It is important to grip an object accurately and recognize the shape of the objects for the potential future robot hand. We have developed the multichannel flexible tactile sensor having finger print applicable for an artificial robot hands. The tactile sensor was consisted of four pneumatic sensors and a flexible plastic cap with a finger print structure that was made of silicone resin and was separated from four air spaces. We evaluated the slipping characteristics of the tactile sensor. Evaluation tests showed that the multiple soft tactile sensors detected the friction between the tactile sensor and the object, and the results that the objects were easy to slip or not.
Following natural disasters, accidents, and shocking incidents, some children experience mental disorder. A respiration control method that relaxes the body and mind may efficiently prevent and improve mental disorder. Therefore, we developed a stuffed toy using two airbags to measure the respiration wave and lead a child's respiration using the up-and-down movement of the toy's abdomen to help them relax. After evaluating that children's respiration could be measured and guided by the stuffed toy, we then performed an experiment of relaxation effect. Participants in the experiment consisted of 9 healthy girls aged 8-10 years old. We measured the difference of heart rate variation between "Hug condition" (just hugging the stuffed toy) and "Respiration-Leading (RL) condition". The results showed that RL condition made children's heart rate variation lower than Hug condition. This means that the relaxation effect of respiration-leading by the stuffed toy is larger than just hugging it.
Master-slave remotely controlled manipulators such as the da Vinci systems have been developed to support for laparoscopic surgery. The manipulator is the all-in-one system constructed by integrating sub-systems such as the tool mounted arm mechanism, the DC servo actuator and the bilateral feedback controller. Thus, it is difficult to apply another development sub-system for the manipulator. To this end, we propose a modularized manipulator constructed by the tool changeable wire-driven flexible arm, the wire wind-up pulley and the actuator unit which can be interchanged between stepper motors and DC motors.
In this study, we propose a monitoring system for elderly people living alone utilizing the interaction between the elderly and robots. Not only sensors but also small robots are arranged all-around the house. The robots encourage the elderly to speak to them spontaneously by holding a relatively simple conversation. The elderly are informed that their approaches to robots are recorded. When there is no approach from the elderly in spite of a detection of existence, a robot speaks to them. If elderly do not respond, the robot gives the notice to guardians. The slow gradual changes in patterns of living are also detected and reported by analyzing the recordings. The guardians can access to the reports whenever they can. In this presentation, we describe the concept of the proposed system, a part of implementation and a method to detect the abnormality from the recordings.
人間がどのように行動を学習し，計画・決定しているのかは，非常に興味深い問題である．またそうした行動計画・決定や言語理解，思考などの高次機能の獲得は，ロボット工学においても依然として未解決の問題である．本稿では，こうした点を解明する一助として，概念・行動・言語・プランニングを同時に学習する枠組みを提案する．これを多層マルチモーダルLDA (mMLDA) を中心に様々なモジュールを統合することにより実現する．mMLDAと強化学習の統合により，理解に基づく行動を可能にする．また，mMLDAとベイジアン隠れマルコフモデルの統合により，文法学習を行い，ロボットが自分の行動を言葉で表現し，ユーザーの発話を理解することを可能にする．さらに，mMLDAと隠れマルコフモデルの統合により，長期的なプランニングが可能になる．実データを用いたシミュレーション環境で実験を行い，提案するモデルの有効性を検証する．