Robot motion coordination by fuzzy control

Abstract

Conventional Robot Motion Coordination is usually based on the tasks of path planning, coordinate transformation and the solution of the inverse kinematic problem. This concept is widely used in industrial applications, where pre-determined paths have to be accurately followed (cartesian path planning) or a Point To Point movement is to be executed without considering any defined cartesian path (high speed movement). The approach presented in this paper prepares a way between the pure cartesian and pure joint-coordinate based motion coordination. By implementing a rule base system for the generation of trajectories, the solution of the inverse kinematic problem can be avoided. In exchange, a Fuzzy control loop is closed between the actual position of the Tool-Center-Point, which is determined by the forward kinematics, and the joint drives. Following this strategy, the joint- and link-configuration can be controlled by a set of linguistic rules, which enables a new approach for the solution of robot body-collision avoidance problems and coordinate transformation problems in connection with redundant manipulators for future investigations. In this paper, a new rule based motion coordination concept is presented, and the properties of the method are demonstated by means of a simulation.