Scholarlink

<p>The increasing complexity of multi-unmanned aerial vehicle (UAV) reconfigurable intelligent surface (RIS) networks in heterogeneous sixth-generation (6G) settings reveals severe constraints in centralized meta-learning paradigms such as single-point-of-failure problems, bottleneck problems with communication overheads during reconfiguration of models in non-stationary distribution shifts. In this paper, I suggest QuantumFed-ARIS as a radically new architecture, a hybrid quantum–classical architecture to integrate quantum-computing with federated continual learning to a full autonomous RIS-UAV constellation management. The framework introduces four key innovations: (i) a Quantum Variational Policy Network (QVPN) that exploits quantum superposition and entanglement to explore exponentially larger joint RIS phase-shift and UAV trajectory solution spaces with provable speedup over classical counterparts; (ii) a Federated Elastic Weight Consolidation (FedEWC) protocol that enables privacy-preserving collaborative learning across UAV constellations while selectively protecting critical knowledge from catastrophic forgetting through Fisher information-weighted parameter regularization; (iii) a Real-Time Digital Twin Engine (RT-DTE) that maintains a physics-informed virtual replica of the entire RIS-UAV-channel ecosystem for predictive what-if analysis and pre-emptive resource reallocation; and (iv) a Hierarchical Intent-Aware Scheduler (HIAS) that translates high-level network operator intents into optimized low-level RIS-UAV control actions through semantic policy decomposition. Massive experiments on twelve different deployment scenarios have shown that QuantumFed-ARIS has achieved 52% spectral efficiency gain, 41% energy savings, 99.7% communication reliability and 78% faster convergence than state-of-the-art centralized and federated baselines, with a transformative foundation of self-governing aerial intelligent networks.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Quantum fed-ARIS: quantum-enhanced federated continual learning for autonomous RIS-UAV constellation networks with digital twin-guided predictive optimization

  • 期刊论文

  • 发表日期：

    2026年7月22日

查看全文
  • J. Girish,
  • Jitendra Kumar,
  • V. Premalatha

摘要

The increasing complexity of multi-unmanned aerial vehicle (UAV) reconfigurable intelligent surface (RIS) networks in heterogeneous sixth-generation (6G) settings reveals severe constraints in centralized meta-learning paradigms such as single-point-of-failure problems, bottleneck problems with communication overheads during reconfiguration of models in non-stationary distribution shifts. In this paper, I suggest QuantumFed-ARIS as a radically new architecture, a hybrid quantum–classical architecture to integrate quantum-computing with federated continual learning to a full autonomous RIS-UAV constellation management. The framework introduces four key innovations: (i) a Quantum Variational Policy Network (QVPN) that exploits quantum superposition and entanglement to explore exponentially larger joint RIS phase-shift and UAV trajectory solution spaces with provable speedup over classical counterparts; (ii) a Federated Elastic Weight Consolidation (FedEWC) protocol that enables privacy-preserving collaborative learning across UAV constellations while selectively protecting critical knowledge from catastrophic forgetting through Fisher information-weighted parameter regularization; (iii) a Real-Time Digital Twin Engine (RT-DTE) that maintains a physics-informed virtual replica of the entire RIS-UAV-channel ecosystem for predictive what-if analysis and pre-emptive resource reallocation; and (iv) a Hierarchical Intent-Aware Scheduler (HIAS) that translates high-level network operator intents into optimized low-level RIS-UAV control actions through semantic policy decomposition. Massive experiments on twelve different deployment scenarios have shown that QuantumFed-ARIS has achieved 52% spectral efficiency gain, 41% energy savings, 99.7% communication reliability and 78% faster convergence than state-of-the-art centralized and federated baselines, with a transformative foundation of self-governing aerial intelligent networks.