Especially in robotics contexts, even seemingly simple planning problems can pose dire challenges. Plans need to be created frequently in short time. We address this challenge modeling the planning domain and proble, as well as the search, directly using Answer Set Programming (ASP). In particular, we model the search to only yield a partial plan with a limited lookahead and apply a useful metric to ensure progress over several planning iterations. We use the multi-shot solving capabilities of the Clingo solver to tackle the most pressing performance issues.

Robots often plan for the same or very similar goals. We exploit this by storing successful plans in a database and then extracting frequent action sub-sequences. We than calculate macro operators to speed up further planning. In this work, we support the STRIPS subset of PDDL.

To bridge the gap between the robotics and planning communities, and to provide a common evaluation testbed in crossover researc in these areas, we have created the Planning and Execution Competition for Logistics Robots in Simulation. This paper gives an overview of that competition.

During operation of robots large amounts of data are produced and processed for instance in perception, actuation, or decision making. Nowadays this data is typically volatile and disposed right after use. But this data can be valuable and useful later. Therefore we propose a database system that taps into common robot middleware to record any and all data produced at run-time. We present two examples using this data in fault analysis and performance evaluation and describe real-world experiments run on the domestic service robot HERB.