Motion Planning for Manipulators with Many Degrees of Freedom – The BB-Method

Baginski, B.
Pub. date
January 1999
195 of Dissertations in Artificial Intelligence
ISBN print
Artificial Intelligence, Computer & Communication Sciences, Computer Science

The motions of robot manipulators in industrial applications are still programmed by human operators. To allow a more flexible use of robots, it is desired to extend the capabilities of robot systems. Given information of the objects surrounding it, a robot has to be able to plan its own motions without collisions, without other interferences, and with high efficiency.

This book introduces a new approach for manipulator motion planning. The principle idea is to consider an entire path as the subject of the planning process. Rating functions are employed to measure the quality of a path with respect to the different objectives. Planning itself becomes the reshaping of the path to improve it.

The planning method is able to handle arbitrary manipulators with any number of degrees of freedom successfully. Planning is based on the CAD-data of the robot and its surroundings, and does not require specific preprocessing or limiting preconditions. The quality of the resulting collision-free motions is very high. They are straight and short - thus usually fast and cheap to be moved along. They keep a safety distance to obstacles wherever possible - thus allowing to compensate for modelling and control uncertainties.