Abstract:
Although the Delta robot has many applications in pick-and-place operations, some
problems limit its use. The difficult programming is one of these limitations. Parallel
robot programming depends on solving the forward and inverse kinematics equations
of the robot. These equations relate the geometric parameters, such as length and angle
of every robot arm, to the position of the end effector and vice versa. The kinematic
equations are hard to be derived. Moreover, any change in the robot geometry, due to a
change in the application condition, requires that new kinematic equations be derived.
This needs a very sophisticated and specialized programmer, who is not always
available. Consequently, this problem limits the use of parallel robots. This paper
discusses the use of ANN and embedded systems in addition to stereo vision to
command delta robot in pick-and-place operations. The method is implemented and
tested giving satisfactory results.
