Motion planning and control methods for nonprehensile manipulation and multi-contact locomotion tasks

Abstract

Many existing works in the robotic literature deal with the problem of nonprehensile dynamic manipulation. However, a unified control framework does not exist so far. One of the ambitious goals of this Thesis is to contribute to identify planning and control frameworks solving classes of nonprehensile dynamic manipulation tasks, dealing with the non linearity of their dynamic models and, consequently, with the inherited design complexity. Besides, while passing through a number of connections between dynamic nonprehensile manipulation and legged locomotion, the Thesis presents novel methods for generating walking motions in multi-contact situations.