Distributed Safe Multi-Agent Reinforcement Learning: Joint Design of THz-Enabled UAV Trajectory and Channel Allocation
Bibliographic record
Abstract
6G is anticipated to play a foundational role in realizing various emerging entertainment applications and critical societal services, such as smart agriculture, public safety, and so on. Providing the underlying communications for these applications will substantially increase demand for data rate, reliability, and resiliency. Given these requirements, Terahertz (THz)-enabled unmanned aerial vehicles (UAVs) are expected to support the essential functions within the future 6G networks. Meanwhile, due to the dynamic environment in a THz-enabled UAV-assisted network with multiple UAVs, using model-free multi-agent deep reinforcement learning (MADRL) becomes a promising approach to optimize scarce resources. However, satisfying cooperative safety constraints is not guaranteed in most previous MADRL techniques. Consequently, this paper aims to address the cooperative safety guarantee challenge in the joint design problem of UAV trajectory and channel allocation in a THz-enabled UAV-assisted network. Specifically, the goal of the proposed scheme is to maximize energy efficiency, considering the quality of service of each IoT device, the speed limitation of UAVs, and the collision avoidance constraint between UAVs. The problem is mixed integer nonlinear programming (MINP), known as NP-hard to solve. Thus, we propose a novel distributed safe MADRL (DSMADRL) approach for the safe trajectory design of UAVs using a data-driven uniformly ultimate boundedness stability method. Moreover, we theoretically show that the DSMADRL approach guarantees satisfying the cooperative safety constraint of collision avoidance between UAVs as well as maximizing energy efficiency. Furthermore, the matching method is used to allocate THz sub-bands distributively. Finally, the effectiveness of the algorithms is evaluated by comparing them with the existing multi-agent RL algorithms.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".