人間のあいまいな表現を学習する音声制御ロボット

概要

論文の詳細を見る
This paper discusses the development of the interpretation system for the mobile robot manipulation which maps linguistic command with vague expression issued by a human operator to an appropriate quantity. The system allows the operator to use only few kinds of vague expressions, as can be seen in the case when a human instructs some other person orally. The problem here is to find an appropriate quantity for each operator because the quantity corresponding to the vague word which is implied by the operator differs when operator changes. To cope with the difficulty, this paper proposes the distance type fuzzy reasoning algorithm with iterative learning scheme that can accommodate individual difference on how to quantify the vague expression. The experiment has been carried out to examine the change in the number of instructions required to guide the mobile robot to destination. The result of the experiment shows that the number of instructions necessary for achieving the guidance task decreases in almost all cases, which implies the validity of the proposed system.