25-07-2012, 04:27 PM
Novel Artificial Neural Network Application for Prediction of Inverse
Kinematics of Manipulator
Novel Artificial Neural Network.pdf (Size: 849.77 KB / Downloads: 54)
ABSTRACT
The robot control problem can be divided into two main areas, kinematics control (the
coordination of the links of kinematics chain to produce desire motion of the robot), and dynamic
control (driving the actuator of the mechanism to follow the commanded position velocities). In
general the control strategies used in robot involves position coordination in Cartesian space by
direct or indirect kinematics method. Inverse kinematics comprises the computation need to find
the join angles for a given Cartesian position and orientation of the end effectors. This
computation is fundamental to control of robot arms but it is very difficult to calculate an inverse
kinematics solution of robot manipulator. For this solution most industrial robot arms are
designed by using a non-linear algebraic computation to finding the inverse kinematics solution.
From the literature it is well described that there is no unique solution for the inverse kinematics.
That is why it is significant to apply an artificial neural network models. Here structured artificial
neural network (ANN) models an approach has been proposed to control the motion of robot
manipulator. In these work two types of ANN models were used.
INTRODUCTION
Robot manipulator is composed of a serial chain of rigid links connected to each
other by revolute or prismatic joints. A revolute joint rotates about a motion axis
and a prismatic joint slide along a motion axis. Each robot joint location is usually
defined relative to neighboring joint. The relation between successive joints is
described by 4X4 homogeneous transformation matrices that have orientation and
position data of robots. The number of those transformation matrices determines
the degrees of freedom of robots. The product of these transformation matrices
produces final orientation and position data of an n degrees of freedom robot
manipulator. Robot control actions are executed in the joint coordinates while
robot motions are specified in the Cartesian coordinates. Conversion of the
position and orientation of a robot manipulator end-effectors from Cartesian space
to joint space, called as inverse kinematics problem, which is of fundamental
importance in calculating desired joint angles for robot manipulator design and
control. In most robotic applications the desired positions and orientations of the
end effectors are specified by the user in Cartesian coordinates.
Background of the work
In this paper, some methods of artificial neural network applied for the
solution of inverse kinematics of 2-link serial chain manipulator. The methods are
multilayer perceptrons and polynomial preprocessor neural network has applied.
The main objective of this thesis is to predict the values of joint angles (inverse
kinematics), as we know that there is no unique solution for the inverse
kinematics even mathematical formulae are complex and time taking so it is better
to find out solution through neural network. There are so many methods in softcomputing,
but in this paper two methods has been taken. After validation of
these methods, we multilayer perceptrons giving better result.
Methodology
In this paper the researchers has proposed two methods for the solution of
inverse kinematics of manipulator, the proposed methods are multilayer
perceptrons and polynomial preprocessor in order to validate the performance of
MLP and PPN for inverse kinematics problem, simulation studies are carried out
by using MATLAB. Many researchers have followed MLP, PPN, RBF and
FLANN with MISO (multi input single output) system. Here in this paper we
have applied MLP and PPN with MIMO (multi input multi output) system. A set
of 130 data sets were first generated as per the formula equation (10) for this the
input parameter X and Y coordinates in inches. Using these data sets was basis for
8
the training and evaluation or testing the MLP and PPN models. Out of the sets of
130 data points, 100 were used as training data and 30 were used for testing for
MLP. Back-propagation algorithm was used for training the network and for
updating the desired weights. In this work epoch based training method was
applied.
Summary
The robot motion problem involves in bringing the end-effectors of the
manipulator from the present to the desired position and orientation in the global
coordinates while following a prescribed trajectory in either the joint coordinates
or global coordinates. Since the desired position is usually specified in the global
coordinates, whereas the actuators used to drive the system are to be commanded
with desired joint values, the inverse kinematics must be solved. There are several
types of soft computing methods available which can be used for finding the
solution of the inverse kinematics and further we’ll discuss about them in next
chapter.