Importance of real-world properties in chasing task: simulation and analysis of dragonfly's behavior

摘要

Recently, robots that operate in complex real environments, such as rescue robots and nursing-care robots, have been proposed and developed. However, controlling a robot in a complex environment remains very difficult, and therefore, the study of autonomous control of a robot is one of the most important topics in robotics. In this study, we modeled a dragonfly's behavior and conducted dynamic simulations of chasing behavior. To validate the role of real-world properties, we analyzed the simulation results and showed that by using real-world properties, the computational cost of the autonomous controller can be reduced considerably and interception behavior can be realized by a simple controller without estimating the movement of the prey.