Abstract
This paper presents a framework for optimizing the 3D placement and mobility of multiple Unmanned Aerial Vehicles (UAVs) serving as aerial base stations for energy-constrained Internet of Things (IoT) networks. By jointly optimizing device association, power control, and UAV trajectory, the proposed approach achieves a 45% reduction in IoT device transmit power and a 28% improvement in system reliability compared to stationary deployments.
Introduction
The proliferation of the Internet of Things (IoT) involves numerous battery-limited sensors that often struggle to maintain long-range communication with terrestrial base stations. Unmanned Aerial Vehicles (UAVs) offer a flexible solution as "flying base stations." However, their deployment involves complex trade-offs between sensor power consumption, UAV flight energy, and network interference. This research addresses the joint optimization of UAV positioning and mobility to maximize network efficiency.
System Model and Methodology
The system considers a set of IoT devices distributed across a geographical area, served by multiple UAVs. The optimization problem is tackled through a two-step iterative framework.
1. Static Snapshot Optimization
At any discrete time interval, the system optimizes UAV placement and device association.
- Association & Power Control: For fixed UAV locations, an iterative algorithm determines the optimal device-to-UAV mapping and minimizes transmit power to satisfy Signal-to-Interference-plus-Noise Ratio (SINR) requirements.
- 3D Placement: UAV coordinates are optimized using Sequential Quadratic Programming (SQP) for interference-heavy scenarios or convex optimization for interference-free environments.
2. Dynamic Mobility & Timing
To account for varying device activation patterns, the framework determines optimal "update times" for UAV repositioning.
- Activation Models: The system uses Periodic and Probabilistic (Beta distribution) models to predict when sensors will be active.
- Trajectory Planning: The Hungarian Method is employed to solve the 3D trajectory mapping problem, minimizing the energy expended by UAVs during movement between optimal locations.
Results and Discussion
Simulations in a 1 km x 1 km urban environment with 500 IoT devices demonstrate significant performance gains. The mobile UAV approach reduced total IoT transmit power by 45% and improved reliability by 28% over stationary alternatives. In interference-free scenarios, power savings reached up to 80%. Furthermore, the iterative algorithm demonstrated high computational efficiency, converging in fewer than five iterations and operating 500 times faster than brute-force methods with only an 11% optimality gap.
Conclusion
While the framework significantly enhances IoT energy efficiency, future work is needed to address dynamic obstacles, mobile IoT devices, and backhaul link constraints. Additionally, integrating battery replenishment logistics for the UAV fleet remains a critical area for practical deployment.
Summary updated