2603.22127

Total: 1

#1 DQN Based Joint UAV Trajectory and Association Planning in NTN Assisted Networks [PDF] [Copy] [Kimi] [REL]

Authors: Afsoon Alidadi Shamsabadi, Cosmas Mwaba, Thomas Nugent, Jie Gao, Pablo Madoery, Halim Yanikomeroglu, Subhadeep Pal

Advanced Air Mobility (AAM) has emerged as a key pillar of next-generation transportation systems, encompassing a wide range of uncrewed aerial vehicle (UAV) applications. To enable AAM, maintaining reliable and efficient communication links between UAVs and control centers is essential. At the same time, the highly dynamic nature of wireless networks, combined with the limited onboard energy of UAVs, makes efficient trajectory planning and network association crucial. Existing terrestrial networks often fail to provide ubiquitous coverage due to frequent handovers and coverage gaps. To address these challenges, geostationary Earth orbit (GEO) satellites offer a promising complementary solution for extending UAV connectivity beyond terrestrial boundaries. This work proposes an integrated GEO terrestrial network architecture to ensure seamless UAV connectivity. Leveraging artificial intelligence (AI), a deep Q network (DQN) based algorithm is developed for joint UAV trajectory and association planning (JUTAP), aiming to minimize energy consumption, handover frequency, and disconnectivity. Simulation results validate the effectiveness of the proposed algorithm within the integrated GEO terrestrial framework.

Subject: Systems and Control

Publish: 2026-03-23 15:54:25 UTC