Kyutacar : Kyushu Institute of Technology Academic Repository
Not a member yet
9746 research outputs found
Sort by
スマート農業におけるプラントデータネットワーク
Get PDF九州工業大学博士（情報工学）1 Introduction| 2 Existing Transmitting Data in Traditional Farming| 3 Internet of Things (IoT) and Cloud Computing-based Smart Plants Watering
System| 4 Mycelium Network Simulation for the Study of Communication Systems Concept| 5 Implementation of a virtual Mycelium Network for Smart Farming Security| 6 Conclusion and Future WorksSince it was first introduced in 1999 by computer scientist Kevin Ashton, the Internet of Things (IoT) has recently experienced rapid development. Sensor devices with a smartphone or other integrated devices bring enormous benefits and have changed human behavior, particularly time efficiency and effectiveness.
Those development processes still face many constraints, particularly data distribution, security, and node placement into the IoT infrastructure.
Those that make IoT vulnerable to intrusions can create damage in mistaken infrastructure management that can threaten IoT networks. We attempt to deal with those problems with several approaches to build a novel network architecture that can be implemented in an agricultural environment to monitor and collect data for increasing agricultural products.
The thesis studies novel data network architectures in smart farming. Traditional methods like note-taking, oral transfer of knowledge, and plain sensor data collection are still prevalent. New data network architectures need to be correctly integrated.
The thesis follows two approaches here. The first approach is the design of smart watering system based on predicion of soil moisture by using a ”Long Short-Term Memory” (LSTM) neural network, and LORA based remote communication.
Training data where collected for 3 months on several plants, and as a result, plant stress states could be correctly identified to above 90%. Thus, Deep Learning can become a key intelligent component for the prospected smart watering system also in practice.
The second approach is a novel communication network architecture inspired by the mycelium network. The communication between plants via the underground mycelium nework of fungi (better known as mushrooms) is known for years, but many of its aspects are still not understood. The thesis first explores the relation of such a mycelium network to TCP/IP-based internet and IoT networks. Then, a simulation environment in metaverse environment for mycelium growth and metabolism has been implemented. Experiments of the routing of plant attack signal in such a network were performed.
The experiments show the succesfull routing of such signals to specific other plants. At the receiving plant, the farmer can take protective measures.
Thus, the research points into a complete novel direction of smart farming by linking sensor signals, chemistry-driven routing principles, and virtualization, and showed its potential to improve safe operation of a farm.
The findings from this approach reveal that the concept of a mycelium network in a virtual environment can be used as a solution to overcome the problems of routes in the network, distribution of data and signals, and communication between sensors in network infrastructure.
This report proposes a more effective and efficient network system adapted from mycelium networks for smart agriculture. Furthermore, the network system concept we have developed can be a reference framework for building strong network infrastructure within the scope of IoT, especially Smart Farming. Furthermore, based on previous experiments, the results of predicted data collected from several types of sensors for the level of water content in the soil can be identified with an accuracy of up to 90%.
This result is the main component for building a smart watering system in an agricultural environment.九州工業大学博士学位論文 学位記番号:情工博甲第400号 学位授与年月日:令和6年6月28日令和6年度doctoral thesi
遺伝子摂動応答トランスクリプトームを用いた疾患横断的な治療標的予測
No full text九州工業大学博士（情報工学）第1章 序論| 第2章 ターゲットリポジショニング：遺伝子摂動応答トランスクリプトームを用いた治療標的予測| 第3章 GWASとTWAS の融合による希少疾患に対する治療標的分子の予測| 第4章 総括九州工業大学博士学位論文 学位記番号: 情工博甲第397号 学位授与年月日:令和6年3月25日令和5年度doctoral thesi
判別モデルを使用した小児前・後屈時腰痛の原因推定
Get PDF九州工業大学博士（情報工学）1.序論| 2.本研究の目的| 3.実験方法| 4.結果| 5.考察| 6.結論| 7.今後の展望近年 腰痛（low back pain: LBP）を持つ小児が増加し、世界的にも重要な問題となっている。LBPの85%は認識できる既知の症状に起因しない非特異的なLBPだとされる。 そのため、痛みの発生部位や原因推定を目的に、前屈時に疼痛が増強する前屈時 LBP （LBP during anteflexion：LBPAF）と、後屈時に疼痛が増強する後屈時 LBP(LBP during postflexion：LBPPF)に分類することが推奨され、分類による治療効果向上の報告が多 くある。しかしながら、小児 LBPAF と LBPPF の原因を調べた研究はない。小児 LBP に効 果的に治療介入するためには、小児 LBPAF と LBPPF の原因を明らかにする必要がある。 そこで本研究は、小児 LBPAF と LBPPF の原因を明らかにすることを目的に実施した。
目的達成のため、4-15 歳の小児 319名（女児 171名）、平均年齢 11.7歳（標準偏差 ±2.9歳）を対象に、アンケートと身体計測を実施してデータ収集を行った。アンケート結果から、LBPAF は 10歳以降で保有者がみられ、保有率は全体の 6％（女児 58％）で あった。LBPPF は、7歳以降で保有者がみられ、保有率は全体の 12％（女児 57％）であった。本研究では、全ての LBPAF と LBPPF 保有者を用い、アンケートと検査結果を含む 15変数を入力変数とした multiple logistic regression（MLR）を実施し、LBPAF と LBPPF の要因に寄与する変数を調べた（MLR_all）。さらに、本研究の LBPAF や LBPPF 保 有者のデータを過学習することを予防するため、汎用的で過学習性の少ない判別モデルを構築し、より一般的な小児 LBPAF と LBPPF の原因推定を目指した。判別モデルは、ステップワイズ法と Elastic Net（ENET）を用いた 2種の MLR と、Conditional Inference Forests（CIF）を構築し、交差検証を用いて判別性能と過学習性を評価した（MLR_CV、 ENET-MLR_CV、CIF_CV）。さらに MLR_all の結果と汎用的で過学習性の少ない判別モデルの結果を比べることで、判別モデルを用いた原因分析の有用性の確認も行った。
本研究で構築した 3つの判別モデルのうち、判別の性能や過学習性の観点から、LBPAF では MLR_CV が、LBPPF では ENET-MLR_CV が最良の判別モデルであると考えられた。上記で述べた MLR_all と MLR_CV および ENET−MLR_CV は、一部異なる説明変数を原因として推定した。構築した MLR_CV と ENET-MLR_CV のうち、判別性能が高いモデルのみを抽出して原因変数を調べ、小児 LBPAF においては、スポーツ実施時間の増加、大腿前面筋の柔軟性増加、大腿後面筋の柔軟性と脊柱の可動性低下、LBP 歴をもつことが小児 LBPAF の原因だと考えられた。さらに小児 LBPPF においては、LBP 歴を持つこと、大腿前面筋 の柔軟性増加、大腿後面筋の柔軟性と脊柱の可動性増加が小児 LBPPF の原因だと考えられた。
以上から、LBPAF と LBPPF は多因子が影響して発生すること、また2つの LBP の原因は異なり、原因に応じた介入が必要なことが示唆された。さらに、LBPAF の MLR_all と MLR_CV は一部異なる変数を原因として推定した。今回判別に寄与する変数を確認した MLR_CV と ENET-MLR_CV のモデルは、未知の LBPAF と LBPPF 保有者を判別できるモデル 2 であった。そのため、上記に記載した 2つの LBP の原因は、未知のデータにも適応する 結果だと推察でき、一般的な原因の推定に判別モデルを使用することは、有用であることが示唆された。九州⼯業⼤学博⼠学位論⽂ 学位記番号：生工博甲第490号 学位授与年⽉⽇: 令和6年3⽉25⽇令和5年度doctoral thesi
Facial Asymmetry Analysis Based on Temporal Changes in Landmarks and 3D Structure
Get PDF口唇裂は胎児の成長過程で唇が正常に形成されない先天異常であり，複数回の外科手術が行われる．本論文では，その手術の前後で顔の非対称性を定量的に評価するための手法を提案する．はじめに，疑似動画像の作成方法を述べる．次に，動画像を解析する方法を二つ述べる．一つは，フレーム間マッチングによって顔のランドマーク点の移動方向を検出し，それをベクトルで表すことによって表情の非対称性を解析する方法である．もう一つは，口唇部の領域を分割し，各小領域の重心座標によって形状の非対称性を解析する方法である．提案手法を疑似動画像に適用した結果，比較手法より高い精度でランドマーク点のマッチングを行えること，および感覚的な評価に矛盾しない非対称性評価を行えることが確認できた．We propose a method for the facial asymmetry analysis using a video image of 3D point cloud. First, we describe how to create pseudo video image. Secondly, we evaluate the asymmetry in two ways. One is the analysis based on the temporal changes of the face landmarks obtained by the matching process, which evaluates how asymmetrical the lip movement is. In the matching process, face landmarks in different frames are corresponded by comparing FPFH values. The other is the analysis based on the center of gravity, which evaluates how asymmetrical the lip shape is. We applied our method to ten pseudo video images. As an experimental result, it is shown that our method can match face landmarks with higher accuracy and the result of the asymmetry analysis does not contradict our sense.journal articl
ゲート電荷制御によるパワーMOSFETのスイッチング波形設計に関する研究
Get PDF九州工業大学九州工業大学博士学位論文（要旨）学位記番号: 生工博甲第493号 学位授与年月日:令和6年9月25日thesi
エラスターゼの効率的な近赤外検出とバイオイメージングを目指した新規スクアレイン-ペプチド複合体の研究
Get PDF九州工業大学九州工業大学博士学位論文（要旨）学位記番号: 生工博甲第500号 学位授与年月日:令和6年9月25日thesi
オープンソースのリモートセンシングデータと機械学習アプローチを活用したラオス人民民主共和国ナム・グム川流域（NNRB）における洪水危険度マッピング
Get PDF九州工業大学博士（工学）1 INTRODUCTION| 2 LITERATURE REVIEW| 3 STUDY AREA| 4 RESULTS AND DISCUSSION| 5 FLOOD RISK ASSESSMENT ON LAND COVER, POPULATION, AND CRITICAL FACILITIES IN NAM NGUM RIVER BASIN, LAO PDR| 6 CONCLUSION AND RECOMMENDATIONFrequent floods caused by monsoons and rainstorms have significantly affected the resilience of human and natural ecosystems in the Nam Ngum River Basin, Lao PDR. A cost-efficient framework integrating advanced remote sensing and machine learning techniques is proposed to address this issue by enhancing flood susceptibility understanding and informed decision-making. This study utilizes remote sensing geo-datasets and machine learning algorithms (Random Forest, Support Vector Machine, Artificial Neural Networks, and Long Short-Term Memory) to generate comprehensive flood susceptibility maps. The results highlight Random Forest’s superior performance, achieving the highest train and test Area Under the Curve of Receiver Operating Characteristic (AUROC) (1.00 and 0.993), accuracy (0.957), F1-score (0.962), and kappa value (0.91), with the lowest mean squared error (0.207) and Root Mean Squared Error (0.043). Vulnerability is particularly pronounced in low-elevation and low-slope southern downstream areas (Central part of Lao PDR). The results reveal that 36– 53% of the basin’s total area is highly susceptible to flooding, emphasizing the dire need for coordinated floodplain management strategies. This research uses freely accessible remote sensing data, addresses data scarcity in flood studies, and provides valuable insights for disaster risk management and sustainable planning in Lao PDR
In addition, the generated flood susceptibility map is used to analyze the possible effect on the different land use/land cover classes, populations and critical facilities. ANN and DNN outperform LSTM, achieving higher accuracy based on Receiver Operating Characteristics. The resulting flood susceptibility maps identify critical zones within the Nam Ngum River Basin at high risk of flooding, revealing that 36-53% of the basin area is highly susceptible, especially in low-elevation and low-slope regions. Additionally, 85-93% of the population is highly vulnerable to flooding within 261 to 296 km² of built-up area. Almost all of the critical facilities for health and education lie within the area, which is highly susceptible to flooding.九州工業大学博士学位論文 学位記番号:工博甲第594号 学位授与年月日:令和6年9月25日令和6年度doctoral thesi
竹由来リグニン：バイオマス由来グラフェン誘導体の作製とその機能性
Get PDF九州工業大学博士（工学）1 Introduction and Literature Review| 2 Lignin Extraction from Bamboo Biomass Via Autoclave-Assisted Organosolv Treatment| 3 Synthesis of Graphene Derivatives Utilizing Bamboo Fractionated Lignin as a Precursor| 4 Practical Application of Bio-Derived Graphene Derivatives in The Fabrication of Nanofiltration Membranes and as Bioadsorbents for Pollutant Removal| 5 Practical Application of Bio-Derived Graphene Derivatives in the Preparation of Sulphonated Catalysts for Biodiesel Production Using Microwave-Assisted Transesterification Method| 6 Conclusion and RecommendationsLignocellulosic biomass has been the subject of thorough investigation due to the increasing need for sustainable materials, specifically as a potential feedstock for advanced bio-based materials. As a novel source for graphene production, this present thesis scrutinizes the prospective applications of fractionated lignin from bamboo biomass, which is renowned for its carbon-dense structure and renewable nature.
Chapter 2 examines lignin extraction via autoclave-assisted methods utilizing a variety of acid catalysts and solvents, revealing that HCl has a lower purity but a higher recovery rate than H2SO4. The utilization of structural analysis improves understanding and offers valuable insights for optimizing extraction procedures. Chapter 3 of the paper delves into the synthesis of graphene derivatives from bamboo lignin using metal-catalyzed catalytic pyrolysis. The superior properties of Fe-catalyzed graphitic carbons are emphasized. Furthermore, the process of sequential conversion into bio-based graphene products, such as carbon quantum dots (CQDs) featuring distinctive structures, is clarified.
Chapter 4 explores into the utilization of graphene derivatives derived from bamboo lignin for the purpose of removing pollutants through the implementation of bioadsorbents and nanofiltration membranes. The results exhibit notable water permeability and effective expulsion of saline solutions and organic pigments. The successful elimination of phenol and tannic acid is observed to occur under particular conditions. In Chapter 5, an examination is conducted on bio-based graphene derivatives utilized as catalyst supports in microwave-assisted biodiesel production. The results demonstrate
improved catalytic activity, increased efficiency, decreased consumption, and shortened reaction time. In particular, efficacy is validated by kinetic and thermodynamic analyses conducted at elevated temperatures.
This study introduces bamboo fractionated lignin as a sustainable and adaptable precursor for bio-graphene derivatives, providing significant knowledge regarding its synthesis, potential applications, and far-reaching implications in various disciplines.九州工業大学博士学位論文 学位記番号:生工博甲第495号 学位授与年月日:令和6年9月25日令和6年度doctoral thesi