Distributed Manipulation of Flat Objects With Two Airflow Sinks

Distributed manipulation systems induce motions on objects through the application of forces at many points of contact. Current forms of distributed manipulation include multiple mobile robots, vibrating plates, actively controlled arrays of air jets, and planar micro- and macro-mechanical arrays of actuators. The authors have presented a new form of distributed manipulation using passive airflow fields. This paper lays out infrastructure for manipulation algorithms using logarithmic potential fields applicable to passive airflow distributed manipulators. It uses a line-integral form of the lifted force equations, and provides a numerical approach to check the uniqueness of the robust pivot point for given objects in a logarithmic potential field. The numerical method is proved analytically to require a finite resolution to find all robust pivot points. It also proposes a squeeze-like sequential manipulation algorithm to bring an object with a unique robust pivot point to a unique final pose using airflow fields without sensors. The algorithm has been verified by experiments which are conducted for three different starting orientations, and end up with a unique final pose at the end of the manipulation sequence