Analysis of environmental factors limiting humans' obstacle avoidance ability

Abstract

We conjecture that action intelligence of humans could contribute to the path-planning problem of autonomous mobile robot. In order to achieve the imitation of such action intelligence on a robot, besides action knowledge acquisition, we need to further learn more about the characteristics of humans' action intelligence. In this paper, from the viewpoint of attributes of an obstacle such as number, size and speed, we focus on analysis of environmental factors that could restrict humans' obstacle avoidance ability. Based on an obstacle avoidance simulator, the characteristics of humans' obstacle avoidance are analyzed and the limitation of humans' obstacle avoidance was verified, especially such limitation could be quantified to some parameters, mainly the number of obstacles in environment. Finally, the guidelines for knowledge acquisition are summarized: under the environment with static obstacles, it is enough to extract the knowledge under the environment with one obstacle; under the environment with mobile obstacles, it is enough to extract the knowledge within limitation of humans for obstacle avoidance.