A non-contact control architecture for micro-components assembly

This paper presents a control system to assemble hybrid micro-components with a micro-factory. The assembly system is based on image processing techniques that feed relevant information to an automata based control system. The scene containing the micro-components that must be assembled is analysed taking advantage of the finite number of typologies of components that permits to implement a statistical based recognition and classification approach. The objects features, such as their position and orientation, together with the status of the micro-factory actuators provide the inputs to an event-driven finite state automaton that implements the assembly procedures previously established.