AI-Powered Drones Boost Wireless Speeds in Crowded Networks

In an era where fast, reliable wireless connections are essential for everything from streaming video to smart city applications, maintaining strong signals in crowded or obstructed areas remains a challenge. This research introduces a smart solution that uses drones equipped with intelligent reflectors to significantly improve wireless performance, making it a critical advancement for next-generation networks like 5G and beyond.

The key discovery is that an unmanned aerial vehicle (UAV) carrying an intelligent reflector (IR) can enhance multi-user downlink transmissions over millimeter-wave (mmWave) frequencies. By dynamically adjusting the reflector's position and settings, the system maximizes the sum data rate for multiple users, even when obstacles block direct lines of sight. This approach ensures that signals are reflected to maintain clear connections, addressing common issues like signal attenuation from buildings or trees.

To achieve this, the researchers employed a distributional reinforcement learning (DRL) method. This AI technique enables the UAV-IR to learn from the environment and make real-time decisions on its location and reflection coefficients, optimizing for long-term communication performance. The base station's precoding matrix is also jointly optimized with the reflector's settings, focusing on maximizing data rates without requiring extensive manual configuration.

Simulation results demonstrate substantial improvements. The proposed method achieves a line-of-sight probability of over 90%, compared to lower rates for non-learning approaches. In terms of data rates, it delivers more than double the performance of direct transmission schemes and outperforms non-learning UAV-IR and static IR methods by 25% to 50%. For instance, as transmit power increases from 20 to 40 dBm, the downlink data rate rises significantly, with the AI-enhanced system maintaining higher averages. In worst-case scenarios, where received power is low, the DRL model selects optimal actions, such as increasing altitude, to boost signals effectively.

This innovation matters because it enhances the reliability and speed of wireless networks in urban or dense environments, where mmWave signals are easily disrupted. For everyday users, this could mean fewer dropped calls, faster internet in crowded areas, and better support for emerging technologies like autonomous vehicles and IoT devices. It also offers energy efficiency, as the reflector consumes minimal power compared to traditional relay stations.

However, the study has limitations. It focuses on outdoor scenarios with static users and does not address mobile users or multiple UAV-IRs in complex environments. Future work will explore these areas to further improve adaptability and scalability.