Large Language Models (LLMs) demonstrate impressive capabilities in natural language understanding and generation, but incur high communication overhead and privacy risks in cloud deployments, while facing compute and memory constraints when confined to edge devices.Cloud–edge inference has emerged as a promising paradigm for improving privacy in LLM services by retaining sensitive computations on local devices.However, existing cloud–edge inference approaches apply uniform privacy protection without considering input sensitivity, resulting in unnecessary perturbation and degraded utility even for non-sensitive tokens. To address this limitation, we propose Privacy-aware Routing for Inference with Semantic Modulation (PRISM), a context-aware framework that dynamically balances privacy and inference quality. PRISM executes in four stages: (1) the edge device profiles entity-level sensitivity; (2) a soft gating module, also on the edge, selects an execution mode -cloud, edge, or collaboration; (3) for collaborative paths, the edge applies adaptive two-layer local differential privacy based on entity risks; and (4) the cloud LLM generates a semantic sketch from the perturbed prompt, which is then refined by the edge-side small language model (SLM) using local context.Our results show that PRISM consistently achieves superior privacy-utility trade-offs in various scenarios, reducing energy consumption and latency to 40–50% of baseline methods such as Uniform and Selective LDP, while maintaining high output quality under strong privacy constraints. These findings are validated through comprehensive evaluations involving realistic prompts, actual energy measurements, and heterogeneous cloud–edge model deployments.

This paper presents a robust Model Predictive Control (MPC) framework for trajectory tracking of unmanned aerial vehicles (UAVs), enhanced by a reinforcement learning (RL) policy for warm-start initialization and a sliding mode observer (SMO) for disturbance estimation. The proposed architecture addresses multifaceted performance degradation, including residual trajectory tracking errors, reduced control responsiveness, and slow early-stage convergence, primarily caused by external disturbances, model uncertainties, and time-varying environmental conditions. An enhanced adaptive super-twisting SMO is developed to estimate lumped disturbances in real time. These estimates are integrated into the MPC prediction model to compensate for mismatches between nominal dynamics and true system behavior. To accelerate control convergence, particularly in improving early stage performance, an offline RL warm-start policy is used to generate an initial control sequence for MPC. The overall framework preserves the constraint handling and predictive capabilities of conventional MPC, while significantly improving robustness, stability, and tracking accuracy. The simulation results validate the effectiveness of the proposed method in achieving reliable and precise trajectory tracking under challenging and uncertain operating scenarios.