| 摘要: | 無人機（Unmanned aerial vehicles, UAVs）與行動通訊網路的整合對於網路覆蓋率、負載量和效率帶來許多好處，其中，飛行特設網路（Flying ad-hoc networks, FANETs）和空中基站（Air base stations, ABSs）是以任務導向的應用服務中使用無人機來促進資料傳輸的案例。其一，FANETs的無人機可以使用直接存儲和轉發或存儲-攜帶-轉發（Store, carry and forward）的資料傳輸技術將失去連線的節點和孤立的網路互連；另一方面，5G/6G網路的空中基站允許無人機在行進間向地面設備提供無線連線，然而，為了建立一個有效的無人機輔助系統，必須考慮一些限制因素，如有限的電池蓄電、資料緩衝區和通訊資源問題等。因此，我們以三個面向來探討此議題：無人機團隊工作、資源配置和路徑規劃最佳化。無人機團隊合作包括任務分配、角色定位、資訊共用以及時間同步；資源配置則是討論資源的公平和有效分配，如頻寬、電池電力和存儲空間；路徑規劃優化涉及最佳化電池電力的消耗和滿足地面設備對服務品質要求的資料收集路線。綜合上述，一個有效的無人機輔助系統必須克服這些問題，並探討如何實現最佳性能。為此，本論文的研究提出了一組在無線感測網路中無人機放置和任務卸載的機制，提供無人機間的電池電力使用平衡和服務時間之聯合優化，這套機制的組成包括兩部分：（1）在資料收集任務中無人機放置問題與電池電力平衡之優化；（2）無人機輔助資料收集和任務卸載的服務時間和電池電力使用成本之聯合優化。
在大規模的無線感測網路中，部署無人機可以避免資料收集和傳遞受到地面物理通信的限制，相較於傳統的多跳點無線網路上的路由方法，無人機輔助的資料傳遞是直接並且可管理的。為此，我們利用無人機群的概念，在這種環境中進行團隊合作的資料收集，同時，為了最大限度地延長無人機群的壽命，我們提出一個公平和電力使用平衡的無人機群放置策略，我們將其命名為Fair and Energy-Balanced Ferry Fleet Placement Scheme（FEB），以共同考慮電池電力消耗的效率、平衡和公平性。該方案有兩個相互作用的階段。第一階段使用Power-Voronoi演算法劃定網路中的服務區域；第二階段使用基於基因演算法的旅行推銷員問題（Travelling Salesman Problem, TSP)，決定每個服務區域內感測器之間的最短路徑行程。因此，該方案能夠確保公平的任務分配，平衡電力消耗，並延長團隊合作中多台無人機的使用壽命。實驗性能的量測結果顯示，所提出的方案在累計電力計算、剩餘電力分佈、公平指數關於電池電力的利用、存活的無人機數量以及無人機團隊合作期間的任務執行總時間等方面優於幾個典型方法，包括Native、K-Means和Spiral等方法。
在缺乏基礎設施的網絡環境中，無人機可被用來收集感測數據，並成為許多任務型導向應用的方法。然而，無人機配備有限的資料緩衝空間和電池容量，在此前提下，經由無人機協助的任務將不可避免地受到資源利用以及成本效益的影響。在本論文中，我們針對一個在無人機輔助的無線感測網路中，闡述從目標興趣點（Points of Interest, PoIs）蒐集所有感測數據的任務時間最小化之問題。我們首先定義了無人機在飛行、等待和充電卸載三種狀態下電池電力消耗的成本和剩餘資料緩衝空間大小，接著將其表述為一個混合整數線性規劃（Mixed-Integer Linear Programming, MILP）的最佳化問題，當這個問題出現非決定性的多項式時間計算成本時，即所謂的NP-hard，我們提出套名為 Mission Time Optimization with Energy and Buffer Constraints for Multi-UAV Deployment（MEBD）的最佳化策略。其中，Weighted TSP（W-TSP）演算法旨在發現加權最短路徑以滿足各種數據延遲預算的要求，而 Non-Linear Least Squares based Recharging Route（NLLSR） 演算法則引入非線性最小平方法來改進無人機的部署，提高電池電力和資料緩衝空間的利用效率。經由模擬仿真的實驗結果，我們的MEBP在總任務時間、電力耗損以及訊息遺失率優於Convex hull, RRT, RRT+, DTP, TSP, grid and AHTP-RL等其他方法。
本論文研究在多無人機輔助資料收集系統中顯示出電池電力使用的平衡及任務時間妥善分配的最佳化結果，後續的研究規劃將朝向整合本論文研究成果於新興的無線通訊系統，如Multi-Access Edge Computing（MEC）, Mobile Crowdsourcing and Space-Air-Ground Integrated Network（SAGIN）.
;The integration of unmanned aerial vehicles (UAVs) into communication networks offers numerous advantages in terms of network coverage, capacity, and efficiency. Flying ad-hoc networks (FANETs) and aerial base stations (ABSs) serve as network devices that leverage UAVs to enhance data delivery in mission-centric applications. Specifically, UAVs in FANETs interconnect disconnected nodes and isolated networks using techniques like direct store-and-forward or store-carry-and-forward. Meanwhile, ABSs in 5G/6G networks enable UAVs to deliver wireless connectivity to terrestrial devices on the move. Despite these benefits, challenges like energy, buffer, and communication resource limitations. Our study in this dissertation discusses above issues in three pivotal aspects: UAV teamwork, resource allocation, and path planning optimization. UAV teamwork encompasses task assignment, role determination, information exchange, and synchronized operations. Resource allocation aims for equitable and effective distribution of resources, such as bandwidth, energy, and storage. Path planning optimization focuses on energy efficiency and designing data gathering routes that satisfy the QoS demands of ground devices.
To address these challenges and attain optimal results, this dissertation introduces two key strategies for optimizing UAV-assisted systems in mobile networks: (1) Energy-Balanced Optimization on Flying Ferry Placement for Data Gathering, and (2) Joint Optimization of Service Time and Energy Cost on UAV-Assisted Data Gathering and Task Offloading.
Deploying flying UAVs in large-scaled wireless sensor networks can prevent data gathering and distribution from physical communication restrictions on the ground. UAV-assisted data distribution is straightfoward and manageable, as compared with traditional ad hoc routing over weak and multi-hop wireless networks. We exploit the notion of a UAV fleet that conducts teamwork for data gathering in such environments. To maximize the lifetime of a UAV fleet, we propose a fair and energy-balanced ferry fleet placement scheme, named as FEB for brevity, as jointly considering the efficiency, balance, and fairness of energy consumption. This scheme operates two mutual phases. The first phase demarcates service regions in a network using the powered-Voronoi diagram. The second phase decides a shortest-path itinerary across sensors in every service region using a genetic-based alternative of the traveling salesman problem (TSP) algorithm. Thus, this scheme is able to ensure fair task assignment, balance energy consumption, and prolong the lifetime among multiple ferries in teamwork. Performance results show that the proposed scheme outperforms several typical schemes, including Native, K-Means and Spiral, in terms of cumulative energy computation, residual energy distribution, Jain′s fairness index on energy utilization, the number of alive ferries, and the total of task execution times during UAV teamwork.
Using UAVs to collect sensing data in environments without existing network infrastructure is emerging as a favorable approach for various mission-oriented applications. Given that UAVs come with limited buffer and energy capacities, the effectiveness of resource utilization significantly influences UAV-assisted task execution. Our study accounts for data gathering challenges and aims to minimize the mission time required for collecting all sensing data from a set of target points of interest (PoIs) in a UAV-supported wireless sensor network. Our study initially determines the energy cost and the remaining buffer size across three UAV states: flying, waiting, and charging-offloading. We then express this issue of minimizing the mission time as a mixed-integer linear programming (MILP) optimization problem. Faced with the inherent complexity of this problem, which can lead to non-deterministic polynomial-time hardness (NP hard), we introduce an enhanced strategy termed "mission time optimization with energy and buffer constraints for multi-UAV deployment (MEBD)′′. Within this strategy, the weighted TSP (W-TSP) algorithm seeks a weighted shortest path that adheres to varied data delay budgets. Concurrently, the non-linear least squares based recharging route (NLLSR) algorithm employs a non-linear least squares method to refine UAV deployment, optimizing energy and buffer usage. Our simulation results reveal that the MEBP approach surpasses other baseline methods e.g., convex hull, RRT, RRT+, DTP, TSP, grid and AHTP-RL, achieving significant improvements in total mission time, energy consumption, and drop rate metrics.
Therefore, the contribution of this dissertation can achieve the energy-balanced mission time optimization in multi-UAV assisted data gathering systems. We believe that these efforts can be integrated to the current and new emerging wireless communication systems such as multi-access edge computing (MEC), mobile crowdsourcing, and space-air-ground integrated network (SAGIN). |
