Plan execution in a temporal logic environment

This extended abstract describes essential aspects of a current application scenario which was initiated in order to investigate the applicability of research results of the RAP project in a rather real-world environment. The work on this scenario is not completed which means that reported results are rather preliminary. The project RAP (_Reasoning About Plans) (Bauer al. 1996) aims at a logic-based shell which supports the consistent modeling of planning knowledge and provides a generic module for the logic-based reasoning about plans which can be integrated into a large variety of application systems. The RAP shell, for example, enables users to graphically configure their own complex reasoning procedures and control systems based on basic reasoning services like plan generation, plan validation and verification, temporal projection, plan optimization, plan modification, symbolic plan execution, plan interpretation, etc. Our current application scenario is about an autonomous robot indoor simulation environment in an extended logistic setting. There is a building consisting of a shop with several devices for the analysis and manipulation of objects, and there are differently classified storage rooms according to different properties objects can have. The modeled robot tasks include movements, simple processing, storage, and disposal of certain objects. In order to control the activities of the robot we have built a plan execution control system as a special plan reasoning system in the RAP shell. The graphical simulation environment has been integrated as a front-end to the control system. Figure 1