Learning-Based Trajectory Optimization for 5G mmWave Uplink UAVs

Abstract

A Connectivity-constrained based path planning for unmanned aerial vehicles (UAVs) is proposed within the coverage area of a 5G NR Base Station (BS) that uses mmWave technology. We consider an uplink communication between UAV and BS under multipath channel conditions for this problem. The objective is to guide a UAV, starting from a random location and reaching its destination within the BS coverage area, by learning a trajectory alongside achieving better connectivity. We propose simultaneous learning-based path planning of UAV and beam tracking at the BS side under urban macro-cellular (UMa) pathloss conditions, to reduce its sweeping time with apriori computational overhead using the deep reinforcement learning method such as Deep Q-Network (DQN). Our results show that our proposed learning-based joint path planning and beam tracking method is on par with the learning-based shortest path planning, besides beam tracking comparable to heuristic exhaustive beam searching method.

Performance Optimization in UAV‐Assisted Wireless Powered mmWave Networks for Emergency Communications

“…However, UAVenabled cellular architecture over mmWave frequency band has been recognized as one of the best solution for ondemand high data rate service. Therefore, the many existing works focused on UAV-assisted mmWave communications [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38].…”

“…The authors of [26] focused on BS-to-UAV backhaul communications and proposed a beam tracking method, which reduces training overhead by adopting wider beam width at the cost of lower beamforming gain. The authors of [27] proposed a connectivity constraint-based path planning and beam tracking method, by which the UAV can start from a random location and reach its destination within a BS coverage by learning a trajectory while keeping better connectivity.…”

A Survey of Wireless Networks for Future Aerial COMmunications (FACOM)

“…Moreover, a number of works propose machine learning-based algorithms to capture the dynamic 3D environment for UAV Trajectory Planning. For instance, the authors propose deep Reinforcement Learning (RL) methods to plan the path of UAVs based on its connectivity constraints [206], [262], to minimize UAV mission completion time [207] and to minimize interference to ground UEs [263].…”

Deep Reinforcement Learning for Dynamic Band Switch in Cellular-Connected UAV

“…Besides, exploiting their mobility, UAVs can establish short Line of Sight (LoS) communication links, that represents an ideal situation to transmit at millimeter Wave (mmWave) band. A mmWave link offers a wide spectrum and enables the use of directional beamforming, providing high data rate [8]. The work in [8] addresses the cellular connected UAV communication-aware trajectory problem learning simultaneously the mmWave beam and trajectory via Deep Q-Network (DQN) to improve the Uplink (UL) performance.…”

“…A mmWave link offers a wide spectrum and enables the use of directional beamforming, providing high data rate [8]. The work in [8] addresses the cellular connected UAV communication-aware trajectory problem learning simultaneously the mmWave beam and trajectory via Deep Q-Network (DQN) to improve the Uplink (UL) performance. However, due to the severe attenuation and sensitivity to blockages, mmWave links are highly intermittent, leading to frequent radio failures at low UAV altitudes [9].…”