Skip to main content
European Commission logo print header

Whole-body Adaptive Locomotion and Manipulation

Project description


Robotics, Cognitive Systems and Smart Spaces, Symbiotic Interaction

The WALK-MAN project aims to develop a humanoid robot that can operate in buildings that were damaged following natural and man-made disasters. The robot will demonstrate new skills:
-Dextrous, powerful manipulation skills - e.g. turning a heavy valve of lifting collapsed masonry,
-robust balanced locomotion - walking, crawling over a debris pile, and
- physical sturdiness - e.g. operating conventional hand tools such as pneumatic drills or cutters.
In addition, the robot will have the sufficient cognitive ability to allow it to operate autonomously or under tele-operation in case of severe communication limitations for remote control due to limited channel bandwidth and/or reliability).
The robot will show human levels of locomotion, balance and manipulation and operate outside the laboratory environment. Disaster sites may include buildings such as factories, offices, houses.
This project will significantly advance the current walking and locomotion capabilities of humanoid systems so that the robots will be able to walk in and through cluttered spaces (walking in a crowded environment) and maintain their balance against external disturbances, such as contact and impacts with objects or people. The robot will achieve this by exploiting all its limbs (hands, arms, legs, feet and trunk) to demonstrate whole-body motion dynamics and making use of surrounding workspace constraints (handrails, walls, furniture etc).  In parallel, we plan to advance the manipulation capabilities of existing humanoids by developing new hand designs that combine robustness and adaptability. This will take advantage of recent developments in mechanical design and materials that allow the creation of less fragile and delicate end-effectors that are capable of grasping/manipulating traditional handtools.
To achieve these goals, we the project will develop four powerful enabling ideas:
1) an integrated approach to whole-body locomotion and manipulation (termed loco-manipulation), where all body parts (arms, hands and legs) can be used to ensure more stable and balanced motion. This combined use of arms and legs to create motion (arms and legs working together to achieve a task beyond the limits of a single limb) will also allow the robot to produce the large manipulation forces needed in a disaster environment (eg to move rubble and heavy masonry); 
2) the development of a system of loco-manipulation behaviours that control the  robot’s perception, cognition and action;
3) the use of soft, compliant actuator technologies, to provide more natural adaptability, interaction and robustness, and
4) efficient planning algorithms exploiting a robust and consistent control hierarchy based on the theory of motion description languages and symbolic control.

Despite the remarkable developments in robotics, and the many years and the millions Euros spent, robotics could not yet demonstrate readiness to help in recent catastrophic situations. This project aims to develop a humanoid that will be able to operate in de-structured spaces, demonstrating apt and strong manipulation, robust locomotion, and physical sturdiness. We aim to advance the locomotion capabilities to permit humanoids to walk and balance against disturbances by exploiting their whole body motion and the workspace constraints. We target to advance the manipulation skills of humanoids by looking on new hand that combine robustness and adaptability allowing less delicate use of the hands for manipulation as well for locomotion/balancing assistance. We rely on four powerful enabling ideas: 1) an integrated approach to whole-body locomotion and manipulation (loco-manipulation), where all body parts contribute to produce and balance motion, as well as to produce the large manipulation forces needed in a disaster environment; 2) loco-manipulation affordances mediating perception, cognition and action; 3) soft robotic actuators technologies, to afford adaptability and robustness, and 4) ) efficient any-time planning algorithms exploiting a robust and consistent control hierarchy based on the theory of motion description languages and symbolic control. We will define stringent and plausible validation scenarios, including but not limited to those of the DARPA robotic challenge. The Consortium will generate a European entry to the DARPA Robotic Challenge to participate in the final trials in December 2014, to which we have been officially invited. However, our work will not be targeted to a system adjusted to the requirements of this or any specific challenge. In the second part of the project civil defense corps will define specifications for real-world use. Our true goal is when the need will unfortunately arise again, robotics will be there to help.

Call for proposal

FP7-ICT-2013-10
See other projects for this call

Coordinator

FONDAZIONE ISTITUTO ITALIANO DI TECNOLOGIA
EU contribution
€ 2 131 433,00
Address
VIA MOREGO 30
16163 Genova
Italy

See on map

Region
Nord-Ovest Liguria Genova
Activity type
Research Organisations
Administrative Contact
SIMONA VENTRIGLIA (Ms.)
Links
Total cost
No data

Participants (4)