Abstract
This paper studies the improvement of the achievable end-to-end data rate of ground users assisted with unmanned aerial vehicles (UAVs) and tethered balloons (TBs). The goal is to maximize the end-to-end throughput of a network suffering from the absence of terrestrial infrastructure. First, we solve an integer linear programming problem to optimize the associations. Then, we solve the UAVs transmit powers optimally by converting the problem into a convex one. Subsequently, an efficient algorithm is proposed to optimize the UAVs placement. Finally, our mathematical formalism is illustrated with some selected numerical results that show the advantages provided by our proposed scheme.
Introduction
In scenarios where terrestrial infrastructure is destroyed or non-existent, such as natural disaster zones or remote regions, Unmanned Aerial Vehicles (UAVs) have emerged as an effective solution for rapid-deployment "flying base stations." However, UAV-based networks face two critical challenges: establishing reliable backhaul connectivity to the core internet and managing limited resources efficiently. While multi-hop UAV links are often proposed, they suffer from instability and management complexity.
This research proposes a two-tier aerial network architecture that introduces high-altitude Tethered Balloons (TBs) as stable backhaul anchors. By leveraging the high altitude of TBs, the system ensures a high probability of Line-of-Sight (LoS) connectivity. To maximize the network's end-to-end throughput, we develop a robust framework to jointly optimize 3D UAV placement, user associations, and transmit power allocation.
System Model and Methodology
The system considers a downlink network using Orthogonal Frequency Division Multiple Access (OFDMA) to eliminate intra-cell interference. The architecture consists of TBs at fixed high altitudes providing backhaul to multiple UAVs, which in turn serve a set of ground users.
The optimization goal is to maximize the minimum end-to-end throughput across all users, ensuring fairness and network reliability. Given that this results in a complex Mixed Integer Non-Linear Programming (MINLP) problem, we decompose it into three iterative steps:
- Association Optimization: We utilize Integer Linear Programming (ILP) to determine the optimal connections between users and UAVs, as well as between UAVs and TBs.
- Power Allocation: With fixed associations, the problem is transformed into a convex form. We apply Lagrangian Duality and the subgradient method to derive an optimal power distribution strategy, akin to a "water-filling" approach constrained by backhaul capacity.
- Placement Optimization: To determine the optimal 3D coordinates for the UAVs, we propose the Shrink-and-Realign (S&R) algorithm. This heuristic method iteratively evaluates candidate positions and refines the search radius to converge on the most effective deployment locations.
Numerical Results and Analysis
The proposed framework was evaluated through numerical simulations in a $1000\text{m} \times 1000\text{m}$ area with 20 to 30 randomly distributed users. The simulation parameters included two TBs at 200m altitude and four UAVs at 100m altitude, utilizing standard urban path loss models.
Key findings from the simulation include:
- Throughput Improvement: The joint optimization scheme achieved a 58% improvement in throughput compared to random association benchmarks and a 21% gain over optimized association with uniform power allocation.
- Backhaul Bottleneck: The results confirm that the end-to-end rate is limited by the weakest link. Increasing UAV transmit power only improves performance until the access link speed matches the backhaul capacity provided by the TBs.
- Algorithm Efficiency: The S&R algorithm demonstrated rapid convergence, typically within 6 to 9 iterations, proving its suitability for dynamic, real-time deployments.
Conclusion and Future Work
This paper demonstrates that integrating Tethered Balloons into UAV-assisted networks significantly enhances throughput and reliability in infrastructure-less environments. The proposed three-step optimization effectively manages the complexities of resource allocation and 3D placement. Future research will explore managing inter-cell interference in high-density deployments and the potential of Free-Space Optical (FSO) links to further alleviate backhaul bottlenecks and provide energy harvesting opportunities for the UAVs.
Summary updated