Reduction of Information Requirements in Mechanical Systems

Todd Murphey

University of Colorado - Boulder

ABSTRACT: In autonomous systems task assignment, control, and sensing needs must be balanced against mechanical and sensing capabilities. In many cases this balance of information and physics is explicit--by designing a mechanical system with information requirements in mind, the information requirements can drastically change. This has been well known for several decades in the sensorless manipulation community where problems considered are typically static or quasi-static in some sense. I will briefly discuss some examples of systems where dynamic mechanical control properties have a direct role in determining necessary sensing capabilities, including self-assembly and coordination. Then I will discuss general results in kinematic reduction of mechanical systems that lead to a natural reduction in the uncertainty description of a mechanical system that experiences stick/slip transitions. I will finish with a discussion of how these results can be applied to vehicle and manipulation planning.

    
Deepak Ramachandran
Last modified: Fri Aug 8 11:38:17 CST 2006