TitleMotion Primitive Based Random Planning for Loco–Manipulation Tasks
Publication TypeConference Paper
Year of Publication2016
Conference NameIEEE International Conference on Humanoid Robots (HUMANOIDS 2016)
AuthorsSettimi, A, Caporale, D, Kryczka, P, Ferrati, M, Pallottino, L
Conference LocationCancun, Mexico, 15-17 Nov. 2016
ISBN Number978-1-5090-4718-5

Several advanced control laws are available for
complex robotic systems such as humanoid robots and mobile
manipulators. Controls are usually developed for locomotion or
for manipulation purposes. Resulting motions are usually executed
sequentially and the potentiality of the robotic platform
is not fully exploited.
In this work we consider the problem of loco–manipulation
planning for a robot with given parametrized control laws
known as primitives. Such primitives, may have not been
designed to be executed simultaneously and by composing
them instability may easily arise. With the proposed approach,
primitives combination that guarantee stability of the system
are obtained resulting in complex whole–body behavior.
A formal definition of motion primitives is provided and a
random sampling approach on a manifold with limited dimension
is investigated. Probabilistic completeness and asymptotic
optimality are also proved. The proposed approach is tested
both on a mobile manipulator and on the humanoid robot
Walk-Man, performing loco–manipulation tasks.

This work is supported by the European commission project Walk-Man EU FP7-ICT no. 611832 and the ECs Horizon 2020 robotics program ICT-23-2014 under grant agreement 644727 (CogIMon)
Refereed DesignationRefereed
PDF icon [PDF]9.89 MB