Ein flexibles Greifsystem für Roboterassistenten im Haushalt

The goal of this thesis was to find approaches for flexible gripping systems with low complexity. The gripping system should be enable robot assistants to interact with their environment and to adapt to different situations regarding the handling of objects. To reach this goal the so called Finray-effect, a bionic principle for self adapting systems, was theoretical investigated, kinematics models and planning algorithms were developed and experiments were conducted to ensure the feasibility for the use in flexible gripping systems. Starting from the state of the art of robot assistants, industrial gripping systems and artificial hands three scenarios for robot assistants in household environments were analysed. For the subsystems of flexible grippers different concepts were developed. An innovative solution consisting of three fingers was developed and combined with the above mentioned Finray-effect. Due to the lack of basic theoretical information about the Finray-effect in combination with flexible grippers the basic geometrical and kinematical coherences were developed. Therefore the kinematics of the Finray-effect was described as a set of links. With this model it was possible to describe a finger which consists of any number of links. Based on this model equations which describe the movement of the finger and the resulting forces were developed. To get information about the behaviour of the Finray-effect while gripping an object several simulations were conducted. Based on the developed mechanical model and simulation results an iterative algorithm was developed which delivers the resulting angles and forces between a finger and an object. To investigate the feasibility of the developments two gripping systems were realized and evaluated regarding their possibilities and their limitations: The IPA-hand I and the IPA-hand II. The IPA-hand II consists of the self-adaptive finger kinematics. Experiments with the IPA-hand II show a very good flexibility regarding different object geometries. The very high spectrum of objects which can be handled and the low complexity of the IPA-hand II makes it very interesting for gripping devices of robot assistants