The scenario shows the ZAFH service robots "Kate" and "Larry" acting as butler. Kate takes orders from persons and hands over parts to Larry. While Kate makes a cup of coffee, Larry fetches the sugar dispenser from within a closed sideboard.
Both orders demonstrate important skills such as mobile manipulation, object recognition and task coordination. With Larry opening the door of the sideboard its capability to perform constrained manipulation tasks is demonstrated. This scenario extends the previous butler scenario with Larry assisting Kate.
Both systems are developed using the SmartSoft MDSD Toolchain.
It is one example of how to build a new platform with new tasks by reusing and composing already existing software components as well as already existing action plots for coordination.
The following deployment diagram shows that only few (hardware driver) components had to be exchanged to build the new platform.
One part of the scenario shows an example of an constrained mobile manipulation task. Larry opens a sideboard door to fetch objects from within. The manipulation trajectory planning is done in OpenRAVE using CoMPS for constraints.
Object recognition uses a Kinect 3D sensor and fuses the result of several recognition algorithms. The mesh representation of obstacles is taken into account to safely manipulate in the environment.
Coordination of the components is done with SmartTCL, which handles all contingencies during execution.