Distributed Drone Base Station Positioning for Emergency Cellular Networks Using Reinforcement Learning.

Area of Science:

• Telecommunications Engineering
• Artificial Intelligence
• Disaster Management

Background:

• Natural disasters necessitate rapid deployment of functional communication networks for emergency response.
• Traditional infrastructure is vulnerable, highlighting the need for mobile and adaptable solutions.
• Unmanned aerial vehicles (drones) offer a promising platform for emergency communication infrastructure.

Purpose of the Study:

• To propose an intelligent solution for optimal drone small cell (DSC) placement in emergency scenarios.
• To maximize user coverage within network constraints using reinforcement learning.
• To evaluate the performance of the proposed Q-learning approach against other methods.

Main Methods:

• Developed a reinforcement learning algorithm, specifically Q-learning, for dynamic DSC positioning.
• Modeled the emergency scenario with constraints on backhaul and radio access networks.
• Compared the Q-learning solution against baseline approaches for performance evaluation.

Main Results:

• The proposed Q-learning solution significantly outperformed all other evaluated approaches.
• Maximized user coverage was achieved under realistic network limitations.
• Demonstrated superior performance across all considered metrics.

Conclusions:

• Intelligent drone small cells (DSCs) provide an effective solution for rapid emergency communication deployment.
• Reinforcement learning, particularly Q-learning, is a viable method for optimizing DSC placement.
• This approach enhances network resilience and reduces potential loss of life during disasters.