Predição dos arranques mínimos de um manipulador robótico utilizando aprendizado de máquina para aplicações em cirurgias médicas

Abstract

The usage of robots to perform automated and/or repetitive tasks has been expanding beyond the industrial scope, and these are currently used in areas such as agricultural production, service sectors and hospitals: to ensure that the robot’s performance is the best possible, the field of study focused on trajectory planning optimization was created. One of the biggest problems faced in this field is the modeling of minimum jerks so that robots can be operated without them going into resonance and, consequently, suffering wear in their components: however, this modeling process commonly uses optimization algorithms, such as PSO and NLP. In this work, modeling methods, such as Artificial Neural Networks (ANN) and Machine Learning algorithms, are evaluated to determine the minimum jerks of a robotic manipulator with application in medical surgery, assembled in a simulation. The focus of this work is to prove that simpler and less computationally costly methods can be used for this purpose, with results as good as or better than those of the methods currently used.