Energy-Efficient Aerial STAR-RIS-Aided Computing Offloading and Content Caching for Wireless Sensor Networks

Area of Science:

• Wireless Communication
• Network Engineering
• Artificial Intelligence

Background:

• Unmanned Aerial Vehicle (UAV)-based wireless sensor networks (WSNs) offer mobility and line-of-sight links for ground sensors.
• Mobile Edge Computing (MEC) in UAV-WSNs addresses transmission and coverage limitations of traditional base stations.
• Simultaneously Transmitting and Reflecting Reconfigurable Intelligent Surfaces (STAR-RISs) enhance propagation but are often fixed or underutilized.

Purpose of the Study:

• To propose an energy-efficient framework for UAV-based WSNs integrating STAR-RIS for computing offloading and content caching.
• To address limitations of fixed STAR-RIS positions and underutilized UAV computing/caching capabilities.
• To minimize overall system energy consumption through joint optimization.

Main Methods:

• Formulation of an energy consumption minimization problem.
• Decomposition into subproblems: content caching, computing offloading, UAV hovering, and STAR-RIS resource allocation.
• Application of a Deep Reinforcement Learning (DRL)-Successive Convex Approximation (SCA) combined algorithm for near-optimal solutions.

Main Results:

• The proposed framework effectively optimizes resource utilization in UAV-based WSNs.
• Significant reduction in overall system energy consumption demonstrated.
• Numerical results validate the efficiency of the DRL-SCA algorithm.

Conclusions:

• The aerial STAR-RIS-aided framework offers a flexible and efficient solution for UAV-WSNs.
• Joint optimization of caching, offloading, positioning, and beamforming is crucial for energy efficiency.
• The DRL-SCA approach provides a low-complexity method for solving complex optimization problems in this domain.