Abstract
〈Vol.6 No.4(2013.7)〉
Titles
[Contributed Papers]
▲
■ Reconstruction of Object Surface from Multi-View Images Improved by
Image Forces
Silesian University of Technology・Karolina NURZYNSKA, Kanazawa
University・Mamoru KUBO and Ishikawa National College of Technology・Ken-ichiro MURAMOTO
Currently available techniques for surface reconstruction from multi-view
images require a large number of images in order to assure high surface
resolution and precise detail reconstruction. Obtaining such big number
of images in some cases is impossible; moreover, it may impose high preparation
cost. Therefore, to address this drawback, a novel method is presented.
This method is based on the visual hull approach. It assumes to be able
to reconstruct three dimensional manifolds without edge from images, the
number of necessary images should be as small as possible, and the quality
of reconstructed mesh should be good. Applying this method for four images
from a dino dataset proved the proposed method works well.
▲
■ Skill Level Evaluation for Taijiquan Based on Curve Fitting and Logarithmic
Distribution Diagram of Curvature
Waseda University & CREST・Toshiyuki MURAO, Advanced Institute of Industrial
Technology・Yasuyuki HIRAO and Hiroshi HASHIMOTO
This paper proposes an objective skill level evaluation method for Taijiquan.
The proposed skill level evaluation method utilizes both curve fitting
to spiral motion and classification based on a logarithmic distribution
diagram of curvature for a human body part. In order to demonstrate a possibility
of the proposed method, the paper presents experimental results with a
Kinect sensor for three subjects, i.e., a beginner, intermediate and expert.
▲
■ A Design Method of Compensator to Minimize Model Error
Kumamoto University・Hiroshi OKAJIMA,Hironori UMEI,Nobutomo MATSUNAGA
and Osaka University・Toru ASAI
The robust control design method has been studied in recent decades. A
control system works well under the modeling errors and disturbances if
controller design is based on the robust control method. However, it is
well known that in control systems, generally, there exists a trade-off
between control performance and robustness. To overcome the trade-off problem,
this paper proposes an internal model type compensator structure that minimizes
the modeling gap between the nominal model and actual plant dynamics. By
using the proposed compensator, the dynamics of the compensated system
closes to that of the nominal model. Then, a design method of the compensator
parameters is also proposed for minimizing a set of plant dynamics. The
proposed design method is reduced to the standard μ design control problem.
If we use the proposed compensator for control systems instead of the plant
itself, the output performance might be better despite plant uncertainty.
Given that the proposed compensator can be used for the control of not
only linear but also nonlinear plants, we can easily achieve robust control
of nonlinear systems. The effectiveness of the proposed method is shown
by numerical examples.
▲
■ Acceleration of Reinforcement Learning by Controlled Use of Options
Given as Prior Information
Kyushu University・Kento TERASHIMA,Hirotaka TAKANO, and Junichi MURATA
Reinforcement learning is a method with which an agent learns an appropriate
action policy for solving problems by the trial-and-error. The advantage
is that reinforcement learning can be applied to unknown or uncertain problems.
But instead, there is a drawback that this method needs a long time to
solve the problem because of the trialand-error. If there is prior information
about the environment, some of trial-and-error can be spared and the learning
can take a shorter time. The prior information can be provided in the form
of options by a human designer. But the options can be wrong because of
uncertainties in the problems. If the wrong options are used, there can
be bad effects such as failure to get the optimal policy and slowing down
of reinforcement learning. This paper proposes to control use of the options
to suppress the bad effects. The agent forgets the given options gradually
while it learns the better policy. The proposed method is applied to three
testbed environments and two types of prior information. The method shows
good results in terms of both the learning speed and the quality of obtained
policies.
▲
■ Energy Resolution Enhancement of Retarding Type Electron Analyzer for
Photoelectron Thermometry
Yokohama City University・Ikuo KINOSHITA and AIST・Juntaro ISHII
A novel retarding type electron energy analyzer implementing an electrostatic
lens system to achieve a high energy resolution sufficient to measure thermodynamic
temperature has been investigated based on the electron trajectory analysis.
The energy resolution of the analyzer depends linearly on the retarding
energy with a modulus of 0.05 %. The voltages to be applied to the electrodes
in the analyzer are found to be controllable linearly against the retarding
voltage. The analyzer has threshold energies, whose effect can be excluded
in practical measurements.
▲
■ An Improvement of Transient Response of Dynamic Certainty Equivalent
Model Reference Adaptive Control System Based on New Smooth Parameter Projection
High Order Tuner
National Defense Academy・Masataka SAWADA and Keietsu ITAMIYA
A special adaptation law which has the ability to constrain adjustable
parameters of an adaptive controller to the specified convex space is often
used in an adaptive control system (In the following, this is referred
to as a projection adaptation law). In particular, the projection adaptation
law which guarantees the existence of time derivative or high order time
derivatives of adjustable parameters is referred to as a smooth projection
adaptation law and it is used for an adaptive control system based on the
adaptive back-stepping method or the dynamic certainty equivalent principle.
In the conventional projection adaptation law, the convex set to be constrained
is a hyper-rectangular or a hyper-sphere in order to accomplish a realization
of an adaptive controller, an efficient estimation and robust stability
of adaptation loop. This paper indicates new two smooth adaptation laws
(high order tuner) to improve the transient response of a control system
in addition to the ability of conventional restraint. Also, the authors
propose design schemes of a model reference adaptive control system using
them. They can constrain adjustable parameters to an internal space of
hyperparallelogram.
Effectiveness of the proposed scheme is illustrated by the theoretical
analysis and simple numerical simulation results.
▲
■ Data-Based Ground Fault Diagnosis of Power Cable Systems
Kyoto University・Iftikhar AHMAD,Hiroyuki MABUCHI,Manabu KANO,Shinji HASEBE,
Kansai Electrical Safety Inspection Association・Yoshikazu INOUE and Hiroaki UEGAKI
Ground faults are major problems of power cable systems. Time-series data
of voltage and current are available for diagnosis when a ground fault
occurs. In the present work, a data-based fault diagnosis system of power
cable systems was developed. In order to achieve the high fault diagnosis
performance, new feature variables were generated by using wavelet analysis
and cepstrum analysis. In addition, six classification techniques, i.e.,
k-nearest neighbor (k-NN), artificial neural network (ANN), boosted ANN
(B-ANN), random forest (RF), classification and regression trees (CART),
and boosted CART (B-CART), were compared. B-ANN and B-CART were combined
with the naive Bayes
classifier to cope with multiclass problems. The results of applying the
proposed methods to real ground fault data show that B-ANN and B-CART with
the naive Bayes classifier can achieve the best diagnosis performance,
which satisfies the requirement for its industrial application.
▲
■ Game Theoretic Receding Horizon Cooperative Network Formation for Distributed
Microgrids: Variability Reduction of Photovoltaics
Tokyo Institute of Technology・Yasuaki WASA,Takeshi HATANAKA,
Masayuki FUJITA, and The University of Tokyo・Hideaki TAKENAKA
This paper investigates cooperative energy network formation for distributed
autonomous microgrids based on receding horizon control and game theoretic
cooperative control. In particular, we focus on photovoltaics and aim at
minimizing its temporal and spatial variability while reducing transmission
losses over the whole network by forming an appropriate network of power
transmissions. We first formulate a novel optimal network formation problem
in the form of resource allocation games so that the welfare function reflects
the above objectives. Then, the problem is reduced to a potential game
through an existing utility design technique. The paper next presents a
variation of a learning algorithm presented in one of the authors’ previous
works and newly provide a proof of convergence in probability to potential
function maximizers. Moreover, we consider real time implementation of
the presented framework based on receding horizon control, where it is
shown that the information processing of the learning algorithm is almost
distributed with the helps of a solar radiation forecasting/estimation
system. Finally, this paper illustrates the effectiveness of the present
approach through simulation using real data of a solar radiation estimation
system.
▲
■ Racket Control for a Table Tennis Robot to Return a Ball
Nagoya University・Chunfang LIU,Nagoya University &・RIKEN・Yoshikazu HAYAKAWA,
and Nagoya University・Akira NAKASHIMA
This paper mainly proposes a racket control method for returning a table
tennis ball to a desired position with a desired rotational velocity. The
method determines the racket’s state, i.e., the racket’s striking posture
and translational velocity by using two physical models: the racket rebound
model and the aerodynamics model. The algorithm of determining the racket’s
state is derived by solving nonlinear equations and solving a two-point
boundary value problem of a differential equation. But this is not suitable
for a real-time process because of large computing time. The paper proposed
a modified algorithm which could be used for a real-time process by introducing
a simple aerodynamics model. Numerical simulations and experimental results
show effectiveness of the proposed methods.
|