<p>Emergency response operations require autonomous UAV swarms to maintain reliable performance under highly dynamic and unpredictable conditions. The operational environment for these systems tests their performance because it creates situations where sensors fail, communication systems break down and individual drones go missing. Current swarm designs treat perception, coordination and fault tolerance as separate components, which limits their ability to operate effectively in dynamic and uncertain environments. This research presents a digital twin model that functions as the primary operational system. The Quantu-Twin offers a digital twin framework that combines amplitude-based probabilistic representations with classical implementation to enable near-real-time updates between physical UAVs and their digital counterparts. The proposed framework supports reliable perception under uncertainty, predictive planning and automatic system recovery. The framework also incorporates a lightweight probabilistic decision mechanism to support reliable operation under uncertain conditions. The combination of these two frameworks enables resource-constrained UAV swarms to carry out effective uncertainty management, fault recovery, and short-term hazard-aware planning operations. The framework is validated through high-fidelity simulations and real-world field experiments using UAV swarms of 10 to 15 agents operating in dynamic disaster environments. Experimental evaluation demonstrated notable improvements in overall swarm performance, including approximately 20% lower mission completion time and about 42% reduction in communication overhead. The framework also achieved better perception accuracy and maintained mission success rates above 92% even under GPS degradation and packet-loss conditions exceeding 45%. The results from ablation studies show that predictive digital twins function as shared control systems, which enables multiple agents to coordinate effectively under uncertain conditions.</p>