{"title":"A Parameter Identification Method for Single-Input and Single-Output Linear Systems with Colored Noises Using Noise Gain Function","authors":"N. Komatsu","doi":"10.5687/ISCIE.34.29","DOIUrl":"https://doi.org/10.5687/ISCIE.34.29","url":null,"abstract":"A new identification method for single-input and single-output linear systems with colored input and output noises is proposed. Variances and autocovariances of the input noise are assumed to be known and those of the output noise are assumed to be unknown. The proposed method uses the eigenvector method and the noise gain functions which are derived in this paper. By using the noise gain functions, the variances and autocovariances of the output noise generated by the input noise can be obtained from those of the input noise. This algorithm uses an iterative calculation and it consists of two parts. One part is the identification of the system parameters using eigenvector. The other part is the estimation of the variance and autocovariace of output noises using the noise gain function. Some results of the simulation show the effectiveness of the proposed method.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122916725","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Consensus-based Approach to Distributed Scheduling Problem","authors":"T. Miyamoto, T. Umeda, S. Takai","doi":"10.5687/ISCIE.34.58","DOIUrl":"https://doi.org/10.5687/ISCIE.34.58","url":null,"abstract":"direction method of for to and show that the job shop scheduling problem (JSP) can be formulated by the proposed method. In distributed optimization, the optimization of is repeated until the convergence condition is but the scheduling problem is nonconvex optimization problem, the convergence by the proposed method is not guaranteed. In some a large number of iterations may be required to satisfy the convergence condition. In this paper, we propose two schemes to obtain a feasible solution within a smaller number of iterations. Then, the method is evaluated by computer experiments using benchmark instances of JSP.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117187703","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yoji Takayama, T. Urakubo, Takaki Tominaga, H. Tamaki
{"title":"Accuracy Enhancement of GNSS/INS Positioning in Dense Urban Environments with NLOS Signal Rejection based on Geometric Model","authors":"Yoji Takayama, T. Urakubo, Takaki Tominaga, H. Tamaki","doi":"10.5687/ISCIE.34.37","DOIUrl":"https://doi.org/10.5687/ISCIE.34.37","url":null,"abstract":"In this paper, we propose a NLOS (Non-Line-Of-Sight) signal rejection method to improve the positioning accuracy of integrated GNSS (Global Navigation Satellite System) and INS (Inertial Navigation System) system for a vehicle in dense urban environments. NLOS signals caused by reflection and diffraction always have positive measurement errors in pseudo-ranges, and they should be excluded in the Kalman filter of GNSS/INS, because the filter assumes that the measurement errors are zero-mean. In the proposed method, the positive errors in pseudo-ranges are geometrically estimated by simplifying the environments around a vehicle, and the signal that is supposed to be a NLOS signal based on the estimated errors is excluded from the measurements of the Kalman filter. We apply the proposed method to the measurement data obtained by actual driving in dense urban environments, and demonstrate that the positioning accuracy is improved by the proposed method.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129902568","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Time-series Modeling of Evacuees in Large-scale Disasters Considering Number of Potential Evacuees","authors":"Hiroshi Okajima, S. Kai, Shinnya Tokunaga","doi":"10.5687/ISCIE.34.47","DOIUrl":"https://doi.org/10.5687/ISCIE.34.47","url":null,"abstract":"Management of evacuation centers and flexible support of supplies are required when a large-scale disaster occurs. It is necessary to predict demand in order to transport relief goods quickly and appropriately to evacuation centers. Therefore, it is effective to predict the number of refugees in evacuation centers. Generally, the supply and demand trends vary not only due to the scale of the earthquake disaster but also depending on the occurrence situation, power outages and other conditions. In this study, we propose a dynamic model for predicting the number of evacuees. The effectiveness of the proposed model is illustrated by numerical example about case study of the Kumamoto earthquake.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114946789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Development of Automatic Verification for GUI System using Image Feature","authors":"Masatoshi Arai, Kazuhito Ito","doi":"10.5687/ISCIE.34.23","DOIUrl":"https://doi.org/10.5687/ISCIE.34.23","url":null,"abstract":"組込みシステムの設計現場では UMLや Simulinkを 用いるモデルベース設計が導入されている [1–3].モデル ベース設計の利点は,設計仕様のモデル化とシミュレー ションをPC上で行うことにより,モデルが設計仕様を 満たしているか容易に検証できることが挙げられる.検 証済みモデルは実機用プログラムコードに自動変換して 実装するが,PCと実機はハードウェア構成が異なり,計 算形式の違い(浮動小数点と固定小数点の違いなど)に よる計算誤差が発生する可能性がある.PCにおける計 算結果(仕様)と実機における計算結果との誤差が公差 内(許容範囲内)か否かを検査し,正しく実装できたこ とを検証する実装検証が必要となる. GUIを使った車載メータは運転者の視認性に基づき表 示の位置や大きさの設計仕様が決められている.PCと 実機のハードウェア構成の違いにより,画面表示位置等 のデータが公差内であってもメータ実機上の表示に公差 を超える差異が発生する可能性がある.そのため画面表 示の実装検証が必要であるが,PC画像(仕様)と実機 画像の単純な比較では公差を認める検証ができず,目視 による検証を行っている.GUIを伴う組込みシステムの 実証検証の効率向上が求められている. 本研究では,PCと実機の画面表示から抽出した特徴 点のマッチングにより表示の位置や大きさのずれを求 めることでGUIシステムを実装検証する手法を提案す る.提案手法が車載メータの実装検証に利用可能であり, GUIシステムの自動実装検証に有効であることを示す.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133384205","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Masayuki Tsuji, T. Isokawa, Masaki Kobayashi, N. Matsui, N. Kamiura
{"title":"Gradient Descent Learning for Hyperbolic Hopfield Associative Memory","authors":"Masayuki Tsuji, T. Isokawa, Masaki Kobayashi, N. Matsui, N. Kamiura","doi":"10.5687/ISCIE.34.11","DOIUrl":"https://doi.org/10.5687/ISCIE.34.11","url":null,"abstract":"This paper proposes a scheme for embedding patterns onto the Hyperbolic-valued Hopfield Neural Networks (HHNNs). This scheme is based on gradient descent learning (GDL), in which the connection weights among neurons are gradually modified by iterative applications of patterns to be embedded. The performances of the proposed scheme are evaluated though several types of numerical experiments, as compared to projection rule (PR) for HHNNs. Experimental results show that pattern embedding by the proposed GDL is still possible for large number of patterns, in which the embedding by PR often fails. It is also shown that the proposed GDL can be improved, in terms both of stability of embedded patterns and of computational costs, by configuring the initial connection weights by PR and then by modifying the connection weights by GDL.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132177552","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Probabilistic Modeling for Nonlinear and Non-Gaussian Systems Using Kernel Density Estimation and Gaussian Process","authors":"Y. Kaneda, S. Suzuki, Y. Irizuki","doi":"10.5687/ISCIE.33.303","DOIUrl":"https://doi.org/10.5687/ISCIE.33.303","url":null,"abstract":"Recently, in order to model systems with uncertainty, probabilistic modeling with conditional probability distribution are widely noticed. Modeling the systems as probability density function enables us to apply strategies considering unsertainty to control systems. One of the most famous modeling methods is Gaussian process regression (GPR). GPR is often used in the field of control systems and its effectiveness is demonstrated in many papers. However, since GPR can represent only Gaussian distribution, GPR cannot always achieve good performances for systems under non-Gaussian noise. This paper focuses on kernel density estimation that can approximate arbitrary probability distributions and proposes a probabilistic modeling method for non-Gaussian systems including ones with dynamic characteristics. At that time, a proposed method assumes that measurement is written as a sum of noise and function. Moreover, a prior distribution of the function assumes to be Gaussian. Under the assumption, the proposed method estimates the function as hidden variables by variational Bayes methods. The proposed method can model arbitrary probability density function for single output systems from data. Numerical simulations demonstrate the effectiveness of the proposed method.","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126821328","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Reduction of Learning Time for Differential-Wheeled Mobile Robots by Knowledge Transfer for Real-Time Learning","authors":"N. Kotani","doi":"10.5687/ISCIE.33.317","DOIUrl":"https://doi.org/10.5687/ISCIE.33.317","url":null,"abstract":"強化学習 [1]は,エージェントが環境から得られる報酬 を頼りに自律的に行動を獲得する手法である.これまで に,ロボットの動作生成への活用を目指した研究等,さ まざまな領域への展開が模索されている.強化学習をロ ボットの動作生成に用いる利点は,設計者がロボットの 具体的な動作を設計しなくてもよい点である.反面,強 化学習は,エージェントが試行錯誤を通して,環境に適 した行動を自ら学習するため,一般的に,多くの試行回 数を必要とする問題がある.とくに,エージェントが複 数のタスクを扱うような場合,未知のタスクに遭遇する たびに学習し直すことは効率が悪い. この問題に対して,過去に獲得した知識や,あらかじ め与えられた知識を応用することによって,学習にかか る試行回数を抑制することが考えられる.このような考 え方を実現する方法として転移学習があり,近年の機械 学習分野の研究でも注目されている話題の一つである. この転移学習を強化学習に適用することで,学習能力の 向上を目指した研究事例が報告されている [2–4]. 筆者らは,これまでに,遺伝的アルゴリズムの交叉・ 淘汰・突然変異の考え方を取り入れた強化学習手法を提 案してきた [5].そして,数値シミュレーション上の多リ ンク型ロボットアームによる経路獲得問題を対象として, 学習能力の向上とシミュレーションに要する時間を大幅 に削減できることを示し,提案手法の有効性を明らかに してきた.しかしながら,これまでのシミュレーション は,摩擦やすべり等の物理的な作用を反映しておらず, 現実環境下での学習を想定した場合の有効性については 明らかではなかった.この点において,実ロボットを用 いて実験することも考えられるが,実ロボットを用いた 実験には環境ノイズなど,さまざまな要因が含まれるた ∗ 原稿受付 2020年 8月 28日 † 大阪工業大学 情報科学部 Faculty of Information Science","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129377537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Active Sensing and Control of a Mobile Robot Considering Uncertainly of Path Prediction","authors":"Tomohiro Kitaoka, Y. Minami, M. Ishikawa","doi":"10.5687/ISCIE.33.314","DOIUrl":"https://doi.org/10.5687/ISCIE.33.314","url":null,"abstract":"近年,自動車や移動ロボットの自動化や知能化に向け 多くの取り組みが行われている.自律的な移動を実現す るためには,周辺環境の「認識・予測」とロボットの「行 動制御」が求められる.環境認識には,カメラや LRF, LiDARなどのセンサが用いられる [1].しかし,各種セ ンサから得られる情報は,ノイズやハードウェアの制約 などから不完全である場合が多い [2]. このような不完全情報のもとで,自律的な移動を実現 するために二つのアプローチが考えられてきた.一つは, 認識・予測の精度を高めること,もう一つは,認識・予 測の不確実性を補う制御系の構築である.認識・予測の 精度向上を目的として,ハードウェア面ではセンサの高 性能化に向けた開発,ソフトウェア面では,不完全な観 測情報の補間や予測に関する研究が数多く行われてい る [3].また,ロバスト制御 [4]やモデル予測制御 [5]を用 いて,認識・予測の不確実性を補う制御系の構築に向け た研究が進められている. 一方で,センサ系に運動を付加することによって,セ ンシング領域を拡大したり,精度を向上させたりする能 動的センシングという方法が知られている [6].これは, 認識・予測と制御をつなぐ技術として期待されている. 一般に自動車や移動ロボットでは,センサ位置は固定さ れている.これにより,センシング領域が制限されるこ とで認識・予測の精度が低下する場合がある.この問題 に対応するためには,たとえば,一時的に速度を落とし て認識を繰り返しながら移動するということが考えられ る.つまり,不確かさに応じて車やロボット本体のふる まいを制御し,能動的センシングを実現するということ","PeriodicalId":403477,"journal":{"name":"Transactions of the Institute of Systems, Control and Information Engineers","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114681759","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}