A Pilot Contamination-Aware Lightweight HQCNN Algorithm for mURLLC in AI-Agent Communication Networks

Abstract

Artificial intelligence-agent communication networks (ACNs) have emerged as a key scenario in the future sixth-generation communication systems, where autonomous agents require massive ultra-reliable and low-latency communications (mURLLC) for real-time collaboration under cell-free architecture. To support mURLLC requirements, this paper investigates pilot contamination in cell-free massive multiple-input multipleoutput systems. We propose an ACN architecture, derive a closed-form expression for pilot contamination, and propose a light-weight hybrid quantum-classic convolutional neural network (HQCNN) algorithm that leverages quantum parallelism and nonlinear learning to mitigate the pilot contamination effectively. Simulation results show a 99.71% reduction in error probability and a 31.8% improvement in connection density, showing great potential in industrial metaverse and swarm robotics in the next-generation ACNs.