Ant Colony Optimization and image model-based robot manipulator system for pick-and-place tasks

Abstract

A visual servo control system combines with the model-based image segmentation and an Ant Colony Optimization (ACO) algorithm to design an excellent six-Degree-of-Freedom (6-DOF) robot manipulator for solving the complicated combinations of pick-and-place tasks. A simple but efficient vision-based segmentation methodology is developed to extract the object information by getting appropriate feature of the controlled platform when the robot is tracking the manipulated image patterns. The evolutionary ACO learning algorithm explores the near-optimal path selections to drive the 6 ROF robot arm kinematics model for completing the Pick-and-Place tasks as soon as possible. Inverse orientation kinematic machine is proposed to successfully guide the robot manipulator into the desired position. Several software simulations include image segmentations, the shortest path selection, and the performance validation in various experiments. These results are described and presented to demonstrate that the designed image model-based robot manipulator wins the excellent Pick-and-Place task. Not only the software simulation, the practical robot synchronously performed in real-world to reach the higher feasible functions in the eye-to-hand experiments.