Abstract
A fixed geometric trajectory ensuring the success of an assembly task to be performed by a robot cannot be determined a priori when the geometric uncertainty is significant. In order to decrease the influence of the geometric uncertainty, passive compliance devices have been extensively applied; nevertheless, the general solution appears to be the use of active compliance, which involves reaction force/torque feedback, and gives rise to some particular problems in task planning and execution. Configuration and force/torque sensory information becomes a key point in this field. This chapter describes how this sensory information can be used in an assembly fine-motion planner and also during task execution; two approaches, one analytical and the other learning approaches, are presented and compared.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
