Initial population generation strategy for evolution of planar hybrid manipulators

摘要

A strategy for generation of random population of planar hybrid manipulators with specific number of degrees of freedom (dof) is proposed in this paper. For a given number of dof and fixed maximum number of links in use, the possibilities of generating diverse hybrid morphologies is explored. The aim of the work is to randomly generate basic morphologies of hybrid robotic manipulators with specific number of dof, to serve the evolution of an optimal mechanism for a given task using evolutionary algorithms. This paper focuses at the former part of the problem and presents an algorithm for population generation. A concept of Mechanism Assembly Matrix is introduced which helps both for random generation and for representation of the morphologies. Some specific number of cell locations of the proposed matrix are randomly selected and filled with one to represent a corresponding joint. Collectively, these cells represent one morphology. The planar mechanisms obtained from the random generation are successfully matching to the possible solutions for the given number of degrees of freedom.