Development of a vector-based fuzzy logic approach for motion planning

Abstract

Due to its potential ability in reasoning linguistic terms, the fuzzy logic approach has been widely applied to robot motion planning. The existing fuzzy logic planners for robot motion planning mainly deal with planar manipulation or navigation. Further, the existing planners operate only on fuzzy scalar variables. However, for a general robot motion in 3-D Cartesian space, both the pose of a manipulator end-effector and the position of a mobile robot relative to other objects or coordinate frames contain both magnitude and orientation information. In this paper, as an attempt towards a general and effective robot motion planning in 3-D space, a new fuzzy logic approach operating on fuzzy vector variables is proposed. For this purpose, a set of new membership functions suitable to operate fuzzy vector variables is defined. Simulation studies on an underwater robotics vehicle (URV) in a simulated oceanic scenario, a typical example operating in 3D space, have demonstrated the effectiveness of this new fuzzy motion planner