Nonlinear path planner for industrial manipulator

Abstract

Consulta en la Biblioteca ETSI Industriales (Riunet)

[eng] This project investigates path planning algorithms for industrial manipulators. The implementation of any path planning strategy relies on the understanding of the kinematic and kinetic characteristics of the robot. Hence the first part of the thesis deals with the development of dynamical models. The chosen planning strategy minimize the energy consumption and the travelling time for a specific path which is build by a sequence of points. First the kinematic model of the manipulator is characterized using direct and inverse kinematics based on the Denavit-Hartenberg notation. The complex inverse kinematic problem for the six revolute joints manipulator is solved by applying the algorithm of Manocha and Canny. With the kinematic model in place, the effective workspace are and the singularities inside it are determined. The six degrees-of-freedom kinetic model of the industrial manipulator is then derived within the Lagrangian mechanics framework. In the second part, two path planning methods are presented. The first of them, based on potential fields techniques, treats the manipulator as a set of points attracted or repelled by sinks or obstacles depending on the robot configuration, and it exploits the kinematic model. The second method is based on the optimization of a specific cost function, and it uses the kinetic model that take into account the dynamics of the system. Both methods have been tested trough simulations in the MATLAB environment, providing the testing results. The robotic platform taken as a reference for this project is the industrial manipulator UR5 provided by the Danish company Universal Robots.