AL SHAFIQ HAKIMI BIN ZAINAL FITHRI

1018147

Literature review on inverse kinamatic

Inverse kinematic solutions are used in manipulator controllers to determine corrective joint motions for errors in end-effector position and orientation(Wampler,1986).

As Wampler (1986) explain: In general terms, an inverse kinematic problem for a serial-link manipulator is the problem of finding joint motions that will produce a given motion of the end-effector. The solution can be used in the control of the manipulator to determine joint motions that will correct measured errors in the position and orientation of the end-effector. For example, resolved-rate control [1] is based on an inverse velocity/angular velocity solution, and resolvedacceleration control [2] is based on an inverse acceleration/ angular acceleration solution. An inverse kinematic solution is also useful when a human operator specifies the velocity of a telemanipulator's end-effector via a joystick or other device. Closed-form solutions to the problem of finding joint positions corresponding to a given end-effector position and orientation are known for only a few simple nonredundant manipulator geometries, but for redundant arms and nonredundant arms with complex geometries, the inverse position/ orientation problem can be solved iteratively using inverse velocity/angular velocity solutions [3]. To date, inverse kinematic solutions have not found widespread use due to three primary reasons:

1) nonredundant arms with closed-form inverse position/ orientation solutions are sufficient for many applications; 2) existing methods for computing inverse kinematic solutions based on the Jacobian matrix are not efficient; and 3) these methods suffer from numerical difficulties near con-figurations where the Jacobian matrix is ill-conditioned (i.e., at kinematical singularities). The methods presented in this paper address the latter two problems, thereby opening the door to the practical investigation of the advantages offered by redundant arms and nonredundant arms with unconventional geometries.

But, this inverse kinamatic also have problem. So, a lot of studies are done to improve this technique. The problem include Real-Time Inverse Kinematics Techniques, Calibration of Stewart Platforms and Other Parallel Manipulators by Minimizing Inverse Kinematic Residuals, Redundant Manipulators, Singularity Analysis of Closed-Loop Kinematic Chains, .comiputational techsioue for Inverse Kinematic, Vector Formulations and Damped Least-Squares Methods.

Computer-controlled robots are usually servoed inthe joint space, whereas the desired task is normally expressed in a "world" coordinate system (taskoriented space). Therefore, some computations must be performed to achieve the required position and orientation of an end effector. This is known as the inverse kinematics problem, which deals with the transformation of coordinates from a world to a joint space (McKerrow, 1991).

Generally, the inverse kinematics problem for robots with offset and reduced wrists can be solved by using a variety of techniques (inverse transform, screw and quaternian algebra, dual matrices, geometrical methods and others). Detailed descriptions of these techniques have been given by Paul et al. (1981), McKerrow (1991), Craig (1989), Lee et al. (1991) and other authors. But, as has been shown by Ragavan and Roth (1990), the final result can be represented by a high-degree polynomial equation in one unknown, and closed-form expressions for the remaining unknowns.

Parallel manipulators have attracted considerable interest in the robotics research community. Two special issues were published in the Journal of Robotics Systems.1, 2 This paper concentrates on the alibration of parallel manipulators in general and Stewart platforms in particular (Hanqi, Jiahua & Masory, 1989).

Many methods of solution for this problem have been discussed over the past 30 years and, as computational speeds have improved, there has been greater expectation that this calculation will be performed online. This paper compares the merits of many of the methods already presented and describes a newapproach that leads to a fast and numerically wellconditioned algorithm (Lucas, Tischler & Samuel, 2013)

As Gosselin (1990) explain: A closed-loop kinematic chain consists of a set of rigid bodies connected to each other with joints where at least one closed loop exists. The chain is also characterized by a set of inputs (denoted here by an n-dimensional vector e), which correspond to the powered joints, and by a set of output coordinates (denoted here by an m-dimensional vector x). These input and output vectors depend on the nature and purpose of the

kinematic chain. For instance, in a parallel manipulator, the input vector I9 represents the set of actuated joints, and the output vectorx represents the Cartesian coordinates of the gripper. However, in general, the output need not be a set of Cartesian coordinates and can also correspond to joint angles or displacements. Furthermore, although the number of inputs and outputs does not have to be equal, the number of independent inputs and outputs will always be the same, except in the presence of redundancies

Geometry of a manipulator link

Tool vectors for a two-fingered gripper

Computer-controlled robots are usually servoed in the joint space, whereas the desired task is normally expressed in a "world" coordinate system (taskoriented

space). Therefore, some computations must be performed to achieve the required position and orientation of an end-effector. This is known as the inverse kinematics problem, which deals with the transformation of coordinates from a world to a joint space (McKerrow, 1991).