In this paper, we propose a method to achieve projection mapping by mobile projector robot. Projection mapping technology become popular and is used in various situations. However, since most of the current projection mappings are based on pre-calibrated fixed projectors, projectable areas become small fixed regions. We have been researching Ubiquitous Display (UD) that is a mobile-projection robot and aims to support the users by projecting various information. In this paper, we propose projection mapping technology by the mobile projector robot. UD has sensors for self-localization, but their accuracy are not enough for achieving projection mapping. Thus, in our method, we first models projection target beforehand. Then, UD moves to the front of the target and obtains point cloud observations of the scene. After that, the pose and position of the model is estimated. By using the estimated pose and position, we can realize the accurate projection mapping based on the mobile projector. Preliminary experiment shows the effectiveness of the proposed method.

In this research, we are developing a care training robot for quantitative evaluation of care operation. In this paper, we propose a care training robot and how to evaluate the care operation. The proposed robot aims to imitate elderly people body and a care worker is able to gain experience or training by using the robot. It is possible to evaluate elderly care activity quantitatively by using sensors placed on the robot's body. Therefore, the system can return accurate feedback, and it is possible to nurture a care worker who acquired effective caretaking operation efficiently. Currently, in order to develop care training robot, we are developing a robot that imitates elderly elbow joints. This elbow joint robot can simulate symptoms that are common in elderly people. In cooperation with a care worker, we measured the joint angle and the force during the improvement movement of contracture, and evaluate the availability of the proposed robot.