Application of a new Iterative pseudo-inverse Jacobian Neural Network Matrix technique for controlling geckodrive DC motors of manipulators

Abstract

In controlling arobot's end-effector and joints movement space, one of the most important problems is to know the joints' relative displacements with extreme precision. The control of angular displacements in joints must be done by accurately commanding the movement of each motor in the joints after applying the forward and inverse kinematics problem. While this problem could be solved online, it is usually solved offline. In this paper a very precise proper technique has been developed to solve the inverse kinematics problem in robots with redundant chains. The technique is called Iterative Pseudo Inverse Jacobian Matrix Method (IPIJMM) coupled with Sigmoid Bipolar Hyperbolic Tangent Neural Network with Time Delay and Recurrent Links (SBHTNN-TDRL). Solving the inverse kinematics problem is very difficult when the degrees of freedom increase and in many cases becomes impossible due to the existence of redundant solutions. All such cases must be solved by numerical iterative approximation technique, as our proposed technique combined with intelligence algorithms. In this paper the virtual LabVIEW instrumentation is used to obtain the space curves in different Euler planes by using anarm robot, acquisition board, Geckodrive for each joint, controlled by the proposed algorithm. As a result of applying the proposed technique, the errors of the space movement of the robot's end-effector was less than 0.001 mm. The presented technique and the used virtual instrumentation (VI) are general and can be used for any robot types and application, as well as for any other conventional and unconventional space curves.