Transactions of the Institute of Systems, Control and Information Engineers
Online ISSN : 2185-811X
Print ISSN : 1342-5668
ISSN-L : 1342-5668
Search
Advanced search
OR
Browse
Search
Volume 29, Issue 3
Displaying 1-7 of 7 articles from this issue
Show all abstracts
    • |<
    • <
    • 1
    • >
    • >|
Papers
  • Optimal Hyper-parameter Decision Method for Support Vector Machine Using Improved Response Surface Method
    Takayuki Sekine, Eitaro Aiyoshi
    2016 Volume 29 Issue 3 Pages 105-113
    Published: March 15, 2016
    Released on J-STAGE: June 15, 2016
    DOIhttps://doi.org/10.5687/iscie.29.105
    JOURNAL FREE ACCESS
    Show abstractHide abstract
    Support Vector Machine (SVM) has been studied actively as one of discriminators for a large number of data. In order to obtain a high identification rate, it is necessary to decide the best parameters of the SVM, and the problem to decide the parameters are formulated as an optimization problem, where the objective function is not described by mathematical formula, and a function call to evaluate the objective function value requires calculation quantity. In this study, to break through the problems, it is considered to adopt the Response Surface Method (RSM) with meta-heuristic algorithm to solve the SVM’s parameter optimization problem.
    View full abstract
    Download PDF (1589K)
  • Joint Angle Estimation of Upper Limb based on Evolutionary Multi-criterion Optimization using 3D Image Sensor
    Takenori Obo, Junya Kusaka, Naoyuki Kubota
    2016 Volume 29 Issue 3 Pages 114-121
    Published: March 15, 2016
    Released on J-STAGE: June 15, 2016
    DOIhttps://doi.org/10.5687/iscie.29.114
    JOURNAL FREE ACCESS
    Show abstractHide abstract
    In this research, we have developed a rehabilitation support system for analysis of upper body motion using 3D image sensor. Such a sensor has embedded systems to detect and track joint positions, but it is difficult to directly apply the data to calculation of inverse kinematics to estimate the joint angles because of errors in measurement. To solve the issue, some researchers have proposed methods of joint angle estimation using kinematic model and search technique. In these works, objective functions were defined to evaluate the similarity of posture between a person and kinematic model, and the problem was typically considered as a single-objective optimization problem. However, the data basically includes not only feature of personal behavior patterns but also noise. Therefore, we address a multi-objective optimization problem in order to scale back the influence of noise. This paper proposes a method of joint angle estimation of upper limb base on Evolutionary Multi-criterion Optimization (EMO). Furthermore, we apply a feedforward neural network to motion pattern modeling to realize a predictive search, combining with the EMO.
    View full abstract
    Download PDF (1762K)
  • Approximation Bayesian Reinforcement Learning based on Estimation of Plant Variation and its Application to Peg-in-Hole Task
    Kei Senda, Toru Hishinuma, Yurika Tani
    2016 Volume 29 Issue 3 Pages 122-129
    Published: March 15, 2016
    Released on J-STAGE: June 15, 2016
    DOIhttps://doi.org/10.5687/iscie.29.122
    JOURNAL FREE ACCESS
    Show abstractHide abstract
    In a general reinforcement learning problem, a plant, i.e. state transition probabilities, is estimated, and a learning policy for the estimated plant is applied to a real plant. If there is a difference between the estimated plant and the real plant, the obtained policy may not work well for the real plant. In this study, the real plant variation is parameterized by an interpolation of several estimated plants. This study proposes a reinforcement learning method based on estimation of parameter variation, and applies this method to 2-dimensional Peg-in-Hole Task. The effectiveness of the proposed method is demonstrated by numerical and experimental results.
    View full abstract
    Download PDF (1165K)
  • Privacy-preserving Crowd Movement Analysis with Fuzzy k-member Clustering-based Anonymization of Face Images
    Katsuhiro Honda, Masahiro Omori, Seiki Ubukata, Akira Notsu
    2016 Volume 29 Issue 3 Pages 130-135
    Published: March 15, 2016
    Released on J-STAGE: June 15, 2016
    DOIhttps://doi.org/10.5687/iscie.29.130
    JOURNAL FREE ACCESS
    Show abstractHide abstract
    Privacy preservation is an important issue in such personal information analysis as crowd movement analysis with face image recognition. This paper proposes a novel framework for estimating crowd movement characteristics without exactly distinguishing each person, in which personal authentication is performed in eigen-face spaces after fuzzy k-member clustering-based k-anonymization of feature vectors. An experimental result demonstrates that, supported by fuzzy partitioning, the novel framework can improve not only the noise sensitivity and anonymization quality of the conventional k-member clustering but also the reproducibility of crowd movement.
    View full abstract
    Download PDF (1134K)
  • Incremental Learning for a Calibration of a High-precision SAR-ADC by using the Inverse Calibration and Bayesian Regression
    Keiji Tatsumi, Toshimasa Matsuoka, Sadahiro Tani
    2016 Volume 29 Issue 3 Pages 136-142
    Published: March 15, 2016
    Released on J-STAGE: June 15, 2016
    DOIhttps://doi.org/10.5687/iscie.29.136
    JOURNAL FREE ACCESS
    Show abstractHide abstract
    Recently, a high-precision and low-power analog-to-digital converter (ADC) is required for a wearable biomedical measurement sensor which is driven by a battery. For achieving the purpose,a software-level calibration method was proposed for the successive approximation register ADC (SAR-ADC), which trains a calibration function with an incremental learning, and selects additional training data by using the Bayesian linear regression. Some numerical experiments showed that the desirable precision is obtained and the method needs a small amount of training data. However, since the additional training data specified by the selection method cannot necessarily be obtained in the practical viewpoint, its straightforward implementation is extremely difficult. Therefore, we propose a selection method of additional training data which can be obtained by using an inverse calibration,and moreover, we show the effectiveness of the proposed method through numerical experiments.
    View full abstract
    Download PDF (1040K)
  • Path Following Control for a Reference Path Consists of a Sequence of Points and an Application to a Boll & Plate System
    Kazuki Umemoto, Tadashi Egami
    2016 Volume 29 Issue 3 Pages 143-151
    Published: March 15, 2016
    Released on J-STAGE: June 15, 2016
    DOIhttps://doi.org/10.5687/iscie.29.143
    JOURNAL FREE ACCESS
    Show abstractHide abstract
    In this research, a path-following control system design for a path consist of sequence of points is proposed. To interpolating the points, a smooth path is made from the points as a polynomial curve. Non-inertial coordinate system is constructed that the origin is a reference point on the smooth path and axes are tangential and normal direction on the reference point. Movement of the reference point relates to a reference speed and a tangential component of the non-inertial coordinate. In order to implementing path-following, tracking control is achieved of a moving object to the reference point. Effectiveness of the proposed design is validated by control experiments with a boll and plate experimental system.
    View full abstract
    Download PDF (1399K)
Short Paper
    • |<
    • <
    • 1
    • >
    • >|
feedback
 Top