Sep 2026· International Conference on Internet of Things, Communication Engineering, and Artificial Intelligence· Vol 14373, pp. 1437316 - 1437316-6· 0 citations· 11 references
Engineering
TL;DR
Simulations show the SAC-based approach outperforms benchmarks in task completion rate, energy consumption, and convergence, and validate the maximum entropy mechanism's effectiveness in enhancing exploration and generalization in UAV-assisted ISAC environments.
Abstract
UAVs are key for ISAC due to their mobility and coverage, but existing resource allocation methods lack adaptability in dynamic environments. This paper proposes a dynamic resource allocation framework for UAV-assisted ISAC using the SAC deep reinforcement learning algorithm. The framework constructs a state space with UAV parameters and jointly optimizes power control, channel selection, and task scheduling in a continuous action space. A maximum entropy mechanism balances exploration and energy-efficient decision-making. Simulations show the SAC-based approach outperforms benchmarks in task completion rate, energy consumption, and convergence. Specifically, it achieves 93.7% task completion, 36.9 J energy consumption, and 3.96 bit/J energy efficiency, with a 24.5% gain over DDPG. It converges within 310 seconds, faster than TD3 and PPO. These results validate the maximum entropy mechanism's effectiveness in enhancing exploration and generalization in UAV-assisted ISAC environments.
An adaptive resource allocation method to optimize the delay of UAVs under highly dynamic environments and Quantum Deep Reinforcement Learning (QDRL) is formulated since it can learn an optimal policy that maximizes the earned reward quickly even in dynamic environments conditions.
F. A. Asuhaimi, Hafizal Mohamad, N. S. Suhaimi· 2026 IEEE International Conf...· 0 citations
This paper investigates an energy efficiency maximization problem in a UAV-assisted WPCN by jointly optimizing the UAV trajectory, hovering altitude, and hybrid TDMA/NOMA resource allocation and proposes a deep reinforcement learning-driven Hierarchical Energy-Efficient Data Collection scheme, which is named DRL-HEEC.
Si-Liang Gong, Kai-Yang Qu, Qi-Sen Wang et al.· Electronics· 0 citations
A Deep Reinforcement Learning (DRL) framework that jointly optimizes user scheduling, IoT device transmit power, bandwidth, and UAV movement in a 6G-enabled UAV-relay uplink network, using a deterministic large-scale air-to-ground path-loss channel model.
Alissa Nauman, Sung Won Kim· Italian National Conference...· 0 citations
The growing demand for multimedia services in Internet of Things (IoT) networks has significantly increased the traffic load on backhaul links, making Mobile Edge Caching (MEC) a key technology for reducing content delivery latency. Unmanned Aerial Vehicles (UAVs) can serve as mobile aerial caching nodes that complemen...
Tao Zhang, Tao Xu, Ze-Kai Liu et al.· Journal of Circuits, Systems...· 0 citations
Autonomous vehicles (AVs) and 5G wireless networks need low latency, reliability, and energy efficiency. AVs generate massive amounts of heterogeneous, real-time data, requiring efficient energy and resource allocation for large-scale vehicular communications. Vehicles with fast mobility patterns, channel conditions, a...
M. Thenmozhi, B. Sridevi· Discover Computing· 0 citations
An efficient framework to maximize global energy efficiency (EE) while promoting safe operation through threat-aware clustering and reward-based safety enforcement and demonstrates effective generalization to unseen user distributions, large UAV fleets, and different threat geometries, while maintaining zero safety vio...
F. Al-kamali, Hussein A. Ammar, François Chan et al.· IEEE Internet of Things Jour...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 9, 2026