A rule-based algorithm to predict the dynamic behavior of mechanical part orienting operations

Mechanical part orienting approaches, such as part feeders and manipulator pushing operations, reduce the need for sensors. However, their dynamic behavior is usually difficult to predict due to unpredictable kinematic constraint changes. The authors present a rule-based algorithm to predict the dynamic behavior of these systems automatically. Since the rules allow algorithm to make decisions, no data relating to the kinematic constraint changes or subsequent reformulation of the dynamic equations of motion are required a priori. The rules are developed on the premise that the boundary geometry of the interacting bodies is an integral component of the system dynamics. A part feeder, developed using this algorithm is presented.<
>