PhD students are encouraged to collaborate with other projectes related to their topic and IRTG SIGI. On this page, we present these collaborations and highlight the relation to IRTG.
BAALL overview : Though Manfred Eppe is fully funded by IRTG, there exists a strong collaboration between him and the BAALL-project. The Bremen Ambient Assisted Living Lab is an apartment suitable for the elderly and people with physical or cognitive impairments. With a size of 60 m², it comprises all necessary conditions for trial living, intended for two persons. It is situated at the Bremen site of the German Research Center for Artificial Intelligence (DFKI) in one of the labs of the research department Cyber-Physical Systems in the Cartesium building.
BAALL has been developed in cooperation with Universität Bremen in the SFB/TR8 Spatial Cognition of the Deutsche Forschungsgemeinschaft and in projects funded by EU and BMBF respectively.(source: http://baall.informatik.uni-bremen.de)
Association with IRTG initiative A: Ambient Intelligence Environments such as the BAALL consist of various sensors and actuators such as person trackers, automatic doors, smart kitchen devices and the like. In the case of the BAALL there is also the autonomous robotic wheelchair Rolland which is designed with a great emphasis on indoor navigation in the BAALL and other narrow environments.
The mobility of Rolland and the functionality of other actuators and sensors in a spatial environment like the BAALL demands for a formal semantic treatment of the question “What can be done where and under which circumstances?”. A research field which accounts for a formal treatment of this question is planning, as known from Artificial Intelligence.
Manfred’s Research: Most current A.I. planning approaches assume a completely known world state. However, in highly dynamic environments such as Ambient Assisted Living it is often not the case that every property of every detail in the environment is known. For example, consider the case where the wheelchair has the task to pick up a person and bring the person to its target destination. In Smart Environments, it is not desirable to continuously monitor a person’s location due to privacy issues. Thus, if it is unknown where the person is, the wheelchair first has to find out the person’s location and can then plan to pick it up and bring it to the destination. To account for incomplete knowledge, sensing has to be integrated in an interleaved process of planning and plan execution. Manfred’s research tackles the question about how to interleave the planning and plan execution in dynamic environments such as the BAALL. The same approach can also be applied for other dynamic scenarios such as emergency response planning.