Control of a redundant manipulator using fuzzy rules

Abstract Sophisticated robot tasks require a kinematical redundancy which enables the robot links to avoid obstacles whereas the effector follows a planned path. In this connection the solution of the inverse problem , i.e. the determination of the internal joint coordinates from a given position in the operational space (e.g. Cartesian coordinates), is an essential problem. The special relations between external and internal coordinates lead to an introduction of restrictions for the internal joint coordinates depending on appropriate criteria such as distances of the links from the boundaries as well as from external obstacles. These criteria lead, in a classical way, to corrections from which the desired robot motion can be obtained by using the pseudo inverse Jacobi matrix. A new method of choosing appropriate joint angles using fuzzy logic is presented. In this connection distances and corrections are denoted as fuzzy terms. In this way various criteria for controlling a redundant arm are formulated linguistically by fuzzy production rules. The advantage of the fuzzy method is evident in the presence of multi-criteria decisions. Simulation results demonstrate the practicability of the method.