Optimization of Industrial, Vision-Based, Intuitively Generated Robot Point-Allocating Tasks Using Genetic Algorithms

Abstract

Current industrial robot-programming methods require, depending on the task to be developed, an elevated degree of technical ability and time from a human operator, in order to obtain a precise, nonoptimal result. This correspondence paper presents a methodology used to generate an optimal sequence of robot configurations that enable a precise point-allocating task applicable, for instance, to spot-welding, drilling, or electronic component placement maneuvers. The optimization process starts from a nonoptimal, initial sequence designated intuitively by a human operator using an easy-to-use interface. In this correspondence paper, intuitive programming is considered as the process of defining, in a computer graphics environment and with a limited user knowledge of robotics or the industrial task, the sequence of motions that enable the execution of a complex industrial robotic maneuver. Such an initial sequence is later followed by a robot, very precisely, using a vision-based, calibration-free, robot control method. Further robot path optimization is performed with a genetic algorithm approach. An industrial robot, which is part of the experimental setup, was used in order to validate the proposed procedure.

Path-Tracking Maneuvers With Industrial Robot Manipulators Using Uncalibrated Vision and Impedance Control

“…Current research involving the development of systems for intuitive industrial-robot programming benefits from the use of calibration-free, vision-based techniques [24]. Among them, a method known as camera-space manipulation (CSM) [25]- [28] has demonstrated its effectiveness, especially in the cases where the cameras used to control the maneuver remain fixed in the interval between workpiece detection and task culmination.…”

Hierarchical resource allocation for integrated modular avionics systems

“…Genetic algorithms (GAs), which are parallel and global search techniques for optimization, have been widely used in many fields [19][20][21]. However, the GA is not good at local search and may lead to a premature convergence.…”

Robot Positioning Using Camera-Space Manipulation With a Linear Camera Model

“…Those proposed controllers with joint-velocity estimation can achieve very satisfactory system performance and guarantee the overall closed-loop stability of robot systems in joint-space control. Most importantly, they provided a systematic framework to analyze the overall closed-loop system stability with the help of Filippov solution concept [27] using the so-called reduced-order manifold dynamics. Furthermore, sliding observers are robust to dynamic uncertainties of robot manipulators, which is useful for our system designs under the assumption that the robot dynamics is unknown.…”