Compliant Grasping
Motivation: General use graspers must be designed to perform unknown missions in unknown environments. For autonomous or semi-autonomous operations, many missions require robustness against errors manipulator placement and incomplete object information. The speed with which grasping can be performed is another critical factor in grasping performance, a limitation of current complex anthropomorphic graspers. This research attempts to exploit the advantages of controllable compliance for rapidly grasping a wide range of unmodeled object geometries.
Project Highlights:
- Passive conformability: The grasper is able to rapidly grasp a wide range of objects without sensory data to describe the object geometry. To achieve robustness goals, the grasper passively conforms to the shape of the object as they are pressed against one another. Once an appropriate level of grasp control has been achieved, the conformability of the grasper is actively eliminated to maintain grasp.
- Application of novel locomotion technologies to grasping: Locomotion research has developed unique solutions for achieving enhanced traction on a wide variety of surfaces. The application of such technologies to grasping will expand the space of object surfaces and geometries able to be handled by a grasper.
- Position independence: The grasper is robust to positional errors induced by noisy, erroneous, and/or incomplete sensor data, as well as manipulator placement errors. This large range of acceptable error makes it useful for robots with limited sensors and unsophisticated manipulator arms.