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:
- 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);
- the development of a system of loco-manipulation behaviours that control the robot’s perception, cognition and action;
- the use of soft, compliant actuator technologies, to provide more natural adaptability, interaction and robustness, and
- efficient planning algorithms exploiting a robust and consistent control hierarchy based on the theory of motion description languages and symbolic control.
To ensure the goals are realistic and achievable, we will define strict, real world validation scenarios, that can include (but will not be limited to) those of the DARPA robotic challenge. The Consortium will aim to produce a European entry to the DARPA Robotic Challenge (DRC) to participate in the final trials in December 2014, to which we have been officially invited. However, our work will not bedrivenby the requirements of this challenge or by any agenda other than those set within the EU research framework programmes. It is important to realise that the aims of this project reach far beyond the DRC both in time and in ambition. In the second part of the project, civil defence bodies will be consulted to help define specifications for a true Real-World Challenge (RWC), with realistic and realisable scenarios.
Our ultimate goal is that, when the need will (unfortunately) arise again, robotics will up to the challenge.