical analysis and extensive simulations demonstrate that TAM successfully misleads adversaries, inducing distance estimation errors of hundreds of meters, while fully maintaining the communication and localization performance for legitimate users. While physical-layer security solutions had a limited practical impact in various standards, the use of physical-layer techniques for privacy protection due to sensing is a necessity. We show that the proposed solution can be seamlessly integrated. into current 5G and future 6G standards.
Chen-Hu Kun, P. Popovski, Ana García Armada· IEEE Transactions on Wireles...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
