An Adaptive UAV Network for Increased User Coverage and Spectral Efficiency

Abstract

Unmanned Aerial Vehicles (UAVs) are fast becoming a popular choice in a variety of applications in wireless communication systems. UAV-mounted base stations (UAV-BSs) are an effective and cost-efficient solution for providing wireless connectivity where fixed infrastructure is not available or destroyed. We present a method of using UAV-BSs to provide coverage to mobile users in a fixed area. We propose an algorithm for predicting the user locations based on their mobility data and clustering the predicted locations, so that one UAV-BS would provide coverage to one user cluster. The proposed method, hence is similar to the UAV-BSs following the users to keep them under the coverage region. Simulation results show that the proposed method increases the user coverage by 47%-72% and increases the spectral efficiency by 43%-55% depending on the scenario and in addition, reduces the number of UAV-BSs required to provide coverage.

A Reinforcement Learning Approach for Fair User Coverage Using UAV Mounted Base Stations Under Energy Constraints

“…The authors of [19] proposed a new network architecture called 'Coordinated multipoint (CoMP) in the sky' that mitigated the inter-user interference by the corporation gain of CoMP and used the mobility of UAVs to provide strong channel gains to mobile ground users. A method of dynamically clustering roaming users and energy efficient trajectory planning for a set of UAV-BSs to serve the user clusters was presented in [20]. A UAV-BSs deployment and mobility control algorithm was proposed in [21].…”

“…It is understood that satellite-UAV links are not commercialised yet. However, UAV-BSs being backhaul-connected through satellite links, is a common assumption in literature [1], [20], and [28]. Later, we extend the scenario to use a fleet of UAV-BSs to provide downlink communication to ground users in the said area.…”

Improved genetic algorithm based 3-D deployment of UAVs

“…In this case, the deployed locations of UAVs need to be adjusted to provide effective coverage. To fully reap the benefits of the mobility of UAVs, the authors in [15] predicted the mobile users' locations in a disaster area and deployed an adaptive UAV network to provide coverage for the users with minimum energy. To improve the accuracy of prediction, a deep Q-network based method was proposed in [16] to adjust the locations of UAVs in response to the mobility of ground users.…”

“…To improve the accuracy of prediction, a deep Q-network based method was proposed in [16] to adjust the locations of UAVs in response to the mobility of ground users. Both [15] and [16] assumed that the power of UAVs was fixed, to further reduce power consumption, the authors in [17] investigated an adaptive UAV deployment scheme. In this scheme, the UAVs can adaptively change transmit power or relocate to new positions to guarantee the QoS of ground users as the users move.…”