Abstract
Unmanned aerial vehicles (UAVs) have been envisioned as a promising technology to support wireless virtual reality (VR) applications. However, the huge energy consumption during the rendering process is challenging for UAV-enabled VR systems. The existing energy management strategies neglecting the dynamic and tiling features of VR tasks may not achieve optimal performance. To cope with these issues, we present a dynamic energy optimization framework for UAV-enabled VR systems. In the proposed framework, a tile-based collaborative rendering scheme for multi-UAVs is developed by exploiting the correlation among different tiles of VR users. Further, we formulate an optimization problem by jointly considering the collaboration decision, resource allocation, and UAV trajectory, aiming to reduce the long-term energy consumption under the motion-to-photon (MTP) latency requirement. Due to the tight coupling variables and the dynamic factors, the formulated problem is non-convex and difficult to tackle by adopting the traditional method. To this end, we first decompose the original problem into two sub-problems, i.e., the UAV trajectory and collaboration decision (UTCD) problem, and the resource allocation (RA) problem. Then, we design a dynamic UTCD algorithm based on deep reinforcement learning (DRL) and present an optimal RA algorithm based on the Lagrangian multiplier iterative (LMI) method. To enhance the solving efficiency, a block coordinate descent (BCD)-based solving method is introduced by embedding the RA algorithm in the learning environment of DRL. Finally, simulation results prove that the proposed scheme can obtain good performance for UAV-enabled VR systems.
Funding Information
  • National Key Research and Development Program of China (2022YFB3104900)
  • National Natural Science Foundation of China (62341129)
  • Postdoctoral Fellowship Program of CPSF (GZB20240061)
  • China Postdoctoral Science Foundation (2024M750181)
  • Natural Science Foundation of Hebei Province (F2022402001)
  • Central Fund Project for Guiding Local Science and Technology Development (236Z0401G)
  • NSF (CNS-2107216, CNS-2128368, CMMI-2222810, ECCS-2302469)
  • U.S. Department of Transportation
  • Toyota Motor Corporation
  • Amazon
  • Japan Science and Technology Agency (JPMJAP2326)

This publication has 51 references indexed in Scilit: