All issues
Volume 49, Issue 3
Displaying 1-12 of 12 articles from this issue
- |<
- <
- 1
- >
- >|
Paper
-
Mamoru SAITO, Katsuhisa KITAGUCHI, Gun KIMURA, Masafumi HASHIMOTO2013 Volume 49 Issue 3 Pages 319-325
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSThis paper presents a method for automated human detection using fisheye image. We introduce a probabilistic model to describe the wide variation of human appearance in hemispherical image. In our method, human is modeled as probabilistic templates of body silhouette and head-shoulder contour. These template features are extracted from the human images taken at various distance and orientation with respect to the camera, and form the training data set for model creation. A probabilistic appearance model is built by using the combination of principal component analysis (PCA) and kernel ridge regression (KRR). Finally, the problem of human detection is formulated as maximum a posteriori (MAP) estimation using above model. Experiments are conducted on indoor space where a fisheye lens camera is installed on the ceiling of crossing hallway. The feasibility and accuracy of our method is discussed through the experimental results.View full abstractDownload PDF (1474K) -
Fuyuki HANE2013 Volume 49 Issue 3 Pages 326-335
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSThis paper proposes three methods to reduce a tracking error attributable to tidal current in a curved tracking of track control system for a ship. In the curved tracking because the heading is changed even if the current is constant, the current component acting to the ship is changed. Present linear estimator that derives from reference coordinate system contains yaw elements and cannot be operated because of becoming non-linear in the curved tracking. The first proposal method derives the linear estimator from earth-fixed coordinate system not to contain the yaw elements. To realize a constant radius route using trajectory tracking control, the present method updates the turning rate by the updated speed obtained from the updating condition that compares the forward ground speed to a threshold value. It also corrects the sideways velocity component of the current by a side-slip angle through a feed-forward control. To reduce the present radius error attributable to the discontinuous update operation, the second proposal method modifies the updating condition, the updated speed, and the threshold value. The present feed-forward rudder angle for the side-slip angle correction has a case that exceeds the preset value. The third proposal method analyzes the maximum feed-forward rudder angle and defines a conditional expression not to become the maximum greater than the preset value. The proposed methods verifies the validity by simulations.View full abstractDownload PDF (1356K) -
Masayoshi DOI, Kazuhisa NAGAMOTO, Kenichi IDENAWA, Yasuchika MORI2013 Volume 49 Issue 3 Pages 336-344
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSThe purpose of this study is a design of an automatic peering system with a Controllable Pitch Propeller (CPP) in the case of the stormy weather. A velocity response of CPP has a delay time. Especially, this study verifies the ship's propelling force against the CPP angle and detects the delay time by analyzing the data of the propelling force. The CPP angle is controlled by Generalized Minimum Variance Control (GMVC). This study arranges the GMVC for the automatic mooring system. The proposed GMVC has wind and a tidal effect's terms. The wind effect's term is considered by the ship's drift with the wind. The predicted wind velocity is calculated by the AR model of wind change. The AR model of wind is used for the GMVC of the peering system.View full abstractDownload PDF (1288K) -
Hiroki UMEDA, Noboru SAKAMOTO, Yoshio UMEMURA2013 Volume 49 Issue 3 Pages 345-352
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSThis paper deals with the optimal control problems of the nonlinear system with state or input norm constraints. To address the constrants, nonlinear weighting functions are introduced to a cost function. An optimal control law is solved via the stable manifold approach which is recently proposed for solving the Hamilton-Jacobi equation. The behavior of the proposed controller coincides with that of linear optimal controller within the constraints. The effectiveness of the proposed method is verified by a numerical example and a DC motor experiment.View full abstractDownload PDF (1309K) -
Yasuhisa HIRATA, Yosuke OJIMA, Kazuhiro KOSUGE2013 Volume 49 Issue 3 Pages 353-361
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSIn this paper, we propose a motion control algorithm of multiple passive mobile robots for handling a large object in cooperation with a human. The passive mobile robot called “O-PRP(Omni-directional type Passive Robot Porter)” has passive dynamics with respect to an applied force and its motion is controlled based on the servo brakes. In the proposed algorithm, each O-PRP is controlled as if it has an apparent dynamics, and the motion characteristics of the object supported by the multiple O-PRPs is changed for improving the maneuverability. The proposed algorithm is applied to two O-PRPs, and experimental results illustrate its validity.View full abstractDownload PDF (1152K) -
Kazuto KAMOI, Tomohito YAMAMOTO, Yoshihiro MIYAKE2013 Volume 49 Issue 3 Pages 362-369
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSIn this paper we focused on a pause in the speech, and analyzed the factors affecting pause duration. It has been considered that utterance duration just before the pause is the only factor affecting pause duration (preboundary effect), recently effect of utterance duration just after the pause has also been noticed (postboundary effect). However, the relation between two utterance durations and pause duration sandwiched by the utterances has not been analyzed. Therefore we analyzed these factors inclusively. We proposed a simple utterance phrase (XY utterance phrase) consisting of two words in speech experiment and analyzed the relation between pause duration and utterance durations. The results showed that utterance duration just before the pause affected pause duration, and suggested that the relation between prior and posterior utterance duration affected pause duration. These results mean that not only a pre or postboundary effect but also a pre-postboundary effect may exist in speech, and we discussed its mechanism.View full abstractDownload PDF (942K) -
—Solution of Dilemma Problems—Shota YAMAWAKE, Yasuaki KUROE, Hitoshi IIMA2013 Volume 49 Issue 3 Pages 370-377
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSIn this paper, we propose a swarm reinforcement learning method for dilemma problems of multi-agent tasks in which it is difficult for agents to learn cooperative actions. In the proposed method, multiple sets of the agents and the environments, which are called learning worlds, are prepared and each agent in each world learns through exchanging information with agents in all other worlds. In order to acquire the cooperative actions, we propose a method of information exchange in which the agents in all learning worlds share the state-action values which are estimated to be superior for taking cooperative actions. The proposed method is applied to the N-persons iterated prisoner's dilemma and the Tragedy of the commons that are typical dilemma problems, and its performance is evaluated by investigating the learning processes and results.View full abstractDownload PDF (797K) -
Hideo MUROI, Takashi SHIKIMORI, Shuichi ADACHI2013 Volume 49 Issue 3 Pages 378-385
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSIn this paper, a new Local Linear Partial Least Squares (LL-PLS) method which is a nonlinear system identification method using local models is proposed. A procedure of the local linear PLS method is that first local models are identified using conventional methods like Recursive Least Squares (RLS) method, then a virtual system which describes a nonlinearity of the real system is identified using PLS method. In this paper, a derivation of this method is described, and then effectiveness of the proposed method is shown by numerical simulations.View full abstractDownload PDF (837K) -
Hiroshi SATO, Tsukasa WATANABE, Hiroyuki FUJIMOTO2013 Volume 49 Issue 3 Pages 386-393
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSIn order to evaluate the performance of ultraprecision machining, precise and accurate roundness or form profile measurements of a glass hemisphere are required. In such measurements, the multistep method is a commonly used self-calibration method. This method allows the roundness error of the workpiece to be separated from the spindle error of the machine. Recently, the phase combination method has also been used for high-accuracy measurements. This method produces results equivalent to those obtainable using a large number of steps, by combining two sets of results for different smaller numbers of steps. However, there have been no reports describing the uncertainties associated with the estimated roundness profiles using these methods. If the required accuracy is above the specifications, it is necessary to measure of a larger number of steps, which is both time consuming and leads to wear of the workpiece. This paper describes the estimated uncertainty in the profile for roundness measurements using both the multistep and phase combination methods. Based on the estimated uncertainty, the optimum number of steps to achieve a specific level of precision can then be predicted. Furthermore, for the phase combination method, the optimum combination of step numbers is suggested.View full abstractDownload PDF (1069K) -
Shunsuke ISHIMITSU, Kenji TAKAMI, Yoshiharu SOETA, Seiji NAKAGAWA2013 Volume 49 Issue 3 Pages 394-401
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSThe traditional production concept for car engine sound has changed from finding a solution to unwanted noise, to designing a particular sound. Although many studies have investigated the development of comfortable car-engine sounds, the psychoacoustic effects of time-varying rate for accelerating engine sounds remain unclear. Thus, we investigated the effects of increasing the frequency of interior car noise on auditory impressions, using psychological and neurophysiological methods. We carried out two investigations. First, we examined the relationship between the impression of 'sportiness' from the dynamic characteristics of the engine sound and brain magnetic fields. Complex harmonic tones simulating acceleration noise were used as stimuli, and subjective evaluations were examined using the semantic differential method. Neuronal activity in the auditory cortex was measured using magnetoencephalography (MEG). Second, we examined the relationship between subjective preference and brain magnetic fields for the simulated interior noise of an accelerating car. Subjective evaluations were examined using a paired-comparison method. The MEG alpha-wave range (8-13Hz) was measured and analyzed using the autocorrelation function. Neuronal activity was longer following stimulation by a preferred sound. Our results indicated that the particular characteristics of engine sounds have a significant effect on subjective impressions and neuronal activity in the auditory cortex.View full abstractDownload PDF (1157K) -
Masaki INOUE, Masayasu SUZUKI, Kenji KASHIMA, Jun-ichi IMURA, Kazuyuki ...2013 Volume 49 Issue 3 Pages 402-407
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSThis paper is concerned with instability analysis of Lur'e systems. First, a definition of absolute instability is introduced. Then, linear matrix inequalities and equivalent frequency domain inequalities are derived for absolute instability. They are analogous to well-known circle criteria for absolute stability.View full abstractDownload PDF (415K)
Short Paper
-
Koji ISHIHARA, Yuki OHMURA, Tadashi HORIUCHI2013 Volume 49 Issue 3 Pages 408-410
Published: 2013
Released on J-STAGE: March 26, 2013
JOURNAL FREE ACCESSIn this research, model-based reinforcement learning was applied to the six-legged robot. The objective for this robot was to acquire efficient walk movement by model-based reinforcement learning. By setting several constraint conditions on motors of this robot, we considered the reinforcement learning problem with three-dimensional state and action space. Through experiments, the effectiveness of our method using value iteration was revealed in comparison with Q-learning.View full abstractDownload PDF (380K)
- |<
- <
- 1
- >
- >|