Bayesian modeling of the risk of non-repeatability for the networked robotic system

Authors:
Leonard Stepanskiy;Yongjin (James) Kwon
Affiliations:
School of Biomedical Engineering, 3141 Chestnut Street, Drexel University, Philadelphia, PA 19104-2875, USA;Industrial Engineering, Ajou University, Suwon, Zip 443-749, South Korea
Venue:
Computers and Industrial Engineering
Year:
2013

Citing 10
Cited 0

Bayesian Network Classifiers

Machine Learning - Special issue on learning with probabilistic representations
Queueing networks and Markov chains: modeling and performance evaluation with computer science applications

Queueing networks and Markov chains: modeling and performance evaluation with computer science applications
Bayesian inference for polyhazard models in the presence of covariates

Computational Statistics & Data Analysis
Parametric design optimization of 2-DOF R-R planar manipulator-A design of experiment approach

Robotics and Computer-Integrated Manufacturing
Robot measuring form errors

Robotics and Computer-Integrated Manufacturing
Downstream performance prediction for a manufacturing system using neural networks and six-sigma improvement techniques

Robotics and Computer-Integrated Manufacturing
Screening of factors influencing the performance of manipulator using combined array design of experiment approach

Robotics and Computer-Integrated Manufacturing
Noise analysis of robot manipulator using neural networks

Robotics and Computer-Integrated Manufacturing
Experimental kinematic calibration of parallel manipulators using a relative position error measurement system

Robotics and Computer-Integrated Manufacturing
Enhancing e-quality for manufacture using Kalman Filter calibrated visual robotic control

Robotics and Computer-Integrated Manufacturing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This study addresses one of the most important performance criteria, namely the robot's repeatability, from the e-quality for manufacture (EQM) point of view within the framework of networked robotic system. Real-time monitoring and control of remotely located robots allow the operators to continuously assess the risk of robot's non-repeatability. The elaborated methodology of predictive modeling on the risk of robot's non-repeatability consists of three stages: (1) regression analysis on the association between the disturbing factors and the key performance variables that influence the robot repeatability; (2) probabilistic assessment on the admissible deviations of key performance variables that simulate the robot operations as probabilities of job service without failures in the queueing system; and (3) Bayesian assessment on the risk of non-repeatability for the robot operations. The proposed methodology is expected to reduce the risk of robot's non-repeatability, which is better suited for today's networked, distributed production environment, where quality standards are stringent and customer expectations are high.