Selection of Commercial Robots with Anticipated Cost and Design Specifications using Regression Models

Abstract

The selection of robots used for industry purpose is a crucial practice where various parameters have to be considered during appropriate selection process. The decision strategy of robot selection has a potential research direction to justify the necessity of industrial needs. We have compared three different mathematical models and selected the best method for choosing the industrial robot to provide a complete selection framework to the present article. Principal Component Regression (PCR), Partial Least Square Regression (PLSR) and Linear Regression using Feed Forward Neural Network (FNN) are the three mathematical models used to correlate input with output parameters. During the testing procedure, eleven numbers of distinct parameters are considered to estimate the best possible rank selection. The strata or rank of the robot is approximated by utilizing the proposed algorithm. However, the most approved rank has met the desired genuinity for a targeted application. In addition to the mathematical methodologies applied here, the performance characteristics for selecting the robot is examined by assessment of statistical errors namely Mean Square Error (MSE), Root Mean Square Error (RMSE), and R-Squared Error (RSE).