DefinePK

Abstract

This paper presents two novel methodological approaches to calculate the extrinsic parameters of cameras and the inverse kinematics of robots, two aspects commonly addressed in autonomous robotic operation that require performing three-dimensional (3D) reconstruction and object gripping. The functionality of these new approaches is demonstrated, and a comparison with current methodologies is performed. The first methodology detail how to calibrate the extrinsic parameters of a camera from linear regression, while the second explains how to find the inverse kinematics using the Newton-Raphson numerical method, also, a metric is presented to define the error between two homogeneous matrices. With this, we can show that it is possible to obtain a matrix of extrinsic parameters with an error of , and observe that the calculation of the inverse kinematics with the proposed method reduces the error on average by 99.8% concerning the conventional way.