Abstract
Unmanned aerial vehicles (UAVs) are emerging as key enablers for low-altitude wireless networks (LAWNs), especially when terrestrial infrastructure is unavailable. In such scenarios, the environmental topology is typically unknown, making efficient and safe UAV trajectory design both essential and challenging. To address this problem, we propose a digital twin (DT)-assisted training and deployment framework (DTTDF), in which virtual environments (VEs) are constructed from scratch and progressively refined using real-time sensing data. During training, diverse VEs accelerate and stabilize policy learning, while during deployment a synchronized VE is continuously updated to support safety-aware control and reliable navigation. Built upon this framework, we further develop a simulated annealing and twin-delayed deep deterministic policy gradient-based trajectory design (SATD3TD) scheme, which jointly optimizes discrete user scheduling and continuous UAV trajectory design. By coordinating user scheduling with adaptive trajectory control, the UAV can dynamically respond to newly observed obstacles while maintaining mission efficiency, enabling effective coupled decision-making in dynamic environments. Simulation results demonstrate that the proposed approach achieves faster policy convergence, improved flight safety, and shorter mission completion time compared with baseline methods, validating the effectiveness of SATD3TD within DTTDF for reliable UAV deployment in complex unknown environments.
| Original language | English |
|---|---|
| Journal | IEEE Internet of Things Journal |
| DOIs | |
| Publication status | Accepted/In press - 2026 |
| Externally published | Yes |
Keywords
- UAV trajectory design
- deep reinforcement learning
- digital twin
- low-altitude wireless network
- unknown environments
Fingerprint
Dive into the research topics of 'Trajectory Design for UAV-Based Low-Altitude Wireless Networks in Unknown Environments: A Digital Twin-Assisted TD3 Approach'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver