Identification of Efficient Industrial Robot Selection (IRS) Methods and Their Performance Analysis

Abstract

The quick advancement of mechanical robots and their utilization by fabricating businesses for a variety of applications could be a basic errand for robot choice. As a result, the industrial robot selection (IRS) process for potential clients gets to be greatly complicated since they have to get to numerous parameters for the robot accessible. In this article, a new predictive model-based IRS technique is proposed, and six different optimization techniques are tested using industrial robot specifications. Partial least square regression (PLSR), principal component regression (PCR), scaled conjugate gradient-based backpropagation method, gradient descent with momentum-based backpropagation, fuzzy topsis, and a case-based approach are some of the optimization models examined in this suggested study. As a whole, 11 distinct factors are taken into account as inputs in this suggested technique, and Robot Rank is used to identify the best robot (RR). Using the suggested method, the rank of the preferred mechanical robot is determined from the absolutely best possible robot, providing the most accurate benchmark for robot selection for the given application. Additionally, MSE: mean square error, RSE: R-squared error, and RMSE: root mean square error are used to evaluate the effectiveness of the robot selection methods.