Control of indoor thermal environments in accordance with people's preferences makes an　important contribution to comfort. To keep a suitable temperature environment, we tried to estimate a subject's thermal sensation using biologic signals. We focused on peripheral skin temperature to reduce the number of sensors in order to make an objective thermal sensation measurement system suitable for use in daily life. First we carried out an experiment involving alteration of environmental temperature to reveal the relation between peripheral skin temperature and thermal sensation vote. Next we made a thermal sensation processing algorithm and evaluated the algorithm. When thermal sensation indices estimated by our algorithm were compared with the subject's vote, error of mean squares was l or less in most cases. As a result, the possibility of thermal sensation measurement using peripheral skin temperature is confirmed.

The purpose of the present study is to investigate whether functional near-infrared spectroscopy (fNIRS) can be used as a new psychophysiological index for detection of deception. After completing a mock crime task,14 graduate and undergraduate students participated in a Guilty Knowledge Test (GKT). During the GKT, change in oxyHb level in the prefrontal cortex area was measured by fNIRS. As a result, the increase in the oxyHb level induced by the critical question of the GKT was higher than that induced by the non-critical question. This difference in the oxyHb level detected the lies in the GKT with 91%accuracy. The present study indicates that the fNIRS measurement can detect lies as good as conventional psychophysiological indexes such as respiration, blood pressure, pulse wave, or electrodermal activity.

It has been reported that most novice joggers encounter difficulties to get accustomed to new physical activities due to several reasons. Therefore, this research presents a system which supports novice joggers by keeping their motivation highly. The system was designed and developed with the following technologies; GPS (Global Positioning System, HRM (Heart Rate Monitor and SNS (Social Networking Service. Joggers wear the tiny loggers called GPS and HRM devices that can give physical log information to this system. Communities are organized by the system based on such information. It is also used for the plan and report provided by the system. Such individual reports are prepared and shared inside the communities in order to motivate the participants. Through these functions, two methodologies were considered and tested. Finally, two evaluations of the system revealed that the conceived methodologies were effective for keeping motivation.

When a lot of information are presented using AR annotation, a user may not obtain information efficiently. Display of selected annotations around gaze point would increase visibility of annotations by reducing the amount of information. However, when gazing far area, depth error of gaze point estimation becomes highly larger than horizontal and vertical errors. In this study, a method that estimates the gaze point by calibrating depth with the distance of Purkinje images was proposed. The result of the experiments showed a significant decrease of estimation error and suggested possibility of improvement of visibility with the proposed estimation method in comparison to the traditional one.

Abstract-Eye-fixation-related potentials (EFRPs)associated with the occurrence of fixation pause can be obtained from averaging electroencephalograms (EEGs)at the offset of saccades. A EFRP is a kind of event-related brain potential (ERP), which has a relationship with visual attention and the physical properties of visual stimulus. Auditory P300 is the ERP elicited from probe stimuli in a secondary task. The amplitude of a late positive wave (P300) is inversely proportioned to the amount of perceptual-central processing resources allocated to a primary task. In this experiment, the lambda responses of EFRPs and auditory P300 were examined concurrently with multiple tracking tasks to compare the effects of tasks of different difficulties. Eighteen participants were assigned six different types of tracking task, each for s min. The workload under each tracking condition was different in the task quality (the difficulty of perceptual-motor and/or perceptual-central levels). As a result, the peak amplitude of lambda responses decreased with the level of perceptual-motor workload, while the peak amplitude of P300 decreased with the level of both workloads. The results suggested that ERPs are a useful index of the information processing workload in human computer interaction.

This paper proposes an online chat system using speech balloons changed by user's biomedical signals. The system displays text of chat conversations in speech balloons which express user's emotion like a comic. The figure of a balloon that expresses arousal changes simultaneously with skin conductance (SC) on user's foot, and the color of a balloon that expresses emotional valence changes simultaneously with electromyogram (EMG) on user's forehead and cheek. Therefore, a user can input his/her emotion without superfluous operation. The system solves a problem of previous systems. The previous systems impose not only operations to input messages but also operative burdens to input emotions on users. The users, therefore, cannot keep a rhythm of conversation of a chat because of the superfluous operations to input emotion. We conducted an experiment to compare our system with a not-changing-balloon chat system without emotion. The results show that the operability of our system did not decrease in despite of the operation to input emotion because all users need to do is to smile or to bend his/her brows to input emotion while inputting a message. Also, the results show that our system is easy-to-understand emotions of a speech partner and is very pleasant.

The 'body image' is a sense of ownership of our own body. Some report shows the body image is induced by the integration of cross-modal perception, such as, visual, tactile and proprioceptive perception. Other report indicates that synchronized cross-modal stimuli could induce the body image to visible object and subjects experienced that the visible object was recognizable as a part of their body. Accordingly, the development of new interface system for virtual reality is expected by using the body image. However, the body image is subjective experience and is hard to observe objectively. In this paper, we examine whether the measurement of brain activation using Near Infra-red Spectroscopy (NIRS) would able to observe the body image which is induced by cross-modal stimulation. In this experiment, the brain activation was measured around the temporal lobe to the occipital lobe, and asynchronously visual and proprioception stimuli have given to the subjects synchronously to change their body image. As a result, brain activities were observed around the Superior Temporal Sulcus (STS), the inferior parietal areas and the Extrastriate Body Area (EBA) which is reported that related to distinction of other body. Controversially, Subjects reported that they experienced the body image to the CG model displayed as visual stimuli. Our result shows there are some possibilities that the body image would be induced by some inconsistent cross-modal stimuli.

A human watching moving images displayed on a wide-field display or screen sometimes suffers from the visually-induced motion sickness (VIMS)that induces symptoms related to the autonomic nervous system such as nausea, vomiting, and dizziness. Previous studies indicated that the maximum cross-correlation coefficient (ρmax) between blood pressure and heart rate, whose frequency components are limited to the neighborhood of O.1Hz, is a useful index to evaluate the effects of the VIMS on humans. The present study proposed a new method for obtaining Amax without measurement of neither continuous blood pressure nor ECG but using only finger photoplethysmography. The proposed method was evaluated on the basis of comparison with the conventional method carrying out an experiment in which multiple subjects were simultaneously watching the same swaying video image. The experimental result showed that the proposed method worked as successfully as the conventional method.

At present, 1 out of every s people in Japan experiences anxiety concerning sleep. We consider there are many people find it difficult to get up on time or find the experience of awaking unpleasant. We have developed a method of measuring variable indexes of pulse-to-pulse interval using pulse wave sensor, and a method of estimating the sleep states to reduce the influence of variation in the variable indexes to obtain frequency analysis of pulse-to-pulse interval due to circadian rhythm and the subject's condition. In this study, we describe a method for detecting change in sleep stage of a sleeping subject in real time, which uses variable indexes of pulse-to-pulse interval by pulse wave sensor, and we introduce a wake-up alarm application that is activated during change in sleep.

Noninvasive electroencephalogram (EEG)-based brain-computer interface (BCI)provides a new communication channel to replace an impaired motor function of patients. EEG recordings during right and left motor hand imageries are expected to be available to move a cursor to a target on a computer screen or to move a wheelchair. In this paper, similarities and differencies of EEGs among right/left hand movement sand right/left hand movement imageries have been studied to verify the usefulness of the EEG-based BCI during motor imageries. In the experiments, visual cues ("right hand" or "left hand") were presented for 3 s with the interval time of 4.5-5.5 s. The subjects were asked to move one of their hands or imagine it when visual cues were appeared. EEG signals were recorded from 63 Ag/AgCl scalp electrodes. First, blinking artifacts are eliminated by using ICA. Second, the wavelet transform was applied to obtain the time-varying energy of EEG signals during right/left movement or their imagery. Secondly, the power spectra were normalized. Lastly, topographes were analyzed from them. The event-related synchronizations (ERS)of the beta rhythms were observed after movements or their imageries in the vicinity of the motor area, and demonstrating beta rhythm is a promissing indicator for BCI.

When making Event Related Potential measurements, the intrusion of several artifacts, like eye movements, muscular activities and environmental noise, is an important problem. We assessed an independent component analysis method for dynamically eliminating such artifacts. Using this method we could extract the P3 component of the Eye-Fixation Related Potential (EFRP). Here we present the results of this method and also the effects of the workload caused by information processing on the P3 latency time, and also refer to the possibility to assess the cognitive load using P3 latency in EFRP.

In this paper, we measure electroencephalograms (EEGs) for quantitative evaluation of software usability. Two different (newer and older) versions of same software were selected to intentionally differentiate a subjects' mental state in our experiment. The versions have graphical user interfaces (GUIs) for the main window, which is extremely different from each other, such as different name of menus, commands and different appearance of command dialogs. While the older version is easy to use for our subjects, the newer version of the software is more likely to disturb users' tasks because it requires a user to restudy the interface of the software. Therefore, we can easily compare subjects' EEGs in an uncomfortable situation with that in a comfortable situation. In the experiment, we used alpha-wave, beta-wave, and beta/alpha as metrics of subjects' EEGs. As the result, we found significant differences of alpha-wave, beta-wave and beta/alpha between two versions. The results of questionnaires performed after the experiment sup-ports the differences, and we also found the correlation between the metrics and result of the questionnaire.

Nowadays, many college students possess PC as an educational study tool. However, PC is hardly used for taking notes of a lecture. The main reasons are as follows (a)a text input speed at which the student unfamiliar with a keyboard operation, is slow. (b)In an ordinary PC with keyboard and mouse, it is difficult to input figure information. It is expected that the input of a figure and/or character is easy for a tablet PC with a pen interface, and it is suitable for the tool taking notes of a lecture. In this paper, comparative evaluation of a tablet PC and an ordinary PC was carried out from the viewpoint of the input speed of figure and/or character information. In the experiments, we measured an input-speed of following 3 items: (1) input speed using an ordinary PC with keyboard and mouse, (2)input speed using a tablet PC with the pen interface, and (3)speed of handwriting on a paper. The speed of handwriting on a paper can be regarded as an input-speed of ideal pen interface, so we measured not only the input speed of a tablet PC used in the experiments, but also the input speed of ideal pen-interface.

In recent years, message communication using sound has appeared in various contexts in ordinary life. Sounds like these are called　"sign sounds". The purpose of sign sounds is to communicate a message. However the ease of message understanding and listening comfort have rarely been taken into consideration. Because people have different tastes and attach a different "image" to sign sounds according to their individual sensibility, it is important for each person to be able to make melodies freely and easily, and customize sign sounds. However, creating melodies is not easy for the ordinary per-son. For this reason, in this research we propose a sign sounds family design system using an Interactive Genetic Algorithm (IGA). An "Interactive Genetic Algorithm" (IGA) is a Genetic Algorithm (GA) which simulates the evolution of living things, where the evaluation part of the GA is handled subjectively by a human being. The proposed system produces three kinds of sign sounds which informs the user of start, finish and warning based on one melody generated by the IGA system. Experimental results showed that this system is effective for generating sign sounds family and generated sign sounds are liked by users.

Some video mirror interface systems capture the user's body image and displays it at real-size on the screen in front of him/her. Some virtual objects overlap the video image and the user can select one of them by his/her touching action. The interface is usable for easy and intuitive interactions in various situations. This paper focuses on the influence of the delay time on the object selection or operation in the video mirror interface. There are two kinds of delay time in such operation: image feedback delay and system response one. The experimental result shows that the operation is more sensitive to the image feedback delay than to the system response one. The image feedback delay of 0.03 seconds affects the static target selection while the system response delay of less than 0.15 seconds does not affect the selection. Furthermore, a task to keep touching moving target is affected by the image feedback delay of O.03 seconds but it is not by the system response delay of 1.0 seconds.

The lag effects of utterance to communicative motions in human-robot and in human-CG character greeting interaction have been clarified. In this paper, in-formation presentation timing by an agent based on the time lag effects of utterance to communicative action is analyzed by synthesis for applying to a human-agent interaction support. First, the analysis of a human information presentation action clarifies the average pause and average time delay of utterance. Then, the analysis by synthesis is performed by using an agent of CG character and confirms that the lags of utterance in certain information presentation timing bring communication effects, i.e. the moderate lag is familiar and the longer lag is polite. It is also clarified that it is unnatural to start motions and utterance simultaneously in human-agent interaction. These results demonstrate the importance of timing in human-to-agent embodied interaction and the method to perform an information presentation effectively.

We have proposed a method of embodied self-referable video communication where the virtual face-to-face scene of talkers is presented by using Chroma Key effect, so that two remote talkers can observe their face-to-face interactive behavior and share the embodiment in communication. The previous communication experiment demonstrated the effectiveness of the virtual face-to-face scene by sensory evaluation based on the Picture-in-a-Picture (PIP) scene for free conversation between friends. In this paper, this virtual face-to-face projection is compared with the PIP scene and the mirror scene in which own video image is superimposed on the other talker's video image (virtual mirror scene), for the examination of effective use. The effectiveness of the virtual face-to-face projection is demonstrated by sensory evaluation and human interaction analysis in the role-play of job interview between strangers for 20 pairs of 40 talkers.