부산대학교 도서관

This paper presents a four-wheel-steering four-wheel-drive (4WS4WD) car-like robot, able to autonomously explore unknown environments with flat or uneven terrain. The robot uses a set of sensors and actuators and employs a number of algorithms to effectively map the environment, localize in it, plan and execute motions while utilizing the maneuverability of the four-wheel-steering kinematic model. For the task of autonomous navigation, a ROS-based 4-Tier architecture is proposed consisting of global path planning, dynamic local path deformation for collision avoidance, kinematically feasible local path synthesis and finally path tracking control using a fuzzy logic controller for 4WS car-like robots.