Psychological Evaluation for Rough Shape and Biped Walking of Humanoid Robots Using Virtual Reality

Abstract

Japan is becoming an aging society composed largely of elderly people, and the proportion of aged people in the population is increasing year by year. There is an increasing need for robots that can coexist with people and help them in their daily lives: housekeeping robots, nursing care robots, etc. Such robots are required to have “physical safety” and “mental safety”. Physical safety means that robots do not injure humans. Mental safety means that humans do not feel fear or anxiety toward robots. In addition, it is important that humans do not have feelings of unpleasantness or aversion toward robots. Accordingly, when designing robots coexisting with people and planning their motions, it is necessary to consider the influences of the robots on human impressions and psychology. Mental safety has not yet been fully discussed. This is because the parameters of robots (shape, size, color, motion, speed, etc.) that may affect human psychology have not been clarified, and the method of measuring and evaluating human psychology for the robots has not been established. There have been some researches on the evaluation of human psychology about robots and the interaction between robots and humans. Goetz et al. conducted questionnaire about matching between tasks and appearance of humanoid robots (Goetz et al., 2003). Robins et al. investigated the attitudes of autistic children for robots (Robins et al., 2004), and Woods et al. discussed the design of robots from children’s viewpoint (Woods et al., 2004). Some psychological experiments for wheeled humanoid robot “Robovie” were conducted (Kubota et al., 2003; Kanda & Ishiguro, 2004; Sakamoto et al., 2004). Seal robot “Paro” was also psychologically evaluated worldwide (Shibata et al., 2003; Shibata et al., 2004). Kanda et al. investigated the impressions on real robots with different appearance by the semantic differential method when they did the same tasks (Kanda et al., 2005). These researches are significant because they evaluate the psychological suitability of the existent robots as partner robots or human-friendly robots. But they are not enough to analyze which parameters of robots bring desirable psychological effects. For the purpose of clarifying the relationship between the parameters and their psychological effects, we have proposed the evaluation of human impressions and