近年來無人機由於其成本低廉和功能豐富而在無線通訊和運輸系統中得到了大量的應用，藉由可控的移動性以及部屬的靈活性，讓無人機可以為佈署在複雜、危險區域的通訊節點提供服務，由於這些節點所處的位置不易抵達，以有線方式輸送能量易使線路的架設及維護困難，所以這些節點的運作勢必得依賴本身的電池作為電力來源，長期運作之下可能會導致節點出現電力不足的問題，而利用無人機的無線充電能力則可有效的解決，但無人機的飛行與充電能力受限於本身搭載電池的電力限制，因此有效地規劃無人機通訊的資源分配是一大設計挑戰。 本研究考慮多個獵能節點利用無人機下鏈無線充電來進行上鏈通訊傳輸資料至多台無人機，探討無人機功率控制策略、飛行軌跡以及無人機與節點通訊關聯，以有效管理多台無人機在通訊環境下的同頻干擾，為確保公平性，採用最大化所有節點的最小資料傳輸率作為設計目標。此聯合設計是一個高度非凸問題，並且需要完美知道未來時間的通道狀態資訊，然而這在現實環境中很難預測得知。為克服這些設計難題，本研究提出一種基於凸優化的離線方法，該方法僅利用統計平均的通道狀態資訊，通過應用交替優化和連續凸逼近將問題轉化成三個凸子問題，進而求得無人機功率控制、飛行軌跡以及無人機與節點通訊關聯的離線策略。;In recent years, unmanned aerial vehicles (UAVs) have been used in a large number of wireless communication and transportation systems due to their low cost and rich functionality. The controlled mobility and flexibility of their components allow UAVs to serve communication nodes deployed in complex and hazardous areas. Since these nodes are located in inaccessible locations, the wired transmission of energy makes it difficult to set up and maintain the lines, so the operation of these nodes must rely on their own batteries as a source of power, which may lead to power shortage problems at the nodes over a long period of time, and the wireless charging capability of UAVs can effectively solve the problem. However, the flight and charging capability of UAVs are limited by the power capacity of their own batteries, so the effective planning of resource allocation for UAV communication is a major design challenge. In this thesis, we consider multiple energy-harvested nodes to transmit data to multiple UAVs using UAV downlink wireless power transfer for uplink communication, and investigates UAV power control strategies, flight trajectories, and UAV-node communication links to effectively manage co-channel interference of multiple UAVs in the communication environment. To ensure fairness, the design objective is to maximize the worst-case node total data transfer rate. The joint design problem is highly non-convex and requires the causal (future) knowledge of the channel state information (CSI), which is difficult to predict in reality. To overcome these design challenges, this paper proposes an offline method based on convex optimization that only utilizes the average CSI and solve the problem via three convex sub-problems by applying alternating optimization and successive convex approximation (SCA) to find the offline strategy for UAV power control, flight trajectory, and UAV-node communication association.
