隨著溫室氣體的排放，極端氣候越來越常發生。面對極端氣候的肆虐，地面 通訊設備會遭遇到嚴厲的挑戰。颱風、海嘯及地震等天然災害都會較以往更加劇烈，容易使通訊設備造成損壞。其它像是停電、使用者數過多等皆會造成當地的網路壅塞。如何應對地面基地台超載而造成的通訊困難是本篇論文的方向。 本篇論文使用無人機基地台(UAVBS)來解決地面一些人口壅塞的地區像是活 動現場等。這些地區的地面基地台容易超載，需要無人機基地台來分擔一些負載。無人機基地台具有機動性，而無人機基地台位置的選擇是一個問題。本篇論文使用了機器學習中的分群方法來選擇無人機基地台的位置，包含了K-means 與DBSCAN (Density-based spatial clustering of applications with noise)。在選擇完無人機基地台的位置後，會探討在人口壅塞的地區在無人機基地台超載時使用調整無人機基地台最大發射功率的方式來改變cell coverage 以此來分攤負載的情形以及無人機基地台間的負載平衡。 由實驗結果顯示使用K-means plus DBSCAN 及調整無人機基地台的最大發射功率能達到更好的負載平衡及精準分配UAVBS 在地面基地台超載地區。 ;With the emission of greenhouse gases, extreme weather becomes more and more frequent. In face of extreme weather ravages, communication equipment on the ground will encounter serious challenges. Natural disasters such as typhoons, tsunamis and earthquakes will be more severe than ever, easy to cause damage to communication equipment. Others such as power outages, too many users, etc. will cause regional networks congestion. How to deal with the lack of communication caused by terrestrial base station overload is the direction of my paper. This paper uses unmanned aerial vehicle base station (UAVBS) to solve some of the population congestion areas on the ground, campaign etc. Terrestrial base stations in these areas are easy to overload, and the UAVBSs are needed to s hare some of the load. The UAVBS has the mobility, and the choice of the location of the UAV base station is a problem. This paper uses the grouping method in machine learning to select the location of the UAVBSs, including K-means and DBSCAN (Density-based spatial clustering of applications with noise). After selecting the location of the UAVBS, this paper will analyze the condition of adjusting UAVBS maximum transmit power to change cell coverage while UAVBSs overload on crowded area and analyze the load balancing between UAVBSs. Experiment results show using K-means plus DBSCAN and adjusting UAVBS maximum transmit power can get better load balancing and precisely arrange UAVBS on terrestrial base station overload area.
