In this study, we aim to support the maintenance of opportunities for regular reflection on daily life for college students who have insufficient opportunities for reflection. For this purpose, we focus on the reflection through text chat with a conversational agent, and introduce parent factors to the agent in order to enhance familiarity and ease of talk to the agent and to encourage reflection on daily experiences. In this study, the parent factors are defined as “a casual way of speaking” and “questions that concern child’s current situation.” In the experiment, we verified the impact of repeated reflection through text chat with the agent assigned parent factors on the ease of reflection, motivation to reflect and impression of the conversational agent. So, the experiment was conducted a comparison in which participants were asked to reflect regularly in two conditions; with and without “Parent Factor Conversation.” As a result, we confirmed that with each reflection, the ease of reflection improves in both conditions, and this tendency is higher in the condition applied the parent factors. Additionally, it is shown that the parent factors significantly improve the impression of the agent, irrespective of the number of reflections.

Traffic accidents caused by older drivers have become a social problem in recent years. It is necessary to correct their overestimation of their driving ability for older drivers to continue driving safely. Effective and efficient safe driving education is required as the aging society progresses. Therefore, we proposed an efficient education method to correct older drivers’ overestimation of driving safety. The proposed method includes simulated collision and near-miss experiences with a driving simulator and driving reviews using an app. The app is efficient because it allows education without human intervention. We extracted elements from the dialogue content by humans in our previous study and added teaching remarks at the end of the dialogue to make the education effective with limited dialogue content. In addition, an agent was introduced to review the driving experience to make it easier for the older drivers to accept the content of the dialogue. As a result, the proposed method achieved a higher educational effect than the review done by humans. This result showed that education can be done by the app and without human intervention and that the agent facilitates the acceptance of older drivers and improves the educational effect.

COVID-19 and technological developments have increased opportunities for remote communication, including online meetings. This has led to a focus on “telepresence avatar robots”. The telepresence avatar robot takes an approach to support communication by generating a realistic sense of conversation by incorporating technologies such as camera motion and eye-contact with the speaker. However, these approaches do not provide intention estimation and do not achieve smooth communication Compared to local communication, it is difficult to achieve a sense of presence for the robot operator. In contrast, we implemented eye-tracking, which is behaviored unconsciously by humans, as an autonomous motion, and verified whether the robot operator performs intention estimation by these autonomous motion. Eye-tracking is an action in which one tracks the other’s gaze in a local conversation, and is performed unconsciously. The experimental results suggest that autonomous behavior generates gazing intention estimation, and that this approach may be a useful method for robot operators to obtain a sense of presence in the environment and to predict the behavior of others.

In this study, we have developed a video agent generation algorithm with a hybrid structure that combines multiple functions based on deep neural networks and mathematical models. The hybrid structure is composed of multiple elements in a Low-level neural network-like structure, and the coupling relation of the network smoothly transitions according to the human behavior. The algorithm enables the real-time generation of the video agent and its behavior to change in various ways depending on the actions of a person. As a result of interaction experiments, we confirmed that cooperative behavior between the human and the video agent occurs in the experience of the dynamic system with the algorithm applied.

In human-agent interaction research, agents that build friendly relationships with users in everyday life have been studied. As a result, there has been very little research on non-everyday life like agent, “Inou(extraordinary) agent”. However, the aspect of enjoying unusual experiences is also useful for users. This study reveals how anthropomorphic/zoomorphic creatures of the past, such as “Yo-Kai” can help in designing the interaction of AI and robots. This study also discovers the significance of the unusual experience that seems to be an important to provide enjoyment for users and future society.

In recent years, the introduction of ICT education in the field has sparked interest in the development of educational support robots. While teacher-type robots teach learners, button-based support may foster dependency. To address this, we proposed a method using deep learning to estimate learners’ perplexed state through facial expressions. Additionally, we developed an apprenticeship promotion model, combining cognitive apprenticeship theory and the perplexion estimation method. Experiments confirmed that the model-equipped robot yielded comparable learning effects to conventional robots for university students. In this paper, we design a self-study room for junior high school students, investigating their engagement with the robot.

Self-driving cars are truly autonomous robots in that they control actuators based on information from a variety of sensors. When the autonomous driving system is adapted to be a “social robot”, what kind of interaction and communication with passengers will be possible? In this study, we propose “RunRu”, a Sociable Vehicle with autonomy, sociality, and operability, as a research platform for discussing these issues, and report the results of an experiment to evaluate its configuration.

Our research assesses the gap between people’s true feelings and desires and their behavior and opinions in public. In this research, we focused on “internal desire” as one of the internal states of an agent and discussed a method of expressing desire by combining voluntary behavior and heartbeat expression as an involuntary physiological phenomenon. In this experiment, first, we verified the expression of desire by behavior or heartbeat expression individually. Next, we verified how the strength of internal desire and the suppression of desire change through a combination of agent behavior and heart rate, and we analyzed their relationship. The results showed that both desire behaviors, which are gazing and approaching the object, and the different patterns of heart rate changes affected the intensity of desire. In addition, the combination of large heart rate changes and low desire-level behaviors expressed the agent’s suppressed desire state, which suggested the agent’s true internal state of wanting but holding back.

Existing literature has proposed the use of conversational agents/robots’ non-verbal behaviors to smooth communication between agents/robots and humans. The use of turn-taking behaviors by conversational agents/robots has also been studied to improve communication. However, few studies have investigated whether people distinguish conversational atmosphere types from turn-taking behaviors. In this study, we developed a character conversation simulator by controlling the degree of overlap and investigated the association between turn-taking behaviors (overlaps/gaps) and conversational atmosphere types. The experimental results demonstrate that the major conversational atmosphere types can be determined by changing only the degree of overlap.

We have developed a Tangible Hand-writing System for VR using a magnetic motion capture device and an immersive HMD, which can be operated intuitively by finger touch. In order to evaluate the usability of this system in comparison with a tablet device, we constructed two types of tracing measurement systems, a tangible-type system and a tablet-type system. We conducted an experiment to measure the tracing task of Kanji characters at two different writing speeds. The results showed that the tangible system was superior in subjective evaluation by the participants and in a comparison between starting point gaps.