Abstract
The present paper describes the development of a laser pointer with stabilizing mechanism that aiming to suppress user's tremor effect. This paper discusses the recognition scheme of user's intentional motion to estimate essential desired motion related to user's intention from onboard sensors' information. In experiment, it was seen that measured acceleration profiles had different characteristics in-between "Stationary" phase and "Moving" phase. By applying a machine learning scheme (Random Forests), a recognition scheme was proposed. The proposed scheme successfully categorized the current state into three-groups, "stationary", "right / left-ward" and "up / down-ward" with high accuracy.