Abstract
When intelligent machines coexist with humans, it is important that the intelligent machine's movement in response to human instruction is acceptable by the human state of mind. When a human instructs an intelligent machine with his or her finger, we focus on a simple and easy interface via which conversion of the finger indication to intelligent machine motion is realized by a transfer function. In this paper, we evaluate relations between the transfer functions and human feelings using Semantic Differential method. The results of the SD method show that there are typical forms of transfer functions and parameters to generate comfortable motion for humans. Desirable transfer functions are the first-order or the second-order delay elements in which the parameters must be chosen so that the maximum velocity of generated intelligent machine motion exists in the range of 71-77% of the maximum velocity of instruction movement. In addition, human feelings to the intelligent machine motion generated by human instructions are represented by the factors of “lively”, “mild”, and “interest”.
