Normalized Ordinal Distance: A Performance Metric for Ordinal, Probabilistic-Ordinal, or Partial-Ordinal Classification Problems

摘要

In the present-day scenario of automated industries, the use of robots has increased exponentially. However, the use of robots is limited by the difficulty of positioning the end effectors as per requirements. Previously, robots were physically positioned by an operator, but with automation, robot control has gone into the hands of computers. Robot kinematics is an essential concept in the positioning of robotic manipulators. Generally, for serial manipulators, direct kinematics is easy, and for parallel manipulators, inverse kinematics is simple. The present work tries to take advantage of this, and inverse kinematics is calculated for a serial manipulator and direct kinematics for a parallel manipulator (i.e., the more difficult ones) using backpropagation neural networks. A simple interface has been developed using Visual Basic for easier interaction.