Intelligent Strategies for Compliant Robotic Assembly

Abstract: A review of compliant-motion control strategies is presented, applicable to robotic assembly of industrial components. It is argued that the appropriate framework in which to exert intelligent controls is the use of a virtual attractor and a virtual impedance. In this context, multiple strategies are reviewed, including blind search, interpretation of C-space boundaries, event detection and mode switching, and autonomous learning using a genetic algorithm. It is shown that with such techniques, example components that are currently assembled manually can be assembled faster, gentler and more reliably by robots.