For a humanoid robot to make eye contact and initiate communication with a person, it is necessary to estimate the person's head position. However, eye contact becomes difficult due to the mechanical delay of the robot when the person is moving. Owing to these issues, it is important to conduct a head-position prediction to mitigate the effect of the delay in the robot motion. Based on the fact that humans turn their heads before changing direction while walking, we hypothesized that the accuracy of three-dimensional (3D) head-position prediction from a first-person view can be improved by considering the head pose. We compared our method with a conventional Kalman filter-based approach, and found our method to be more accurate. The experiment results show that considering the head pose helps improve the accuracy of 3D head-position prediction.

The aim of our study is to develop a method by which a social robot can greet passersby and get their attention without causing them to suffer discomfort. Social robots now function in a number of customer service roles, such as receptionists, guides, and exhibitors. However, sudden greetings from a robot can startle passersby. Therefore, we developed a method that allows social robots to adapt their mannerisms situationally based the results of related work. Our proposed method, user-centered reinforcement learning, enables robots to greet passersby without causing them discomfort (p<; 0.01). Our field experiment in an office entrance demonstrated that our method meets this requirement.

Social robots have been providing a number of customer services lately. For example, they can take the place of people as receptionists. When people interact they predict each others decision-making from the others actions and take suitable actions for the sake of the other. However, taking such actions is difficult for modern social robots. Therefore, our initial aim is to solve the problem of how to predict decision-making in human-robot engagement. Choosing a reception system as a case study to approach this goal, we created a new model to predict who will use the system. The model contains a state transition function based on observation studies. We evaluated the model though a controlled experiment and a simulation experiment, using field data on prediction performance and mental effect. The experiment results lead us to believe that the model can predict the decisions of others sufficiently well. We also found that a pedestrian who will not talk with a robot receptionist suffers a negative emotion when the pedestrian is greeted by a robot, but a method we propose prevents this. We suggest that the word suitable is related to negative emotion, a part of usability. By using our model in the future we will attempt to find a method that will enable a robot to learn suitable actions by itself.